Discussion:
Regarding Apple's decision to drop Objective C
(too old to reply)
Michael Gersten
2000-07-06 19:01:45 UTC
Permalink
This is a followup/comment on the Stepwise editorial, "Head for the Ramparts".
The brief summary: We've seen this before. Or: Fool me once, shame on you; fool
me twice, you won't fool me a third time.

The Stepwise article points this out:
Objective C is not being supplanted by java -- Java is not becoming the main
langauge, and objective-C the secondary; OBJECTIVE-C IS BEING THROWN OUT! No
more web objects development in Objective C. AND EOF IS BEING GIVEN TO THE WOF
PEOPLE, NOT THE DESKTOP PEOPLE.

That shout, again: control of EOF is being given to the WOF people; WOF wil be
java-ONLY in 5.0. That means no objective C EOF anywhere.

Yes, objective C will still be around if you use Foundation/Appkit. But no EOF
or DBKit. Anyone want to write a database app around indexing kit?

Think about this. If Apple really wanted to abandon Objective C, because of the
cost of maintaining it, it would be nice if they gave us a release with
functioning ObjC and Java, and the conversion scripts; using that release (call
it 5.0), we could take as much time as needed to get the conversion done. Some
bugs will be uncovered by this, of course; an interm/patch (5.5) would come out.
In 6.0, objective C would all be gone, people had a working version to transfer
their stuff over, everyone is happy.

Instead, in 4.5 we have a buggy java client that cannot handle inherited EO's.
We have a java development environment that is GARBAGE FOR DEBUGGING. It cannot
tell what file you are using for a class when you click on the file (if there
are duplicate names, such as a client side and a parent side class) or when a
breakpoint triggers. I get a number of
Fatal exception: java.io.IOException: Pipe broken APPLICATION EXIT

when I try to debug the java-server side of a java client app. Try using the
'print object' button -- it generate 'print * <var-name>', which jdb doesn't
like (the space between the * and the varname confuse it). I have yet to even be
able to examine something in jdb; I get messages back saying that there is no
symbol 'this' in the current context (yes, this is from jdb, not gdb.).

So, in 4.5, we only have objective C as a reliable development environment. Yet
in 5.0, we have to give it up immediately, no chance for fixes/problems to be
identified.

As annoying as this is, we've seen it before. Not from Apple, from Steve Jobs.
At NeXT, Steve said that they would always support black hardware purchasers
better than intel hardware purchases. That was true for the life of NeXTStep;
but OpenStep was different. To go from 3.2 NS to 4.2 OS required that black
hardware people purchase the same expensive CD that intel people purchased. No
price break of any kind. You couldn't use 4.2 user (cheap, bug fixes) with 3.2
developer. So for some of us (me at least, some others that I knew) there was no
upgrade to openstep.

Back in the black hardware days, my machine was academic. I also, thinking I was
smart, purchased the reinstallation floppies figuring that I 'd have a backup of
the system in case I needed it. Well, sure enough, I did. Later, when I went to
play with DBKit, I found that the reinstallation floppies did not include
sybase. NeXT would not help me out in any way. I had a system with a nice
database programming system, but no usable database; they had adaptors for flat
files, but refused to release it. That was "screw you" number one; the "no
upgrade to openstep" (which, incidently, included EOF which was DBKit's
replacement) was number two. I won't allow a third.

Jobs has dropped software before. Way back when, there was a real nice word
processor called "WriteNow". When NeXTStep went multi-platform, WriteNow was
removed and given to a commercial company. Now granted, giving software to the
community was unfashionable back then, but if that had been done, the software
would still be around, bug fixed, and working. Heck, even if the rumors were
true and it was in 68030 assembly, it would have been ported to the power PC and
fixed up, if not actually C'd.

Now, they are dropping, what? HyperCard? Stepwise reports:

<<
One hacker received a flat "no" when he asked if Apple would release HyperCard
into the open-source arena. Apple cited the time and effort it would take to
tidy everything up in HyperCard and manage the project. Apple would prefer to
allocate such resources to new projects, particularly Mac OS X.

This opened up a larger open-source debate, prompting Glass to state that, "Open
source will not be a dumping ground for dead products."
It takes all of 20 minutes to do a 'make clean; gnutar -cz' to put something
into open source, and let someone else manage it. It doesn't cost Apple anything
to give their old, useless (to them) software away, and it makes them look
better to developers (who know that their old systems will never go away
completely).

There's a bunch more in the stepwise article, but there's this thing (talking
In the end, Apple has put the question entirely to rest post WWDC.
The WebObjects 4.5 licenses have been amended such that you are
no longer allowed to run any other Cocoa applications other than
those provided by Apple for development. The Yellow Box
runtime-license that was available ($3000/10 seats) is
discontinued effective September 9, 2000.
Does this mean that I cannot develop a desktop app for YB/NT, and sell it? Even
if I tell people that they have to purchase a $700 unlimited WOF 4.5 package as
well? (Our program is, eventually, going to be a $500/month license sort of
thing, so we can afford a $700 per customer overhead; but it won't be ready by
Sep 9 :-).

And it's still bad no matter how you look at it. Consider our application, at
$500 per month, is expected to go to 1000 customers the first year (vertical
market). That's $6000 per year, 1000 customers, that $6 million the first year
alone. That's not a small software application. We need to be able to do
distributed display -- running the program and data from a central site,
displaying into a remote terminal. Java client would have been perfect, but
because of the inherited EO bug we need the 5.0 fixes. We can't use Objective C
in 5.0, and java debugging in 4.5 isn't good enough for use. So with no upgrade
path, and no workability now, we have to use something like WinFrame/Terminal
Server instead.

Folks, this is a call for action. Send mail to ***@apple.com, or maybe
try some other addresses (***@apple.com might go somewhere). Apple will
listen to its developers if we make a big enough stink. And if they don't, then
this is a call for action for GnuStep and GnuStep Web; lets port them to NT and
solaris.

It occurs to me that the Java to Objective-c Bridging Specifications was kept in
a .jobs file; now that that bridging wil be going away, presumably the .jobs
files will go away as well. Perhaps this is a sign -- what would it take for a
shareholder (I'm not one) to file a motion to replace Jobs as head of Apple with
someone else? It doesn't matter how good of a salesman is at the head if there
are no programmers writing for it.

Michael

(yes, I'm upset. And yes, I waited the hour or two before posting. Heck, it's
been almost a full week since reading the stepwise article, and I'm STILL upset
about it.)
Pierce T. Wetter III
2000-07-06 20:23:35 UTC
Permalink
Stuff about Java having major problems in 4.0 and 4.5
My point of view:

In the software industry, until you see it with your own eyes, you pretty
much assume that things will remain broken. Java has some major limitations
compared to Objective-C in WebObjects 4.0 and 4.5.

So naturally, when we hear that Apple is forcing everyone to move to Java,
that seems incredibly brain damaged. We can't really use Java now, at the
minimum we have to use a mixture of Obj-C and Java. So that means that Apple
is saying "all your web apps will no longer work". Naturally, we're all
quite upset. Those of us who have used Java in other contexts are even more
upset, Java has a ways to go.

Apple could calm people down pretty easily by listing what they saw as the
limitations of Java, and how they were fixing them in 5.0. If they did it
right, they could spin it as new improved 5.0 features.

Since they haven't done that, naturally, everyone is upset. Those of us in
the software industry pretty much expect that since Java is broken, it will
stay broken.

Specifically:

1. Why not have "Why Wait" compiling instead of Just-In-Time compiling.
SGI does it... WO apps are _server_ apps. There's not a lot of point to
bytecode in a server app.

2. EOModeler should make two classes ala EOGenerator.

3. java.lang.String sucks eggs. Rancid ones.

4. Is my app still going to go away for 20 seconds while the garbage
collector masterbates? I'd rather have 1 millisecond per page spent in
retain/release then 20 seconds all at once...

A pro Java data point:

It takes .5 seconds for FrontBase to return my 10,000 rows I asked for
and turn them into dictionarys for EOF. It takes EOF 10x that to turn them
into objects (5 seconds). I heard that with the Hotspot stuff, an object can
be allocated in 11 instructions. If 5.0 lets me get those 10,000 rows in 1
seconds total, I'd be more then happy to move to Java: 90% of the
application time is spent in EOF.

On the other hand, if EOF isn't faster in 5.0, then Apple has shot
themselves in the foot. Why move to a language that doesn't offer them any
actual benefits for server applications?

Rumor I've heard:

Apple is planning a 5.0 beta soon, so that people will see it actually
working and not be so upset.

My reaction:

They shipped a WOAdaptor that was broken. I suspect the WOF group is
getting farther and farther from their users. Even if 5.0 looks great, I'll
be sceptical until a large site like the Apple Store is running on it, and
in production.
Jobs has dropped software before. Way back when, there was a real nice word
processor called "WriteNow".
WriteNow was always a commercial product and always owned by T/Maker it
was dropped because it was written entirely in 68K assembly and T/Maker
didn't really have that much interest in porting it. Note that they didn't
port it to the powerPC on the mac...
<<
One hacker received a flat "no" when he asked if Apple would release HyperCard
into the open-source arena. Apple cited the time and effort it would take to
tidy everything up in HyperCard and manage the project. Apple would prefer to
allocate such resources to new projects, particularly Mac OS X.
This opened up a larger open-source debate, prompting Glass to state that,
"Open
source will not be a dumping ground for dead products."
It takes all of 20 minutes to do a 'make clean; gnutar -cz' to put something
into open source, and let someone else manage it. It doesn't cost Apple
anything
to give their old, useless (to them) software away, and it makes them look
better to developers (who know that their old systems will never go away
completely).
Apple may have the open source religion, but that doesn't give them the
right to publish other people's code.

If they bought code or technology from someone else and used it in
hypercard, it would take a significant amount of time to "clean" it. "make
clean" doesn't search through the project heirarchy looking for purchased
code. It takes a significant effort to move something as old as Hypercard
from closed to open code. Not worth it for dead projects. Most people into
that thing use SuperCard.

In other words, on the mac side of things, Apple was getting beat up for
dropping STREAMS from the tcp/ip stack in Mac OS X (Open Transport is
STREAMS based). However, their streams code was provided by Mentat. If Apple
"open sourced" open transport, if they didn't pull out the Mentat stuff,
they would be responsible for all financial losses to Mentat, plus they
would have undoubtedly violated various contracts.

Part of what took so long to get Darwin out was that that Apple had to
trace down the source of every single component of NextStep so that that
could isolate the parts that were proprietary to someone other then Apple.

Pierce
Eric Hermanson
2000-07-07 11:31:01 UTC
Permalink
----- Original Message -----
Post by Pierce T. Wetter III
3. java.lang.String sucks eggs. Rancid ones.
As do most of the available Java libraries. Hit java.sun.com and you'll
find literally hundreds of horrendously designed libraries, most suffering
from severe class bloat and a lack of proper design. Private methods that
should be encapsulated are exposed as public, objects that don't have any
business being separate objects, database support that resembles pre-ASP
technology, etc. But that's what happens when a language is pushed because
of hype rather than for technical reasons.

Apple did such a *nice* job of designing their Obj-C code (AppKit,
Foundation, EOF, etc) that I am increasingly dissappointed they're choosing
to associate themselves in an irreversible way with this kind of garbage. I
still do not see why they cannot continue to offer WebObjects in Obj-C after
version 5.0, if not for anything else but to keep their existing X,000
customers happy.

- Eric
Karl Hsu
2000-07-07 16:30:25 UTC
Permalink
Post by Eric Hermanson
As do most of the available Java libraries. Hit java.sun.com and you'll
find literally hundreds of horrendously designed libraries, most suffering
from severe class bloat and a lack of proper design. Private methods that
should be encapsulated are exposed as public, objects that don't have any
business being separate objects, database support that resembles pre-ASP
technology, etc. But that's what happens when a language is pushed because
of hype rather than for technical reasons.
I agree. So the pain is because the available class libraries for Java are
horrid, yes?
Post by Eric Hermanson
Apple did such a *nice* job of designing their Obj-C code (AppKit,
Foundation, EOF, etc) that I am increasingly dissappointed they're choosing
to associate themselves in an irreversible way with this kind of garbage. I
still do not see why they cannot continue to offer WebObjects in Obj-C after
version 5.0, if not for anything else but to keep their existing X,000
customers happy.
So, let's think about this, shall we? (First, AppKit is not part of WO - AK
is still in ObjC, and will be for the forseeable future. ). Did you catch
the bit at WWDC where Apple mentioned providing a goodly chunk of
Foundation in Java? And the bit where WOFrameworks is in Java? So don't use
the class libraries from Java - use Apple's JavaFoundation (unless you
believe that having gone from ObjC to Java, those wacky engineers will
suddenly be unable to do good OO design?) until Sun get's it's act together
and fixes their libraries.

If you want ObjC, use 4.5. That's Apple's way of "keeping" ObjC - the whole
point of 5.0 is pure Java - as they discovered with previous versions, a
half-assed Java implementation is worse than no Java implementation (3.5's
Java support was so poor, it actually caused some companies to _stop_ using
WO entirely - just like Monitor in 4.5). Keeping ObjC "around" in 5.0 would
imply writing a new "reverse" bridge from ObjC to Java - not an easy task.

Karl
Michael Gersten
2000-07-11 22:23:16 UTC
Permalink
Post by Karl Hsu
Post by Eric Hermanson
As do most of the available Java libraries. Hit java.sun.com and you'll
find literally hundreds of horrendously designed libraries, most suffering
from severe class bloat and a lack of proper design.
Apple did such a *nice* job of designing their Obj-C code (AppKit,
Foundation, EOF, etc) that I am increasingly dissappointed they're choosing
to associate themselves in an irreversible way with this kind of garbage. I
still do not see why they cannot continue to offer WebObjects in Obj-C after
version 5.0, if not for anything else but to keep their existing X,000
customers happy.
If you want ObjC, use 4.5. That's Apple's way of "keeping" ObjC
And when new versions have new features that are not available? (think
auto-generation of user interfaces for desktops, or a way to freeze a
java-client display to a .NIB, etc).

And when bugs are discovered and reported and not back-patched/fixed?
(think inherited EO's and java client in 4.5 right now.)
Post by Karl Hsu
- the whole
point of 5.0 is pure Java - as they discovered with previous versions, a
half-assed Java implementation is worse than no Java implementation (3.5's
Java support was so poor, it actually caused some companies to _stop_ using
WO entirely - just like Monitor in 4.5). Keeping ObjC "around" in 5.0 would
imply writing a new "reverse" bridge from ObjC to Java - not an easy task.
Pure java? This is the point?

I'm not familiar enough with the java VM to be 100% certain of what I'm about to
say. I do believe that 'java command' is outdated -- having 'command'
automatically run the java machine is more desirable (already done on linux,
probably done in the 5.0 system / Mac 10 user system).

I'm pretty sure that java permits a 'java myapp' program to use native calls,
even if applets in the sandbox cannot. And, I believe that such apps can use the
java debugger.

Pure Java? Aside from using just the native Java debugger, what does it buy you?
No more call stacks that look like java -> objective C -> java to confuse java
programmers?

That it runs on any java machine? Only those with the apple libraries. Yes, it
means that you are not restricted to NT or Mac 10 deployment boxes -- you can
now use any system. (Hmm -- maybe this is the point -- that Linux people can now
use WOF/EOF for development. Apple, this may be a key point for you to stress).

A reverse bridge? I think it's a great idea.

Other options? How about a dual version -- what if there was both a 100% java,
usable on Linux, version of EOF/WOF, as well as the objective C equivalant for
the existing mixed language implementations/applications on NT and Mac 10? No
resources/no time? No answer to that.

At the least, how about releasing a bug-fixed version of EOF/WOF 4.5/objC-Java
along with the EOF/WOF 5.0 Java? Even if they can't get the new stuff in 5.0
back-ported to objective C, at least get a working mixed language system with
good debugging support at the same time, so that we can port stuff safely if we
have to.

Remember: Proper practice is to have both the old and new running at the same
time until you are satisfied that the new works well enough to discard the old.
By saying 'You can't go forward unless you throw everything away', Apple breaks
that practice.

Michael
Michael Gersten
2000-07-11 21:29:36 UTC
Permalink
Post by Pierce T. Wetter III
Post by Michael Gersten
Jobs has dropped software before. Way back when, there was a real nice word
processor called "WriteNow".
WriteNow was always a commercial product and always owned by T/Maker it
was dropped because it was written entirely in 68K assembly and T/Maker
didn't really have that much interest in porting it. Note that they didn't
port it to the powerPC on the mac...
I was not aware that WriteNow was commercial when it was originally bundled with
NeXTStep. Complaint withdrawn.
Post by Pierce T. Wetter III
...
Apple may have the open source religion, but that doesn't give them the
right to publish other people's code.
If they bought code or technology from someone else and used it in
hypercard, it would take a significant amount of time to "clean" it.
Hmm... Ok, so (A) I'm not a lawyer, and (B) this sort of thing never occured to
me. Couldn't Apple just state something like "We don't have the resources needed
to ensure that there are no 3rd party technologies used here before we release
it"?
Geoff Hopson
2000-07-06 21:08:05 UTC
Permalink
I seem to remember Ernie Prabhakar trying to petition NeXT/Apple
back in the past, and a quick search revealed the following
article...the last paragraph caught my eye.

http://www.omnigroup.com/OldLook/MailArchive/MacOSX-talk/1997/0003.html

In the UK, we call this "poacher turned gamekeeper".....maybe Ernie
should get a petition from developers, just as he was advocating back
in 1997...or set something up so that he, now he is head of Product
Marketing for WebObjects, can listen to us WO/EOF developers.

Apologies for this not being a "dev" post, but I couldn't resist :-)
And apologies to Ernie for raking over the past like this. But
please listen and hear us!

geoff
Post by Michael Gersten
This is a followup/comment on the Stepwise editorial, "Head for
the Ramparts".
Post by Michael Gersten
The brief summary: We've seen this before. Or: Fool me once, shame
on you; fool
Post by Michael Gersten
me twice, you won't fool me a third time.
Objective C is not being supplanted by java -- Java is not
becoming the main
Post by Michael Gersten
langauge, and objective-C the secondary; OBJECTIVE-C IS BEING
THROWN OUT! No
Post by Michael Gersten
more web objects development in Objective C. AND EOF IS BEING
GIVEN TO THE WOF
Post by Michael Gersten
PEOPLE, NOT THE DESKTOP PEOPLE.
That shout, again: control of EOF is being given to the WOF
people; WOF wil be
Post by Michael Gersten
java-ONLY in 5.0. That means no objective C EOF anywhere.
Yes, objective C will still be around if you use
Foundation/Appkit. But no EOF
Post by Michael Gersten
or DBKit. Anyone want to write a database app around indexing kit?
Think about this. If Apple really wanted to abandon Objective C,
because of the
Post by Michael Gersten
cost of maintaining it, it would be nice if they gave us a release with
functioning ObjC and Java, and the conversion scripts; using that
release (call
Post by Michael Gersten
it 5.0), we could take as much time as needed to get the
conversion done. Some
Post by Michael Gersten
bugs will be uncovered by this, of course; an interm/patch (5.5)
would come out.
Post by Michael Gersten
In 6.0, objective C would all be gone, people had a working
version to transfer
Post by Michael Gersten
their stuff over, everyone is happy.
Instead, in 4.5 we have a buggy java client that cannot handle
inherited EO's.
Post by Michael Gersten
We have a java development environment that is GARBAGE FOR
DEBUGGING. It cannot
Post by Michael Gersten
tell what file you are using for a class when you click on the
file (if there
Post by Michael Gersten
are duplicate names, such as a client side and a parent side
class) or when a
Post by Michael Gersten
breakpoint triggers. I get a number of
Fatal exception: java.io.IOException: Pipe broken
APPLICATION EXIT
Post by Michael Gersten
when I try to debug the java-server side of a java client app. Try
using the
Post by Michael Gersten
'print object' button -- it generate 'print * <var-name>', which
jdb doesn't
Post by Michael Gersten
like (the space between the * and the varname confuse it). I have
yet to even be
Post by Michael Gersten
able to examine something in jdb; I get messages back saying that
there is no
Post by Michael Gersten
symbol 'this' in the current context (yes, this is from jdb, not gdb.).
So, in 4.5, we only have objective C as a reliable development
environment. Yet
Post by Michael Gersten
in 5.0, we have to give it up immediately, no chance for
fixes/problems to be
Post by Michael Gersten
identified.
As annoying as this is, we've seen it before. Not from Apple, from
Steve Jobs.
Post by Michael Gersten
At NeXT, Steve said that they would always support black hardware
purchasers
Post by Michael Gersten
better than intel hardware purchases. That was true for the life
of NeXTStep;
Post by Michael Gersten
but OpenStep was different. To go from 3.2 NS to 4.2 OS required
that black
Post by Michael Gersten
hardware people purchase the same expensive CD that intel people
purchased. No
Post by Michael Gersten
price break of any kind. You couldn't use 4.2 user (cheap, bug
fixes) with 3.2
Post by Michael Gersten
developer. So for some of us (me at least, some others that I
knew) there was no
Post by Michael Gersten
upgrade to openstep.
Back in the black hardware days, my machine was academic. I also,
thinking I was
Post by Michael Gersten
smart, purchased the reinstallation floppies figuring that I 'd
have a backup of
Post by Michael Gersten
the system in case I needed it. Well, sure enough, I did. Later,
when I went to
Post by Michael Gersten
play with DBKit, I found that the reinstallation floppies did not
include
Post by Michael Gersten
sybase. NeXT would not help me out in any way. I had a system with
a nice
Post by Michael Gersten
database programming system, but no usable database; they had
adaptors for flat
Post by Michael Gersten
files, but refused to release it. That was "screw you" number one;
the "no
Post by Michael Gersten
upgrade to openstep" (which, incidently, included EOF which was DBKit's
replacement) was number two. I won't allow a third.
Jobs has dropped software before. Way back when, there was a real
nice word
Post by Michael Gersten
processor called "WriteNow". When NeXTStep went multi-platform,
WriteNow was
Post by Michael Gersten
removed and given to a commercial company. Now granted, giving
software to the
Post by Michael Gersten
community was unfashionable back then, but if that had been done,
the software
Post by Michael Gersten
would still be around, bug fixed, and working. Heck, even if the
rumors were
Post by Michael Gersten
true and it was in 68030 assembly, it would have been ported to
the power PC and
Post by Michael Gersten
fixed up, if not actually C'd.
<<
One hacker received a flat "no" when he asked if Apple would
release HyperCard
Post by Michael Gersten
into the open-source arena. Apple cited the time and effort it
would take to
Post by Michael Gersten
tidy everything up in HyperCard and manage the project. Apple
would prefer to
Post by Michael Gersten
allocate such resources to new projects, particularly Mac OS X.
This opened up a larger open-source debate, prompting Glass to
state that, "Open
Post by Michael Gersten
source will not be a dumping ground for dead products."
It takes all of 20 minutes to do a 'make clean; gnutar -cz' to put
something
Post by Michael Gersten
into open source, and let someone else manage it. It doesn't cost
Apple anything
Post by Michael Gersten
to give their old, useless (to them) software away, and it makes
them look
Post by Michael Gersten
better to developers (who know that their old systems will never
go away
Post by Michael Gersten
completely).
There's a bunch more in the stepwise article, but there's this
thing (talking
Post by Michael Gersten
In the end, Apple has put the question entirely to rest post WWDC.
The WebObjects 4.5 licenses have been amended such that you are
no longer allowed to run any other Cocoa applications other than
those provided by Apple for development. The Yellow Box
runtime-license that was available ($3000/10 seats) is
discontinued effective September 9, 2000.
Does this mean that I cannot develop a desktop app for YB/NT, and
sell it? Even
Post by Michael Gersten
if I tell people that they have to purchase a $700 unlimited WOF
4.5 package as
Post by Michael Gersten
well? (Our program is, eventually, going to be a $500/month
license sort of
Post by Michael Gersten
thing, so we can afford a $700 per customer overhead; but it won't
be ready by
Post by Michael Gersten
Sep 9 :-).
And it's still bad no matter how you look at it. Consider our
application, at
Post by Michael Gersten
$500 per month, is expected to go to 1000 customers the first year
(vertical
Post by Michael Gersten
market). That's $6000 per year, 1000 customers, that $6 million
the first year
Post by Michael Gersten
alone. That's not a small software application. We need to be able
to do
Post by Michael Gersten
distributed display -- running the program and data from a central
site,
Post by Michael Gersten
displaying into a remote terminal. Java client would have been
perfect, but
Post by Michael Gersten
because of the inherited EO bug we need the 5.0 fixes. We can't
use Objective C
Post by Michael Gersten
in 5.0, and java debugging in 4.5 isn't good enough for use. So
with no upgrade
Post by Michael Gersten
path, and no workability now, we have to use something like
WinFrame/Terminal
Post by Michael Gersten
Server instead.
Folks, this is a call for action. Send mail to
somewhere). Apple
Post by Michael Gersten
will
listen to its developers if we make a big enough stink. And if
they don't, then
Post by Michael Gersten
this is a call for action for GnuStep and GnuStep Web; lets port
them to NT and
Post by Michael Gersten
solaris.
It occurs to me that the Java to Objective-c Bridging
Specifications was kept in
Post by Michael Gersten
a .jobs file; now that that bridging wil be going away, presumably
the .jobs
Post by Michael Gersten
files will go away as well. Perhaps this is a sign -- what would
it take for a
Post by Michael Gersten
shareholder (I'm not one) to file a motion to replace Jobs as head
of Apple with
Post by Michael Gersten
someone else? It doesn't matter how good of a salesman is at the
head if there
Post by Michael Gersten
are no programmers writing for it.
Michael
(yes, I'm upset. And yes, I waited the hour or two before posting.
Heck, it's
Post by Michael Gersten
been almost a full week since reading the stepwise article, and
I'm STILL upset
Post by Michael Gersten
about it.)
_______________________________________________
EOF mailing list
http://www.omnigroup.com/mailman/listinfo/eof
Scott Anguish
2000-07-07 04:06:40 UTC
Permalink
Let me clarify this... Stepwise didn't write this. It was linked
from MacWeek or MacCentral from Stepwise.
Post by Michael Gersten
<<
One hacker received a flat "no" when he asked if Apple would release HyperCard
into the open-source arena. Apple cited the time and effort it would take to
tidy everything up in HyperCard and manage the project. Apple would prefer to
allocate such resources to new projects, particularly Mac OS X.
This opened up a larger open-source debate, prompting Glass to state that, "Open
source will not be a dumping ground for dead products."
It takes all of 20 minutes to do a 'make clean; gnutar -cz' to put something
into open source, and let someone else manage it. It doesn't cost Apple anything
to give their old, useless (to them) software away, and it makes them look
better to developers (who know that their old systems will never go away
completely).
Completely false. Apple has to do due dilligence on the software.
They must ensure that they have no contracts that prohibit the distribution
of the source. They have to ensure that there is nothng of a confidential
nature that shouldn't be released.

It is not as simple as just doing a make clean; gnutar -cz.

Darwin has much good potential.. but that apple is not inclined to
use it as a dumping ground is a very good thing.
pirmin
2000-07-07 10:49:35 UTC
Permalink
Post by Michael Gersten
Yes, objective C will still be around if you use Foundation/Appkit. But no EOF
or DBKit. Anyone want to write a database app around indexing kit?
I've written a sort of lean EOF in ObjC that only uses the adaptor
level. Should be easy to replace that also and maybe integrate mySQL
client directly.

What's much worse is that Apple refuses to provide YB for Windows
deployment licenses. As EAP and now ADC member I've always been told
about multi-platform advantages. So I developed desktop apps that I
can't sell now. Apple has been actively misleading me, causing financial
loss. Currently I'm investigating possiblities to correct this
situation. If you are in the same position and want to participate, drop
me a mail. If there is already something going on, I'd be happy to join.
Helge Hess
2000-07-07 10:41:29 UTC
Permalink
Post by pirmin
Post by Michael Gersten
Yes, objective C will still be around if you use Foundation/Appkit. But no EOF
or DBKit. Anyone want to write a database app around indexing kit?
I've written a sort of lean EOF in ObjC that only uses the adaptor
level. Should be easy to replace that also and maybe integrate mySQL
client directly.
The GNUstep Database Library (GDL) is a more or less complete EOF 1.x
replacement, that is, it provides the adaptor level but also the
database level. Adaptors are available for quite some databases, eg
Oracle, Sybase, Postgres and FrontBase.
It also works on Windows.

Greetings
Helge
Michael Gersten
2000-07-11 21:43:24 UTC
Permalink
Ok, exactly what is the restriction on this? Our 4.5 license seems to permit
doing YB for windows deployment if we buy one of the $700 full licenses (as it
comes with both development, which permits using the YB libraries and tools, and
the deployment, which lets us run our apps [but if I read it correctly, not the
libraries themselves without another license]) per end-use machine.
Post by pirmin
What's much worse is that Apple refuses to provide YB for Windows
deployment licenses. As EAP and now ADC member I've always been told
about multi-platform advantages. So I developed desktop apps that I
can't sell now. Apple has been actively misleading me, causing financial
loss. Currently I'm investigating possiblities to correct this
situation. If you are in the same position and want to participate, drop
me a mail. If there is already something going on, I'd be happy to join.
That makes two people that have contacted me about this topic.
Philippe C.D. Robert
2000-07-07 09:56:52 UTC
Permalink
Hi,

this is getting endless, I suppose....;-)
Post by Michael Gersten
Objective C is not being supplanted by java -- Java is not becoming the main
langauge, and objective-C the secondary; OBJECTIVE-C IS BEING THROWN OUT! No
more web objects development in Objective C. AND EOF IS BEING GIVEN TO THE WOF
PEOPLE, NOT THE DESKTOP PEOPLE.
Why shout?

While I am a true ObjC 'lover', I understand that the world is moving towards
Java, that you do *not* find a lot of programmers who are *willing* to learn
ObjC anymore, that the existing solution base is in Java, that .... and so on.
We just have to accept that WOF has to be Java, even if this sucks, even if
Java is a move backwards from one or another point of view (while there are
also technical reasons that speak for Java...), but on the 'web' you have to
talk Java thses days period.

The other thing with EOF is IMHO indeed a problem. EOF was design for c/s
applications, not for app server stuff. Therefore it would make sense to write
a new EOF addressing the new situation, call it differenly and let the EOF be
EOF, that is let us use it on the desktop - in ObjC!

BTW if you want to use sth like EOF 2.x/3.x in ObjC, come use GNUstep, gdl2 is
in the work...in ObjC...;-)

sweet dreams, Phil - going back to ObjC hacking...grin
--
Philippe C.D. Robert
Developer - StudioSendai.com
Eric Hermanson
2000-07-07 11:20:47 UTC
Permalink
----- Original Message -----
Post by Philippe C.D. Robert
While I am a true ObjC 'lover', I understand that the world is moving towards
Java, that you do *not* find a lot of programmers who are *willing* to learn
ObjC anymore, that the existing solution base is in Java, that .... and so on.
We just have to accept that WOF has to be Java, even if this sucks, even if
Java is a move backwards from one or another point of view
This statement, and other like it, are _baloney_. People are not "willing"
to learn Obj-C because the language, it's runtime, and Apple's underlying
support frameworks (Foundation, AppKit, etc), are not available for *free*
download from Apple's website like Java is from Sun's website. That's THE
reason, and the ONLY reason why Obj-C has not achieved the marketing hype
that Java has (note the careful use of the term "marketing hype" -- because
that's the only kind of hype that Java currently provides).

In my opinion, everyone is being overly optimistic about Apple's recent
price drop on WebObjects. The fact that they advertise on their website
that WebObjects used to cost $50,000 in the first place is embarrasing. It
shows how little they know about marketing software in post-1995. I don't
know how they expect to gain any more marketshare by lowering the price by a
factor of 70? A price of $10 would still be too much for the runtime and
support frameworks because their entire business model for WebObjects is
inherently incorrect. They should be giving the language and it's support
frameworks away for free, and charging for their development tools
(ProjectBuilder, WOBuilder, etc) just like the rest of the industry. The
fact that Apple curls up like a beaten victim and "acknowledges" that Java
has "won" makes me sick. They are falling into the Java "trap". Once
WebObjects is Java-only, it will no longer be distinguishable from the 200
other Java web application servers. But hey, if they're choosing to lump
themselves in with that group of crappy software, then so be it!

- Eric
Philippe C.D. Robert
2000-07-07 22:11:47 UTC
Permalink
On Fri, 07 Jul 2000, Eric Hermanson wrote:
<sni[ my message>
Post by Eric Hermanson
Post by Philippe C.D. Robert
We just have to accept that WOF has to be Java, even if this sucks, even
if Java is a move backwards from one or another point of view
This statement, and other like it, are _baloney_. People are not "willing"
to learn Obj-C because the language, it's runtime, and Apple's underlying
support frameworks (Foundation, AppKit, etc), are not available for *free*
download from Apple's website like Java is from Sun's website. That's THE
reason, and the ONLY reason why Obj-C has not achieved the marketing hype
that Java has (note the careful use of the term "marketing hype" -- because
that's the only kind of hype that Java currently provides).
Well, gcc is free, the ObjC runtime is free, the language itself is 'free',
there is libFoundation and GNUstep which are free .... THis is IMHO not the
problem that ObjC is not widely adopted. And be honest, you cannot change the
world, you cannot turn off the Java hype and switch the 'ObjC hype button'. We
just can hope that Apple does not forget about its real jewels...

sweet dreams, Phil
--
Philippe C.D. Robert
Developer - StudioSendai.com
Michael Gersten
2000-07-11 21:56:09 UTC
Permalink
Post by Philippe C.D. Robert
Hi,
this is getting endless, I suppose....;-)
Nope, just another round (hey, I'm back from a weekend off :-)
Post by Philippe C.D. Robert
Post by Michael Gersten
Objective C is not being supplanted by java -- Java is not becoming the main
langauge, and objective-C the secondary; OBJECTIVE-C IS BEING THROWN OUT! No
more web objects development in Objective C. AND EOF IS BEING GIVEN TO THE WOF
PEOPLE, NOT THE DESKTOP PEOPLE.
Why shout?
To get people's attention. I understand that shouts are ignored on the web, so
my next line (non-shout) translated that.
Post by Philippe C.D. Robert
While I am a true ObjC 'lover', I understand that the world is moving towards
Java, that you do *not* find a lot of programmers who are *willing* to learn
ObjC anymore, that the existing solution base is in Java, that .... and so on.
We just have to accept that WOF has to be Java
Here I disagree with you.

WOF needs to be usable from Java, yes.
WOF needs to be non-usable from ObjC, No.

WOF needs to be implemented in Java, no.
WOF needs to be implemented in ObjC, no.
WOF can't be implemented in any other language, No.

I don't care what language EOF/WOF is implemented in. Use a Java-native EOF if
you must. Just give us an objective-C brige (a .ojbs file?) as well.
Post by Philippe C.D. Robert
The other thing with EOF is IMHO indeed a problem. EOF was design for c/s
applications, not for app server stuff.
OK, what do you see as the difference between client/server and app server? And,
why do you say that EOF was designed for client/server?

EOF, as I see it, is designed for database client/database server; the app
itself can be desktop, client/server, web page, java client, command line, or
anything.
Post by Philippe C.D. Robert
Therefore it would make sense to write
a new EOF addressing the new situation, call it differenly and let the EOF be
EOF, that is let us use it on the desktop - in ObjC!
BTW if you want to use sth like EOF 2.x/3.x in ObjC, come use GNUstep, gdl2 is
in the work...in ObjC...;-)
I did take a look at GnuStep.
Randy Leonard
2000-07-07 19:31:09 UTC
Permalink
Michael brings up (at least) two distinct points in his original post. The
first is the dropping of Objective-C in WOF 5.0 and the implications that
has on EOF for Cocoa. The second is the dropping of runtime licenses for
WOF 4.5 on NT.
Post by Michael Gersten
Does this mean that I cannot develop a desktop app for YB/NT, and sell it?
Even
Post by Michael Gersten
if I tell people that they have to purchase a $700 unlimited WOF 4.5
package as
Post by Michael Gersten
well?
Is it really the case that developers cannot sell products to those already
using WOF 4.5 on NT after September 2000? Seems bizarre that Apple would
finish WOF/NT 4.5 only to rescind anyone's rights to sell products on that
platform.

Randy Leonard
Ocean Software, Inc.
www.oceansoft.com



-----Original Message-----
From: Michael Gersten [mailto:***@getasia.com]
Sent: Thursday, July 06, 2000 3:02 PM
To: EOF list; webobjects list
Cc: ***@apple.com
Subject: Regarding Apple's decision to drop Objective C


This is a followup/comment on the Stepwise editorial, "Head for the
Ramparts".
The brief summary: We've seen this before. Or: Fool me once, shame on you;
fool
me twice, you won't fool me a third time.

The Stepwise article points this out:
Objective C is not being supplanted by java -- Java is not becoming the main
langauge, and objective-C the secondary; OBJECTIVE-C IS BEING THROWN OUT! No
more web objects development in Objective C. AND EOF IS BEING GIVEN TO THE
WOF
PEOPLE, NOT THE DESKTOP PEOPLE.

That shout, again: control of EOF is being given to the WOF people; WOF wil
be
java-ONLY in 5.0. That means no objective C EOF anywhere.

Yes, objective C will still be around if you use Foundation/Appkit. But no
EOF
or DBKit. Anyone want to write a database app around indexing kit?

Think about this. If Apple really wanted to abandon Objective C, because of
the
cost of maintaining it, it would be nice if they gave us a release with
functioning ObjC and Java, and the conversion scripts; using that release
(call
it 5.0), we could take as much time as needed to get the conversion done.
Some
bugs will be uncovered by this, of course; an interm/patch (5.5) would come
out.
In 6.0, objective C would all be gone, people had a working version to
transfer
their stuff over, everyone is happy.

Instead, in 4.5 we have a buggy java client that cannot handle inherited
EO's.
We have a java development environment that is GARBAGE FOR DEBUGGING. It
cannot
tell what file you are using for a class when you click on the file (if
there
are duplicate names, such as a client side and a parent side class) or when
a
breakpoint triggers. I get a number of
Fatal exception: java.io.IOException: Pipe broken APPLICATION
EXIT

when I try to debug the java-server side of a java client app. Try using the
'print object' button -- it generate 'print * <var-name>', which jdb doesn't
like (the space between the * and the varname confuse it). I have yet to
even be
able to examine something in jdb; I get messages back saying that there is
no
symbol 'this' in the current context (yes, this is from jdb, not gdb.).

So, in 4.5, we only have objective C as a reliable development environment.
Yet
in 5.0, we have to give it up immediately, no chance for fixes/problems to
be
identified.

As annoying as this is, we've seen it before. Not from Apple, from Steve
Jobs.
At NeXT, Steve said that they would always support black hardware purchasers
better than intel hardware purchases. That was true for the life of
NeXTStep;
but OpenStep was different. To go from 3.2 NS to 4.2 OS required that black
hardware people purchase the same expensive CD that intel people purchased.
No
price break of any kind. You couldn't use 4.2 user (cheap, bug fixes) with
3.2
developer. So for some of us (me at least, some others that I knew) there
was no
upgrade to openstep.

Back in the black hardware days, my machine was academic. I also, thinking I
was
smart, purchased the reinstallation floppies figuring that I 'd have a
backup of
the system in case I needed it. Well, sure enough, I did. Later, when I went
to
play with DBKit, I found that the reinstallation floppies did not include
sybase. NeXT would not help me out in any way. I had a system with a nice
database programming system, but no usable database; they had adaptors for
flat
files, but refused to release it. That was "screw you" number one; the "no
upgrade to openstep" (which, incidently, included EOF which was DBKit's
replacement) was number two. I won't allow a third.

Jobs has dropped software before. Way back when, there was a real nice word
processor called "WriteNow". When NeXTStep went multi-platform, WriteNow was
removed and given to a commercial company. Now granted, giving software to
the
community was unfashionable back then, but if that had been done, the
software
would still be around, bug fixed, and working. Heck, even if the rumors were
true and it was in 68030 assembly, it would have been ported to the power PC
and
fixed up, if not actually C'd.

Now, they are dropping, what? HyperCard? Stepwise reports:

<<
One hacker received a flat "no" when he asked if Apple would release
HyperCard
into the open-source arena. Apple cited the time and effort it would take to
tidy everything up in HyperCard and manage the project. Apple would prefer
to
allocate such resources to new projects, particularly Mac OS X.

This opened up a larger open-source debate, prompting Glass to state that,
"Open
source will not be a dumping ground for dead products."
It takes all of 20 minutes to do a 'make clean; gnutar -cz' to put something
into open source, and let someone else manage it. It doesn't cost Apple
anything
to give their old, useless (to them) software away, and it makes them look
better to developers (who know that their old systems will never go away
completely).

There's a bunch more in the stepwise article, but there's this thing
(talking
Post by Michael Gersten
In the end, Apple has put the question entirely to rest post WWDC.
The WebObjects 4.5 licenses have been amended such that you are
no longer allowed to run any other Cocoa applications other than
those provided by Apple for development. The Yellow Box
runtime-license that was available ($3000/10 seats) is
discontinued effective September 9, 2000.
Does this mean that I cannot develop a desktop app for YB/NT, and sell it?
Even
if I tell people that they have to purchase a $700 unlimited WOF 4.5 package
as
well? (Our program is, eventually, going to be a $500/month license sort of
thing, so we can afford a $700 per customer overhead; but it won't be ready
by
Sep 9 :-).

And it's still bad no matter how you look at it. Consider our application,
at
$500 per month, is expected to go to 1000 customers the first year (vertical
market). That's $6000 per year, 1000 customers, that $6 million the first
year
alone. That's not a small software application. We need to be able to do
distributed display -- running the program and data from a central site,
displaying into a remote terminal. Java client would have been perfect, but
because of the inherited EO bug we need the 5.0 fixes. We can't use
Objective C
in 5.0, and java debugging in 4.5 isn't good enough for use. So with no
upgrade
path, and no workability now, we have to use something like
WinFrame/Terminal
Server instead.

Folks, this is a call for action. Send mail to ***@apple.com, or
maybe
try some other addresses (***@apple.com might go somewhere). Apple
will
listen to its developers if we make a big enough stink. And if they don't,
then
this is a call for action for GnuStep and GnuStep Web; lets port them to NT
and
solaris.

It occurs to me that the Java to Objective-c Bridging Specifications was
kept in
a .jobs file; now that that bridging wil be going away, presumably the .jobs
files will go away as well. Perhaps this is a sign -- what would it take for
a
shareholder (I'm not one) to file a motion to replace Jobs as head of Apple
with
someone else? It doesn't matter how good of a salesman is at the head if
there
are no programmers writing for it.

Michael

(yes, I'm upset. And yes, I waited the hour or two before posting. Heck,
it's
been almost a full week since reading the stepwise article, and I'm STILL
upset
about it.)
Craig Brozefsky
2000-07-07 20:23:01 UTC
Permalink
Post by Randy Leonard
Is it really the case that developers cannot sell products to those already
using WOF 4.5 on NT after September 2000? Seems bizarre that Apple would
finish WOF/NT 4.5 only to rescind anyone's rights to sell products on that
platform.
Ever try to sell a client a software solution based on a architecture
that is has no forseeable future in terms of bugfixes, upgrades and
feature additions from it's vendor?
--
Craig Brozefsky <***@red-bean.com>
Lisp Web Dev List http://www.red-bean.com/lispweb
--- The only good lisper is a coding lisper. ---
Randy Leonard
2000-07-07 21:51:25 UTC
Permalink
Ever try to sell a client a software solution based...
I understand your statement, but it misses the point. These developers need
time to migrate their solution to WebObjects Java Client, or some other
platform.

This thread started with an angry developer who has presumably spent several
years developing a solution and is effectively being told he can't sell it.
The reason is not because the platform doesn't work, it's just that Apple
decided not to sell it. I'm sure he would rather be told that he can sell
his product, but not to expect bug fixes or enhancements. In other words,
let him generate reveue off his years of hard work as he migrates to Java
Client or some other platform.

I did revisit Apple's policy at www.apple.com/enterprise/letter.html. It
seems to indicate you can sell AppKit/NT-based products to those customers
who already have WOF 4.5, it's just that Apple won't sell the deployment
license as a stand-alone product. If true, those selling AppKit/NT-based
products could get by selling a $699 WebObjects license with their product
instead of a $300 deployment license. But this seems kind of silly as it
will only burden deployment machines with lots of unnecessary software.

It seems more reasonable for Apple to continue selling the deployment
license for the forseeable future, thus giving developers time to migrate.

Randy Leonard
***@oceansoft.com
www.oceansoft.com



-----Original Message-----
From: Craig Brozefsky [mailto:***@red-bean.com]
Sent: Friday, July 07, 2000 4:23 PM
To: Randy Leonard
Cc: EOF list; webobjects list; ***@apple.com
Subject: Re: Regarding Apple's decision to drop Objective C
Is it really the case that developers cannot sell products to those
already
using WOF 4.5 on NT after September 2000? Seems bizarre that Apple would
finish WOF/NT 4.5 only to rescind anyone's rights to sell products on that
platform.
Ever try to sell a client a software solution based on a architecture
that is has no forseeable future in terms of bugfixes, upgrades and
feature additions from it's vendor?
--
Craig Brozefsky <***@red-bean.com>
Lisp Web Dev List http://www.red-bean.com/lispweb
--- The only good lisper is a coding lisper. ---
sanguish
2000-07-08 01:11:10 UTC
Permalink
Post by Randy Leonard
I did revisit Apple's policy at www.apple.com/enterprise/letter.html. It
seems to indicate you can sell AppKit/NT-based products to those customers
who already have WOF 4.5, it's just that Apple won't sell the deployment
license as a stand-alone product. If true, those selling AppKit/NT-based
products could get by selling a $699 WebObjects license with their product
instead of a $300 deployment license. But this seems kind of silly as it
will only burden deployment machines with lots of unnecessary software.
All the licenses have been modifed to disallow any non Apple Yellow Box
applications to run on WebObjects NT.
Randy Leonard
2000-07-08 01:34:48 UTC
Permalink
I'm understanding the anger expressed in this thread's original note a
little better now.

Guess this means ReportMill won't be available on NT after September? And
my product to compete with ReportMill starting next week can only be sold on
MacOS X Server after September? Lovely.

Thanks,
Randy Leonard
***@oceansoft.com
www.oceansoft.com


-----Original Message-----
From: sanguish [mailto:***@digifix.com]
Sent: Friday, July 07, 2000 9:11 PM
To: Randy Leonard; 'Craig Brozefsky'
Cc: EOF list; webobjects list; ***@apple.com
Subject: Re: Regarding Apple's decision to drop Objective C
Post by Randy Leonard
I did revisit Apple's policy at www.apple.com/enterprise/letter.html. It
seems to indicate you can sell AppKit/NT-based products to those customers
who already have WOF 4.5, it's just that Apple won't sell the deployment
license as a stand-alone product. If true, those selling AppKit/NT-based
products could get by selling a $699 WebObjects license with their product
instead of a $300 deployment license. But this seems kind of silly as it
will only burden deployment machines with lots of unnecessary software.
All the licenses have been modifed to disallow any non Apple Yellow Box
applications to run on WebObjects NT.
PhilipM=?iso-8859-1?q?=F6tteli
2000-07-08 03:31:45 UTC
Permalink
Post by Philippe C.D. Robert
this is getting endless, I suppose....
Post by Michael Gersten
Objective C is not being supplanted by java -- Java is not
becoming the main
Post by Philippe C.D. Robert
Post by Michael Gersten
langauge, and objective-C the secondary; OBJECTIVE-C IS BEING
THROWN OUT! No
Post by Philippe C.D. Robert
Post by Michael Gersten
more web objects development in Objective C. AND EOF IS BEING
GIVEN TO THE WOF
Post by Philippe C.D. Robert
Post by Michael Gersten
PEOPLE, NOT THE DESKTOP PEOPLE.
Why shout?
While I am a true ObjC 'lover', I understand that the world is
moving towards
Post by Philippe C.D. Robert
Java, that you do *not* find a lot of programmers who are
*willing* to learn
Post by Philippe C.D. Robert
ObjC anymore, that the existing solution base is in Java, that
.... and so on.
Post by Philippe C.D. Robert
We just have to accept that WOF has to be Java, even if this
sucks, even if
Post by Philippe C.D. Robert
Java is a move backwards from one or another point of view (while
there are
Post by Philippe C.D. Robert
also technical reasons that speak for Java...), but on the 'web'
you have to
Post by Philippe C.D. Robert
talk Java thses days period.
And tomorrow we gonna go C#...
If it would be that easy...


Phil


--
lic. oec. publ. Philip Mötteli Tel# +41(22)320-3088
c/o Leila Bourahla Fax# +1(425)940-1642
23, Rue Prévost-Martin Tel# +41(78)742-9334
CH-1205 Geneva E-Mail: ***@yifan.net
PhilipM=?iso-8859-1?q?=F6tteli
2000-07-08 03:43:19 UTC
Permalink
But the missing link is that these technologies aren't being
pushed by a
multi-billion dollar company. Do you really think that Java
would have
gained any popularity whatsoever had Sun simply uploaded it to a
random
'ftp' server and expected the masses to download it?
So you are saying that Sun duped everyone with 'hype' into using
Java when in
fact Java has no merit? Perhaps you are not giving people enough
credit to
use their brains and choose a programming language. There are
benefits to
using Java beyond technical (personnel, training, availability of
third party
libraries [don't start with the 'quality' issue on this one, please]).
Critical mass of acceptance is one of the very real advantages,
you know.

Are you dreaming?????????
Or are you really thinking, that Hitler was right?????????
Or is M$ as good as the number of licenses sold?
Lenin (if you know who that is) said once: The people is stupid. I
have a minor in Socio-Psychology with a focus on social influence and
I can tell you, what you said is as wrong as 2+1=21.
Or, to ask you differently: Why do we have e. g. marketing departments?


Phil


--
lic. oec. publ. Philip Mötteli Tel# +41(22)320-3088
c/o Leila Bourahla Fax# +1(425)940-1642
23, Rue Prévost-Martin Tel# +41(78)742-9334
CH-1205 Geneva E-Mail: ***@yifan.net
Philippe C.D. Robert
2000-07-08 11:37:33 UTC
Permalink
Hi,
Post by PhilipM=?iso-8859-1?q?=F6tteli
So you are saying that Sun duped everyone with 'hype' into using
Java when in
fact Java has no merit? Perhaps you are not giving people enough
credit to
use their brains and choose a programming language. There are
benefits to
using Java beyond technical (personnel, training, availability of
third party
libraries [don't start with the 'quality' issue on this one, please]).
Critical mass of acceptance is one of the very real advantages,
you know.
Are you dreaming?????????
Or are you really thinking, that Hitler was right?????????
Or is M$ as good as the number of licenses sold?
Lenin (if you know who that is) said once: The people is stupid. I
have a minor in Socio-Psychology with a focus on social influence and
I can tell you, what you said is as wrong as 2+1=21.
Or, to ask you differently: Why do we have e. g. marketing departments?
Please, no political comparisons....

In computer business, the installed base *is* actually an advantage to make
money, to establish standards and so on... While ObjC and other languages are
great (TOM, SmallTalk, ...), they found their niche market but did not get
mainstream. Today it's C++, VB and Java that rule the world (in terms of
quantity). Nobody can change this (anymore), if you personally don't like them,
you still are free to choose - even ObjC.

So, instead of wasting (not meant to be flaming here) time on discussing this
topic, help us make ObjC, GNUstep, gdl(2) and GNUstep Web to be a valid
alternative!

sweet dreams, Phil
--
Philippe C.D. Robert
Developer - StudioSendai.com
PhilipM=?iso-8859-1?q?=F6tteli
2000-07-08 03:32:20 UTC
Permalink
From: PhilipMötteli<***@moetteli.ch>
Date: 2000-07-08 05:02:47 +0200
To: Eric Hermanson <***@gofish.com>
Subject: Re: Regarding Apple's decision to drop Objective C
X-Mailer-Extensions: SWSignature 1.3.2
Post by Eric Hermanson
Post by Pierce T. Wetter III
3. java.lang.String sucks eggs. Rancid ones.
As do most of the available Java libraries. Hit java.sun.com and
you'll
Post by Eric Hermanson
find literally hundreds of horrendously designed libraries, most
suffering
Post by Eric Hermanson
from severe class bloat and a lack of proper design. Private
methods that
Post by Eric Hermanson
should be encapsulated are exposed as public, objects that don't
have any
Post by Eric Hermanson
business being separate objects,
How do you want to have a good OO design, with an only halfway
object-oriented language? It's simply technically not possible. The
worst is, they even can't learn it. Or even worser, those designers
can't even realise it.Those people do not know the impact of a
completely object-oriented language on the design of their software,
although they try to tinker around by using Gamma Patterns, who were
written actually expecially for that.
Post by Eric Hermanson
database support that resembles pre-ASP
technology, etc. But that's what happens when a language is
pushed because
Post by Eric Hermanson
of hype rather than for technical reasons.
Apple did such a *nice* job of designing their Obj-C code (AppKit,
Foundation, EOF, etc) that I am increasingly dissappointed they're
choosing
Post by Eric Hermanson
to associate themselves in an irreversible way with this kind of
garbage. I
Post by Eric Hermanson
still do not see why they cannot continue to offer WebObjects in
Obj-C after
Post by Eric Hermanson
version 5.0, if not for anything else but to keep their existing X,000
customers happy.
In my eyes Apple is recovering from their period of "splendid
isolation", when they thought, that they just could do without any
concern about what the market wants. Now their slinging into the
other extreme, by licking or better creeping into the ass of their
biggest customers.
But those customers are just letting Apple do their job. The moment
they have an alternative, they will tell Apple, that they're not
sufficently present in the commercial market and they will just drop
Apple like that. And in about a year, we gonna see some alternatives,
or does anybody think, that Apple is the only company that can
program Java? Yes, they have just "Yet Another Java" product.
The art of the game is not to throw away your technical advantages,
but to open a door for the mainstream. NeXT did that, by making
NeXtstep on Intel, by making OpenStep on Windows, by offering Java
for WebObjects and by making adaptors for legacy DBMSs. But never
throw away your superiority. Because afterwards, the only thing that
distinguishes you from the others is price and marketing. And in both
Apple is, modestly said, horrible (I pay three times as much for the
cheapest G4 compared to an equivalent Intel and Apple still has the
reputation of a toy company in between all the people I know).

Me too, I waited several weeks and I'm still very, very angry. Since
early '92 I keep campaigning for that platform and now Apple is
destroying it just like that. And I would like to tell Apple, that
their not alone any more: Gnustep(Web) is quite mature and Linux is a
much more promising market anyway. I just told my clients, who both
have a multi million Dollar project running, that we gonna slowly
move to *BSD. Fortunately Linux runs on G4, not like MOSXS! Not
everybody "Steves".

BTW, I once heard, that Steve Jobs didn't really want NeXTstep to be
OO. It was one of his engineers, who imposed that. Now, old NeXTers,
which of all the products (NeXTstation, NeXTcube, printers, VLSI
chips, Mach, BSD 4.3, OpenStep with all its kits) did survive? Why
did NeXT in the end even make profit? Or, the best question: why
(technically) is Steve Jobs again CEO at Apple? Because NeXT offered
the best programming environment, by having a set of kits and a OS
layer, which were completely OO. And reel OO, thanks to Objective-C.
And what is Steve doing actually? Reducing the OO'ness of his
products!


Thanks for being patient and tolerating my bad mood.
Phil


--
lic. oec. publ. Philip Mötteli Tel# +41(22)320-3088
c/o Leila Bourahla Fax# +1(425)940-1642
23, Rue Prévost-Martin Tel# +41(78)742-9334
CH-1205 Geneva E-Mail: ***@yifan.net
Scott Anguish
2000-07-08 04:32:39 UTC
Permalink
Post by Randy Leonard
I'm understanding the anger expressed in this thread's original note a
little better now.
I don't think so. Keep in mind that those of us with legacy
applications are basically being totally hung out to dry. This was what we
attempted to make clear in the Ramparts article. Unfortunately, too many
people don't seem willing to fight any longer.

Apple doesn't need to resort to Java 'language purity' to improve
WebObjects... however they choose to do so.
Post by Randy Leonard
Guess this means ReportMill won't be available on NT after September? And
my product to compete with ReportMill starting next week can only be sold on
MacOS X Server after September? Lovely.
I'm not sure what this means for ReportMill, or your software.

You could sell deployment for NT, but not the other.
Post by Randy Leonard
Thanks,
Randy Leonard
www.oceansoft.com
-----Original Message-----
Sent: Friday, July 07, 2000 9:11 PM
To: Randy Leonard; 'Craig Brozefsky'
Subject: Re: Regarding Apple's decision to drop Objective C
Post by Randy Leonard
I did revisit Apple's policy at www.apple.com/enterprise/letter.html. It
seems to indicate you can sell AppKit/NT-based products to those customers
who already have WOF 4.5, it's just that Apple won't sell the deployment
license as a stand-alone product. If true, those selling AppKit/NT-based
products could get by selling a $699 WebObjects license with their product
instead of a $300 deployment license. But this seems kind of silly as it
will only burden deployment machines with lots of unnecessary software.
All the licenses have been modifed to disallow any non Apple Yellow Box
applications to run on WebObjects NT.
Randy Leonard
2000-07-08 05:26:14 UTC
Permalink
Scott:

I've received some comments that concerns for WOF/NT deployment licenses may
be overblown. It's probably a good idea for anyone that's developed a
Cocoa-based applicaiton for NT to contact their Apple sales rep for
clarification.

But I'm still having a hard time understanding Apple's official position on
WOF/NT deployment licenses. In the same policy statement
(www.apple.com/enterprise/letter.html) in which Apple announces the end of
deployment licenses for WOF/NT, Apple is also announcing the end of life for
DBKit and NeXTtime. Do you realize how old those products are? Why have an
official position to discontinue selling the most recent version of WOF/NT
so quickly when other products are supported for so long?

Randy Leonard
***@oceansoft.com
www.oceansoft.com


-----Original Message-----
From: Scott Anguish [mailto:***@digifix.com]
Sent: Saturday, July 08, 2000 12:33 AM
To: Randy Leonard
Cc: 'sanguish'; EOF list; webobjects list; ***@apple.com
Subject: Re: RE: Regarding Apple's decision to drop Objective C
Post by Randy Leonard
I'm understanding the anger expressed in this thread's original note a
little better now.
I don't think so. Keep in mind that those of us with legacy
applications are basically being totally hung out to dry. This was what we

attempted to make clear in the Ramparts article. Unfortunately, too many
people don't seem willing to fight any longer.

Apple doesn't need to resort to Java 'language purity' to improve
WebObjects... however they choose to do so.
Post by Randy Leonard
Guess this means ReportMill won't be available on NT after September? And
my product to compete with ReportMill starting next week can only be sold
on
Post by Randy Leonard
MacOS X Server after September? Lovely.
I'm not sure what this means for ReportMill, or your software.

You could sell deployment for NT, but not the other.
Post by Randy Leonard
Thanks,
Randy Leonard
www.oceansoft.com
-----Original Message-----
Sent: Friday, July 07, 2000 9:11 PM
To: Randy Leonard; 'Craig Brozefsky'
Subject: Re: Regarding Apple's decision to drop Objective C
Post by Randy Leonard
I did revisit Apple's policy at www.apple.com/enterprise/letter.html.
It
Post by Randy Leonard
Post by Randy Leonard
seems to indicate you can sell AppKit/NT-based products to those
customers
Post by Randy Leonard
Post by Randy Leonard
who already have WOF 4.5, it's just that Apple won't sell the deployment
license as a stand-alone product. If true, those selling
AppKit/NT-based
Post by Randy Leonard
Post by Randy Leonard
products could get by selling a $699 WebObjects license with their
product
Post by Randy Leonard
Post by Randy Leonard
instead of a $300 deployment license. But this seems kind of silly as
it
Post by Randy Leonard
Post by Randy Leonard
will only burden deployment machines with lots of unnecessary software.
All the licenses have been modifed to disallow any non Apple Yellow
Box
Post by Randy Leonard
applications to run on WebObjects NT.
Malte Tancred
2000-07-08 15:21:50 UTC
Permalink
Post by Karl Hsu
If you want ObjC, use 4.5. That's Apple's way of "keeping" ObjC - the whole
point of 5.0 is pure Java - as they discovered with previous versions, a
half-assed Java implementation is worse than no Java implementation (3.5's
Java support was so poor, it actually caused some companies to _stop_ using
WO entirely - just like Monitor in 4.5). Keeping ObjC "around" in 5.0 would
imply writing a new "reverse" bridge from ObjC to Java - not an easy task.
But 4.5 is *useless*!

We (as in the company I work for) will probably not have too much of a problem
moving to Java. It's not all fun, but we can make it without loosing to much
of the work, effort and *time* we've been putting into our Obj-C solutions.

But, again, 4.5 is *useless*!

Why do I say that?

Today 4.5 runs on Mac OS X Server, Windows NT, Solaris and HP-UX. Most of
the systems we build are deployed on Windows NT and Solaris. We do most of
our development on Mac OS X Server; Windows NT mostly when we're at the
customer's site, as most of our own machines are G4s and PowerBooks.

Now, Mac OS X Server is a *dead product* in its current form. In its next
incarnation it will be Mac OS X with the to-come Server Bundle installed,
AFAIK.

On what hardware will Mac OS X run?
All G3+ hardware? That's great!

On what OSes will 4.5 run?
For development, Mac OS X Server and Windows NT.

Ok, so we have a few (most of them actually) systems built in Obj-C that
we will have to support for a looong time. (and which we of course will have
to move to Java sooner or later, but it might not be a valid option *at the
moment*). How do we do that with the current strategy that Apple has regarding
4.5? (which is the last version supporting Obj-C)

1. Well, we need to keep a room full of old G4 and G3
systems that can run Mac OS X Server so that we will
be able to run 4.5.

OR

2. We can buy all brand new Intel hardware that we will
be able to run WO4.5 on, as well as 5.0, 5.5, etc, etc.

Now, we don't want to buy a lot of Intel hardware. We want to buy Apple hardware.
We want to run Mac OS X. We want to use Aqua. We want to use the new Project
Builder. But we need to be able to *support or existing, installed solutions*, and
we need this ability *on the new Apple hardware using their new OS*.

Is this too hard to understand for Apple?

Again, 4.5 is *useless*, from a developers perspective, as it is a technology that
can only be run on a *dead* OS. (unless we all move to the Intel platform).

4.5 could be VERY useful. Apple, please, please, please, listen to us!
*grovel in the dirt*

And, as always... Cheerio,
Malte

--
Malte Tancred, Oops AB
***@oops.se
http://www.oops.se
PhilipM=?iso-8859-1?q?=F6tteli
2000-07-10 01:58:18 UTC
Permalink
I work writing computer software for 19 years now and I've used
almost all
what companies have proposed us. When I've found YellowBox/EOF/WOF and
Obj-C, I was like I've seen Light ! I decided to move all my
business on it.
Here in Belgium, it was the time Apple pushed YB as a mean to
penetrate the
business world where Windows is king. I've created a company last year
specialized in WO and YB development. I've recruited two guys who
learn now
the whole bunch of tools. It costs me a lot because they're not yet
productive at present time.
And now, Apple tell me to throw all this, all my code, to tell my
guys "hey
Guys, guess what... you go now for Java" when I've tried to
explain them
that I consider Java as immature and that I would never put a cent
on its
future !
But now, how can I explain to Java lovers on this list what are my
feelings
?
Go Gnustep(Web). In my eyes, you only win except three points: You
won't have Quartz, only EOF 2-3 and programming tools (not the
libraries) are not as mature yet. In the whole rest you're winning
(price, performance, hardware, debugging, policy decisions, market
acceptance, reputation, server software, etc.). And the more people
are working on it the quicker it progresses.
Perhaps it will even run one day on Darwin and so also on MOSX.
Would be funny coming back to Apple with perhaps the better
development environment running in parallel with the original, which
in the meantime will run with an extended VB, C# or whatever the few
big clients of Apple want then. (Although I have nothing against C#
if it is as good for program designers as ObjC.)


Phil

--
lic. oec. publ. Philip Mötteli Tel# +41(22)320-3088
c/o Leila Bourahla Fax# +1(425)940-1642
23, Rue Prévost-Martin Tel# +41(78)742-9334
CH-1205 Geneva E-Mail: ***@yifan.net
Kieren Richard MacMillan
2000-07-11 22:54:32 UTC
Permalink
Michael --

First, some disclaimers:

1. I don't want to extend this thread any more than the next guy, but
<no-flame> nobody has been discussing any proactive solutions </no-flame>,
so I thought I'd toss my hat in the ring;

2. I don't know ObjC at all (I just picked up WO, and use Java 100%), so
none of this is directly a show-stopper for me -- I'm only doing it because
I feel your pain, and am a sucker for the underdog, especially when he's
using a superior technology (hence my love of Apple in the first place);

3. Sometimes, by sheer blind luck, my rants come up with something that
actually works, often because I can see the forest more clearly than the
tree-huggers... ;-)
Post by Michael Gersten
I don't care what language EOF/WOF is implemented in. Use a Java-native EOF if
you must. Just give us an objective-C brige (a .ojbs file?) as well.
Why not build the bridge yourselves?

If it's just a reverse of the current Java bridge, I would think that it
wouldn't be a big technical feat for a group of experienced programmers.

It's self-evident that a concerted effort to this end by a bunch of ObjC
programmers would cost less (in terms of development time, not to mention
downtime due to frustration and whining about The Way We Got Screwed) than
requiring every ObjC project being recoded into Java.

Apple shouldn't be uncomfortable giving away just enough proprietary info to
let you do that much -- heck (and here my ignorance bursts forth in all its
glory), maybe there's no need for *any* non-public code to be revealed, if
all you need is (in essence) an API to communicate between your ObjC code
and EOF/WOF.

The only possible downside is support, if ObjC "dies" (i.e. isn't recognized
by anyone except a dwindling population of die-hard programmers) -- however,
if you feel very strongly about ObjC, you'll work it out (e.g. form your own
ObjC support company, etc.), so this is moot in my argument.

You've stuck out your neck and (like it or not!) the blade is dropping -- as
I see it, you have four options:

1. Convince the executioner (Apple) to grab the trigger rope, which means
extra work, rope burns, public embarrassment, etc. -- outlook: bleak.

2. Let the blade drop, cursing the executioner right to the point when your
head rolls into the basket -- outlook: probably feels pretty good, but
definitely bleak.

3. Pull your neck out of the way of the blade and sign up for the
party/religion that put you there in the first place (i.e. move to Java) --
outlook: survivable, but unpalatable to you.

4. Figure out a way to stop the blade from falling any farther than it
already has (q.v. my suggestion, above, or a similar plan) -- outlook:
tough, and you don't have a lot of time to get it done, but much rosier all
around if it works out.

=)

Best wishes,
Kieren MacMillan.
sanguish
2000-07-11 23:36:07 UTC
Permalink
Post by Kieren Richard MacMillan
Post by Michael Gersten
I don't care what language EOF/WOF is implemented in. Use a Java-native EOF
if
you must. Just give us an objective-C brige (a .ojbs file?) as well.
Why not build the bridge yourselves?
It isn't that simple.
Post by Kieren Richard MacMillan
If it's just a reverse of the current Java bridge, I would think that it
wouldn't be a big technical feat for a group of experienced programmers.
According to the engineers that I've spoken to at Apple, it is a major
deal, and not something that they feel that they can get the same level of
functionality in the opposite direction (that is Java to Obj-C rather than
the way it is now).

This is one of the reasons that a pure-java version was written. They
needed to have the ability for the Java developers to use the 1.2 JVM just
as easily as the 1.1 JVM. This bridge caused problems in this area.
Post by Kieren Richard MacMillan
It's self-evident that a concerted effort to this end by a bunch of ObjC
programmers would cost less (in terms of development time, not to mention
downtime due to frustration and whining about The Way We Got Screwed) than
requiring every ObjC project being recoded into Java.
Except. Apple has an Objective-C based product NOW. WOF 5.0 for Java
should be an extension/add-on to WOF 4.5. The functionality is identical
outside of the additinal java features. The development tools are the same.

It is a decision that they need to make.. to support the developers
that have an extensive project base now. At least on a supported Mac OS
(Mac OS X.. Mac OS X Server is virtually an abandoned platform)
Michael Gersten
2000-07-11 23:36:02 UTC
Permalink
[I say Apple would be shooting all its developers and then itself if
it didn't embrace Java support]
Yes.. but they're already doing that. They're already shooting us.
They're not already doing that. They're shooting a very particular
set of developers with the WebObjects 5.0 release, namely those who
continue to rely on Objective-C and who didn't start Java transition
projects when they had the opportunity to do so. Yeah, it sucks to
be you, but Apple's not doomed because of it and neither is
WebObjects -- there are plenty of people who saw the writing on the
wall and chose Java for development. We see them here on the list
occasionally, though they tend to be shouted down by the Objective-C
partisans.
Didn't start java translations on time, when they had the opportunity?

Ok, tell me how I can debug a java app in 4.0 or 4.5?

Tell me how I can use java client with an EOF inheritence tree?

Tell me how I can have a java class split into several different files,
each doing a particular sub-task of that class? (Objective C categories
do this. Java requires that your classes be completely defined in one
large file per class, with no real modularity).

Tell me _why_ the java debugger that project builder uses (in 4.5)
requires that I click on the particular file when a breakpoint is hit,
and multiple files have the same name (think client side class.java and
server side class.java. Think multiple packages with a seperate
duplicate class name, such as "list" or "tree". Etc.)

Tell me _why_ PB's debugger buttons are so broken that they don't work
with java, and were shipped untested? "print * <foo>" "No symbol named
"*" in current context". (might be off slightly, that's from memory).

Now, tell me when I had the opportunity to start switching to java,
given that back in 4.0 days Apple was saying "Objective C will be around
for a long time, we still have too much stuff that uses it, we make
heavy usage of categories to extend NSObject with EOF stuff, etc".

Tell me that if I actually needed poseAs that I even had the CHANCE to
switch to Java.
There is no reason that Java can't be supported in addition to
Objective-C.
Actually, there is. Apple's priority seems to be to get the core of
WebObjects running on any compliant Java virtual machine so they
don't have to maintain all sorts of duplicate tools for different
platforms. How do you propose to make your existing Objective-C code
work in this environment?
Well, lets see. Apple already has a 100% java implementation of most of
the EOF system for java client. They fix that, add the WOF server stuff.
And, they bug-fix what is left of objective C. They then release version
5 with both a mixed language, and a 100% java implementation. **THEN**
They state that this is the last version that will support objective C,
after they have a working version that supports both.

Getting EOF/WOF to run on java VM's does not require throwing away the
working objective C stuff as well.
Kieren Richard MacMillan
2000-07-12 00:02:23 UTC
Permalink
Scott --
Post by sanguish
Post by Kieren Richard MacMillan
Why not build the bridge yourselves?
It isn't that simple.
Ah, I figured it wouldn't be... =\
Post by sanguish
Except. Apple has an Objective-C based product NOW. WOF 5.0 for Java
should be an extension/add-on to WOF 4.5. The functionality is identical
outside of the additinal java features. The development tools are the same.
Kieren Richard MacMillan
2000-07-12 01:06:26 UTC
Permalink
Michael --
Post by Michael Gersten
Well, lets see. Apple already has a 100% java implementation of most of
the EOF system for java client. They fix that, add the WOF server stuff.
And, they bug-fix what is left of objective C. They then release version
5 with both a mixed language, and a 100% java implementation. **THEN**
They state that this is the last version that will support objective C,
after they have a working version that supports both.
Okay, then here's the math, as I've interpreted it from this thread:

a = v5 is functionally no different from v4.5, except that it's 100% pure
Java [Proof: from Scott's post, "The functionality is identical outside of
the additinal java features. The development tools are the same."]

b = v4.5 currently works on MacOS X Server 1.2, Windows NT, and Solaris, and
supports ObjC code [Proof: self-evident];

c1 = MacOS X (client) runs ObjC apps just fine on the desktop [Proof:
self-evident];

c2 = EOF/WOF v4.5 is written (at least partially, if not totally) in ObjC
[Proof: my limited understanding of the technology, plus what I've culled
from this thread -- this might be where my theorem fails...].
Pierce T. Wetter III
2000-07-12 01:17:05 UTC
Permalink
Asked another way, if v4.5 (and, by extension, ObjC) was supported *as is*
on MacOS X client, would this thread even be happening?
No it would not. Everyone is upset because 4.5/ObjC was rather summarily
end-of-lifed. Except for two people, all of the existing WO user base would
be happy at this point if there was one more ObjC release for OSX client, so
there was at least one platform where you could maintain and deploy legacy
WO apps.
Okay, now that I've said all that, tell me the incredibly obvious point I
missed... =\
That Apple hasn't been responsive to the needs of _enterprise_ developers,
which means that we need tools that are around more then 6 months...

Pierce
Michael Gersten
2000-07-12 04:46:17 UTC
Permalink
A third "It isn't enough".
Post by Kieren Richard MacMillan
a = v5 is functionally no different from v4.5, except that it's 100% pure
Java [Proof: from Scott's post, "The functionality is identical outside of
the additinal java features. The development tools are the same."]
Flaw: v5 has bug fixes. V5 has "additional java features" that need to
be ported back to objective C.
Post by Kieren Richard MacMillan
If there is some 'real' reason why the Java part of WO 4.5
can't be ported to
OSX, would a WO4.5 "Lite" (ie. sans Java support) for OSX
appease the situation?
This really doesn't work for us either. My main concern is that I
have a desktop solution (currently YellowBox on NT, but could be
MacOSX with Cocoa) and a web solution both which use business logic
in ObjC. I really don't want to have to maintain businesslogic
frameworks for both Java and ObjC.
For me the big question is what's my desktop future which integrates
well with my web deployment (which is the main reason we choose NeXT
in the first place).
We've also heard that people want to do component by component
conversion to java (which I had also mentioned a while back, not quite
as clearly).

But more than that: 4.5 has bugs.

4.5 Java has some features that 4.5 ObjC does not.
4.5 Java client has a real nice feature -- automatic generation of UI --
that 4.5 ObjC or local java does not.

What I'd need to see:
1. 4.5 with the bugs fixed.
In particular the java client inherited EO.
2. Working mixed development environment
Including debugging support.
Permits porting piece by piece.
Permits a shared java business logic class for desktop, web server, and
web client apps.
2.5: Java client with full features/EOF support, not just a 95% of EOF
support.
Locking, in particular. (What else isn't there in 4.5?)

What I'd want to see:
3. Automatic generation of UI for NIB's
The technology based on d2web (d2JavaClient) can generate the
equivalent of a nib. I'd like to see the nib actually made. If not, I'll
be happy to have a java client front end just for the UI generation.
That is really a huge time saver. The EOF application wizard stinks in
comparison.
3a: Alternately, objective C or local java ability to auto-generate a
display window (IE -- no need to run appletViewer just to get the screen
shown.)

Note that 5.0 is supposed to have a much improved D2JC.

3b: UI generation for many/many joins (select two items in two display
groups, hit "join"). (This may already be there; I haven't played with
it yet).

4. New java stuff ported back to objective C
Includes all the XML stuff. (What else is new to / java-only in 4.5?)

What I'd like to see:
3c: UI generation for N-level joins (not just two level many/many).
In practice, I don't think there would be a need for more than 3. I've
never seen any models that even needed more than 3.
Paul Bucher
2000-07-12 12:53:04 UTC
Permalink
Date: Tue, 11 Jul 2000 21:06:26 -0400
Subject: Re: Regarding Apple's decision to drop Objective C
Asked another way, if v4.5 (and, by extension, ObjC) was supported *as is*
on MacOS X client, would this thread even be happening?
Kieren.
Yes 4.5 on X client would make most of us happy I believe. Most people will
still be upset that Apple didn't do this in the first place and got us all
upset. Which I can't blame them.

It still leaves the problem that there are people using EOF in Obj-C/Cocoa
for Enterprise applications and they don't have a clear path to the future.
But at least with 4.5 on X client there is a "high volume" & supported OS
for developers to deploy there applications on, which some have spent years
developing.

I think everyone is getting the picture that sooner or later they need to
migrate to Java for 5.0 & the future.

I think a piece that a lot of people have missed is that most Database
vendors are not porting there client side software to OS X client. There
was Oracle support on X Server(a EOF adapter) but it looks like it even that
won't make it to X client. With 5.0 WO we will be able to use JDBC thin
clients, we means we will be able to access almost any major RDBMS on the
market via a vendor supported interface. This will make future development
much brighter. Now if we could just have 4.5 & 5.0 on X client with some
kind of commitment to keeping 4.5 running on current versions of X with some
kind of bug fix support. We could end this thread.


- Paul Bucher
Randy Leonard
2000-07-12 14:32:44 UTC
Permalink
Post by Michael Gersten
At the least, how about releasing a bug-fixed version
of EOF/WOF 4.5/objC-Java along with the EOF/WOF 5.0
Java? Even if they can't get the new stuff in 5.0
back-ported to objective C, at least get a working
mixed language system with good debugging support at
the same time, so that we can port stuff safely if we
have to.
BINGO! And allow YB deployment on NT.


-----Original Message-----
From: Michael Gersten [mailto:***@getasia.com]
Sent: Tuesday, July 11, 2000 6:23 PM
To: Karl Hsu
Cc: Eric Hermanson; ***@omnigroup.com; webobjects list;
***@apple.com
Subject: Re: Regarding Apple's decision to drop Objective C
Post by Michael Gersten
Post by Eric Hermanson
As do most of the available Java libraries. Hit java.sun.com and you'll
find literally hundreds of horrendously designed libraries, most
suffering
Post by Michael Gersten
Post by Eric Hermanson
from severe class bloat and a lack of proper design.
Apple did such a *nice* job of designing their Obj-C code (AppKit,
Foundation, EOF, etc) that I am increasingly dissappointed they're
choosing
Post by Michael Gersten
Post by Eric Hermanson
to associate themselves in an irreversible way with this kind of garbage.
I
Post by Michael Gersten
Post by Eric Hermanson
still do not see why they cannot continue to offer WebObjects in Obj-C
after
Post by Michael Gersten
Post by Eric Hermanson
version 5.0, if not for anything else but to keep their existing X,000
customers happy.
If you want ObjC, use 4.5. That's Apple's way of "keeping" ObjC
And when new versions have new features that are not available? (think
auto-generation of user interfaces for desktops, or a way to freeze a
java-client display to a .NIB, etc).

And when bugs are discovered and reported and not back-patched/fixed?
(think inherited EO's and java client in 4.5 right now.)
Post by Michael Gersten
- the whole
point of 5.0 is pure Java - as they discovered with previous versions, a
half-assed Java implementation is worse than no Java implementation (3.5's
Java support was so poor, it actually caused some companies to _stop_
using
Post by Michael Gersten
WO entirely - just like Monitor in 4.5). Keeping ObjC "around" in 5.0
would
Post by Michael Gersten
imply writing a new "reverse" bridge from ObjC to Java - not an easy task.
Pure java? This is the point?

I'm not familiar enough with the java VM to be 100% certain of what I'm
about to
say. I do believe that 'java command' is outdated -- having 'command'
automatically run the java machine is more desirable (already done on linux,
probably done in the 5.0 system / Mac 10 user system).

I'm pretty sure that java permits a 'java myapp' program to use native
calls,
even if applets in the sandbox cannot. And, I believe that such apps can use
the
java debugger.

Pure Java? Aside from using just the native Java debugger, what does it buy
you?
No more call stacks that look like java -> objective C -> java to confuse
java
programmers?

That it runs on any java machine? Only those with the apple libraries. Yes,
it
means that you are not restricted to NT or Mac 10 deployment boxes -- you
can
now use any system. (Hmm -- maybe this is the point -- that Linux people can
now
use WOF/EOF for development. Apple, this may be a key point for you to
stress).

A reverse bridge? I think it's a great idea.

Other options? How about a dual version -- what if there was both a 100%
java,
usable on Linux, version of EOF/WOF, as well as the objective C equivalant
for
the existing mixed language implementations/applications on NT and Mac 10?
No
resources/no time? No answer to that.

At the least, how about releasing a bug-fixed version of EOF/WOF
4.5/objC-Java
along with the EOF/WOF 5.0 Java? Even if they can't get the new stuff in 5.0
back-ported to objective C, at least get a working mixed language system
with
good debugging support at the same time, so that we can port stuff safely if
we
have to.

Remember: Proper practice is to have both the old and new running at the
same
time until you are satisfied that the new works well enough to discard the
old.
By saying 'You can't go forward unless you throw everything away', Apple
breaks
that practice.

Michael
Kieren Richard MacMillan
2000-07-12 13:37:29 UTC
Permalink
Paul --
Post by Paul Bucher
I think everyone is getting the picture that sooner or later they need to
migrate to Java for 5.0 & the future.
Are there not enough ObjC supporters to make it a viable long-term platform
(i.e. with or without assistance from Apple)?

If you had told Linus Torvalds in 1990 that he should "get the picture that
sooner or later he needs to migrate to Windows for the future", he would
have laughed at you! ;-)
Post by Paul Bucher
I think a piece that a lot of people have missed is that most Database
vendors are not porting there client side software to OS X client.
I'm not totally clear on the technical stuff, but if a DBMS runs on BSD
(which a lot of them do), won't it run fine on MacOS X client?
Post by Paul Bucher
We could end this thread.
But... [sniff!]... wouldn't we all miss this? ;-)

Kieren.
sanguish
2000-07-12 18:11:56 UTC
Permalink
Post by Michael Gersten
A third "It isn't enough".
Post by Kieren Richard MacMillan
a = v5 is functionally no different from v4.5, except that it's 100% pure
Java [Proof: from Scott's post, "The functionality is identical outside of
the additinal java features. The development tools are the same."]
Flaw: v5 has bug fixes. V5 has "additional java features" that need to
be ported back to objective C.
How can Version 5 have bug fixes when the entire existing code base is
toast?

The additional java features in 5.0 are thinks like some level of EJB
support, and support for the JVM 1.2. Neither of these are relevant to WOF
4.5 Objective-C development
Post by Michael Gersten
But more than that: 4.5 has bugs.
No argument there..
Post by Michael Gersten
4.5 Java has some features that 4.5 ObjC does not.
Nobody is suggesting that the Java support be removed from 4.5.
Post by Michael Gersten
4.5 Java client has a real nice feature -- automatic generation of UI --
that 4.5 ObjC or local java does not.
This isn't relevant to the issue at hand. DirectToJava is cool.. but
doesn't any anything relevant to the 4.5 <--> 5.0 transition that kills
Obj-C and a good number of development projects.
Michael Gersten
2000-07-12 20:18:28 UTC
Permalink
Post by sanguish
Post by Michael Gersten
Post by Kieren Richard MacMillan
a = v5 is functionally no different from v4.5, except that it's 100% pure
Java [Proof: from Scott's post, "The functionality is identical outside of
the additinal java features. The development tools are the same."]
Flaw: v5 has bug fixes. V5 has "additional java features" that need to
be ported back to objective C.
How can Version 5 have bug fixes when the entire existing code base is
toast?
4.5 has bugs. Some are java-only, some effect both java and objective c.

5.0 will have bug fixes to the java stuff. It won't have those bug fixes
fixed in the objective C stuff.

I'm disagreeing with the statement "v5 is functionally no different from
v4.5". If that was true, I wouldn't be so upset.

It's not true. Java client will work in V5 :-).

Understand, I want java client. I really do.
I want an easy to use GUI. I'd like to see, not just NIB's translated,
but appkit calls translated.
I want the ability to use the same code (business logic) at either end.

I'd prefer if I could just put it all in the server to start, and then
move what is critical into the client. I can't, because they are invoked
differently (explicit invokeRemoteMethod calls).

I can't just have client and server class files because the debugger's
support for it is damaged at best.
Post by sanguish
Post by Michael Gersten
4.5 Java has some features that 4.5 ObjC does not.
Nobody is suggesting that the Java support be removed from 4.5.
No, but I do want to see those new java features available to objective
C users.

I could understand, if after coming out with a java only version,
something new was added that was not put into objective C.

But as long as you have a dual langauge version, why add something to
only one side?

Adding something to objective C can still be used from java server
because of the bridge.

Adding something to java client only makes it only usable from java
client.

Adding something to java server, without adding any cover classes to the
objective C side, well, ??? Yes, in theory, I could set up a bridging
specification, and do it myself, but I have to wonder: Is there
something about it that even Apple couldn't do it, and am I wasting my
time?
Post by sanguish
Post by Michael Gersten
4.5 Java client has a real nice feature -- automatic generation of UI --
that 4.5 ObjC or local java does not.
This isn't relevant to the issue at hand. DirectToJava is cool.. but
doesn't any anything relevant to the 4.5 <--> 5.0 transition that kills
Obj-C and a good number of development projects.
This was just another example of something being added to java in 4.5
without being added to the objective C side.

D2J is basically reading the model, reading a set of rules, deciphering
a UI, and then using XML to send that deciphered UI to someone that will
instantiate a swing-based GUI.

It is little or no different to read the model, the rules, decipher a
UI, and then instantiate an appkit-based GUI.

In fact, from everything I've seen or read so far, swing is a subset of
the appkit functionality. I'd suspect it would be trivial to do.

Now, maybe there wasn't time to finish the appkit version, and they only
had the swing/java client version ready in time to ship. If 5.0 was to
include objective C support, I'd expect to see it there, and I'd say
"Fine, I can wait". Now, I know it isn't coming.

4.5 is a version that has some new features in java, and some bugged
features in java.

There isn't a full featured development environment in 4.5 -- if you use
java client, you lose inherited EO's; if you use objective C, you lose
XML (maybe others as well, that's all I'm aware of that's new in 4.5
java); if you use server-only code, you lose auto-GUI.

And I still haven't heard anyone say that the debugging environment for
java in 4.5 is acceptable. I've heard people say that using a third
party java IDE for java client makes it acceptable; I think that's
pathetic for Apple.

Apple, what do you use to debug your java code?
sanguish
2000-07-12 18:14:52 UTC
Permalink
Post by Paul Bucher
Now if we could just have 4.5 & 5.0 on X client with some
kind of commitment to keeping 4.5 running on current versions of X with some
kind of bug fix support. We could end this thread.
Unfortunately, unless something has changed in the last 12 hours, Apple has
killed the address that we were able to send feedback to
(***@apple.com) about this issue.
P-J Nefkens
2000-07-12 19:37:44 UTC
Permalink
Paul,
Post by Paul Bucher
Yes 4.5 on X client would make most of us happy I believe. Most people will
still be upset that Apple didn't do this in the first place and got us all
upset. Which I can't blame them.
I can't do anything else but to agree with you...
Post by Paul Bucher
It still leaves the problem that there are people using EOF in Obj-C/Cocoa
for Enterprise applications and they don't have a clear path to the future.
But at least with 4.5 on X client there is a "high volume" & supported OS
for developers to deploy there applications on, which some have spent years
developing.
I think everyone is getting the picture that sooner or later they need to
migrate to Java for 5.0 & the future.
I think a piece that a lot of people have missed is that most Database
vendors are not porting there client side software to OS X client. There
was Oracle support on X Server(a EOF adapter) but it looks like it even that
won't make it to X client. With 5.0 WO we will be able to use JDBC thin
clients, we means we will be able to access almost any major RDBMS on the
market via a vendor supported interface. This will make future development
much brighter. Now if we could just have 4.5 & 5.0 on X client with some
kind of commitment to keeping 4.5 running on current versions of X with some
kind of bug fix support. We could end this thread.
However, since Apple reeaallyy asked us quite a bit to vote for WebObjects
on the Java Developers Journal this year, why won't we do a ballot and see
how many people have responded to the vote, and bring it to apple...

Check out http://www.nefkens.net/webobjects/ and put in your vote!

Regards,

PJ Nefkens

PS: I know, it's quite basic.. I'll be extending it in a couple of days..
Send me more requets, and I might put up more questions...
Bruce Fancher
2000-07-12 22:39:47 UTC
Permalink
That online poll is broken. I selected "Objective-C" and clicked on "Vote"
but the results still showed "Objective-C" as having 0 votes.
Post by Randy Leonard
-----Original Message-----
Of P-J Nefkens
Sent: Wednesday, July 12, 2000 3:38 PM
Subject: Re: Regarding Apple's decision to drop Objective C
Paul,
Post by Paul Bucher
Yes 4.5 on X client would make most of us happy I believe.
Most people will
Post by Paul Bucher
still be upset that Apple didn't do this in the first place and
got us all
Post by Paul Bucher
upset. Which I can't blame them.
I can't do anything else but to agree with you...
Post by Paul Bucher
It still leaves the problem that there are people using EOF in
Obj-C/Cocoa
Post by Paul Bucher
for Enterprise applications and they don't have a clear path to
the future.
Post by Paul Bucher
But at least with 4.5 on X client there is a "high volume" &
supported OS
Post by Paul Bucher
for developers to deploy there applications on, which some have
spent years
Post by Paul Bucher
developing.
I think everyone is getting the picture that sooner or later
they need to
Post by Paul Bucher
migrate to Java for 5.0 & the future.
I think a piece that a lot of people have missed is that most Database
vendors are not porting there client side software to OS X
client. There
Post by Paul Bucher
was Oracle support on X Server(a EOF adapter) but it looks like
it even that
Post by Paul Bucher
won't make it to X client. With 5.0 WO we will be able to use JDBC thin
clients, we means we will be able to access almost any major
RDBMS on the
Post by Paul Bucher
market via a vendor supported interface. This will make future
development
Post by Paul Bucher
much brighter. Now if we could just have 4.5 & 5.0 on X client
with some
Post by Paul Bucher
kind of commitment to keeping 4.5 running on current versions
of X with some
Post by Paul Bucher
kind of bug fix support. We could end this thread.
However, since Apple reeaallyy asked us quite a bit to vote for WebObjects
on the Java Developers Journal this year, why won't we do a ballot and see
how many people have responded to the vote, and bring it to apple...
Check out http://www.nefkens.net/webobjects/ and put in your vote!
Regards,
PJ Nefkens
PS: I know, it's quite basic.. I'll be extending it in a couple of days..
Send me more requets, and I might put up more questions...
_______________________________________________
EOF mailing list
http://www.omnigroup.com/mailman/listinfo/eof
Scott Anguish
2000-07-13 04:51:23 UTC
Permalink
Post by P-J Nefkens
However, since Apple reeaallyy asked us quite a bit to vote for WebObjects
on the Java Developers Journal this year, why won't we do a ballot and see
how many people have responded to the vote, and bring it to apple...
Check out http://www.nefkens.net/webobjects/ and put in your vote!
Regards,
PJ Nefkens
PS: I know, it's quite basic.. I'll be extending it in a couple of days..
Send me more requets, and I might put up more questions...
The problem with such 'polls' and 'petitions' is that they lack
context. You have no real knowledge if someone knows what they're voting
for, or is just voting for or against it to vote.

Look at how Apple won this year's JDJ awards. This is a PRIME
example. Mac faithful voted for what the Apple media told them to vote for.
(and I'm aware we had a small hand in this at Stepwise.. mind you our
readership IS developers..)

You need to know who the people are, you need to quantify they're
usage and marketplace. We went through all this when Apple said what they
were going to charge for Mac OS X Server to get the price reduced.. there was
massive back-end data collating required, classifying each respondent as a
certain type of developer/user, their seat requirements and platforms. Plus
the most compelling email was distilled down and summarized for Apple and the
WWW.

The place to tell Apple that you want WOF 4.5 on Mac OS X and longer
term support for Objective-C is ***@apple.com, but unfortunately,
Apple has shut that mailing address down.

My impression is that we have many allies within Apple on this, but
that management hasn't taken the issue seriously.
P-J Nefkens
2000-07-13 05:07:27 UTC
Permalink
Hi,
Post by Bruce Fancher
That online poll is broken. I selected "Objective-C" and clicked on "Vote"
but the results still showed "Objective-C" as having 0 votes.
It was kind a late last night and I made a small type in the config of the
vote... All the people that have emailed me the problem: thank you!

I will check my logs later to see what other votes I have missed on that
one! But I've corrected it now in a quick way...

Thanks,

PJ
P-J Nefkens
2000-07-13 05:16:01 UTC
Permalink
Scott,
Post by Scott Anguish
Post by P-J Nefkens
However, since Apple reeaallyy asked us quite a bit to vote for WebObjects
on the Java Developers Journal this year, why won't we do a ballot and see
how many people have responded to the vote, and bring it to apple...
Check out http://www.nefkens.net/webobjects/ and put in your vote!
Regards,
PJ Nefkens
PS: I know, it's quite basic.. I'll be extending it in a couple of days..
Send me more requets, and I might put up more questions...
The problem with such 'polls' and 'petitions' is that they lack
context. You have no real knowledge if someone knows what they're voting
for, or is just voting for or against it to vote.
Look at how Apple won this year's JDJ awards. This is a PRIME
example. Mac faithful voted for what the Apple media told them to vote for.
(and I'm aware we had a small hand in this at Stepwise.. mind you our
readership IS developers..)
I know, and now we're trying to do it the reverse way.. Apple won this by
letting the faithful mac users vote for webobjects. What if we would tell
all the Mac fanatics at all the websites (stepwise, macrumors, etc) that WO
(for which they voted) will not be supported on that way cool new OS called
OSX... See if we can use the same Mac Faithful base to convince Apple....
;-)
Post by Scott Anguish
You need to know who the people are, you need to quantify they're
usage and marketplace. We went through all this when Apple said what they
were going to charge for Mac OS X Server to get the price reduced.. there was
massive back-end data collating required, classifying each respondent as a
certain type of developer/user, their seat requirements and platforms. Plus
the most compelling email was distilled down and summarized for Apple and the
WWW.
The place to tell Apple that you want WOF 4.5 on Mac OS X and longer
Apple has shut that mailing address down.
BTW, there might be a way to get WO 4.5 running on OSX DP4.. I've heard some
stuff about renaming, but how many of you guys are ADC Member and have tech
support incidents that you're currently not using. There might be a way to
use those tech support incidents and report that WO 4.5 is not running on
OSX.. It says so on the CD (cover)... that we got from the WWDC ;-)

If enough people reported this incident, there might be a way that Apple
will spent time and money at working this out and getting a patch ready..
Post by Scott Anguish
My impression is that we have many allies within Apple on this, but
that management hasn't taken the issue seriously.
I know, but since I don't have the marketing information and/or the userbase
that Apple has who are registered as Developer, putting this only on the EOF
and WO mailinglist should give quite an indication on that we should all
shout to apple that we'd like to have WO on OSX...

Perhaps later it might be possible to cross-reference the poll's names and
email addresses with those of the mailinglists.
That might give some information..

Regards,

PJ
sanguish
2000-07-13 05:20:52 UTC
Permalink
Post by P-J Nefkens
Post by Scott Anguish
Look at how Apple won this year's JDJ awards. This is a PRIME
example. Mac faithful voted for what the Apple media told them to vote for.
(and I'm aware we had a small hand in this at Stepwise.. mind you our
readership IS developers..)
I know, and now we're trying to do it the reverse way.. Apple won this by
letting the faithful mac users vote for webobjects. What if we would tell
all the Mac fanatics at all the websites (stepwise, macrumors, etc) that WO
(for which they voted) will not be supported on that way cool new OS called
OSX... See if we can use the same Mac Faithful base to convince Apple....
;-)
Except that it isn't true. WebObjects 5.0 will be on Mac OS X. Trying
to explain to the masses the importance of Obj-C would be a major sell.

Plus, this same approach can easily work against you. Apple could say
that the results were tainted.

Apple management needs to be convinced. And now that
***@apple.com has been disconnected, we've kinda been given the
message that "they don't want to hear from us".
Kieren Richard MacMillan
2000-07-13 05:34:35 UTC
Permalink
P-J --
Post by P-J Nefkens
BTW, there might be a way to get WO 4.5 running on OSX DP4.. I've heard some
stuff about renaming, but how many of you guys are ADC Member and have tech
support incidents that you're currently not using. There might be a way to
use those tech support incidents and report that WO 4.5 is not running on
OSX.. It says so on the CD (cover)... that we got from the WWDC ;-)
If this is a serious option, I'm happy to donate one of my incidents to The
Cause!

Kieren.
Ronald C.F. Antony
2000-07-13 09:06:16 UTC
Permalink
Post by Michael Gersten
Post by Kieren Richard MacMillan
a = v5 is functionally no different from v4.5, except that it's 100%
pure
Post by Michael Gersten
Post by Kieren Richard MacMillan
Java [Proof: from Scott's post, "The functionality is identical
outside of
Post by Michael Gersten
Post by Kieren Richard MacMillan
the additinal java features. The development tools are the same."]
Flaw: v5 has bug fixes. V5 has "additional java features" that need to
be ported back to objective C.
WO 5.O has bug fixes? Really? Just think about what you're saying
here. WO 4.5 is a true 4.5 release. WO 5.0 really should be
called "WebObjects Pure Java 1.0 functionally equivalent to
mature WebObjects 4.5".
If you think that converting a body of code of the sort of
EOF and WO into pure Java will result in anything but a buggy
nightmare for the first few point releases, then you're
kidding yourself.
WO 5.0 will not be trustworthy for production software,
at least not where advanced features like heavily inheritance
based EOF models, etc. are concerned, for at least until
WO 5.3 or so rolls around.
Worse than that, in an ObjC based WO/WOF product, you can
fix a multitude of the smaller problems with posing,
categories and a host of other features that Java gracefully
decides are undesireable, which means with a Java based
WO you are *stuck* with the bugs until Apple decides to
fix them.
Given past history, certain bugs have been known, reported,
and still went unfixed for several releases, simply because
they happened not to affect the majority or the "current
focus group" of users/developers.
If you are part of the affected minority, with a Java based
product, you will be up you-know-what-creek without a paddle.
Post by Michael Gersten
Post by Kieren Richard MacMillan
If there is some 'real' reason why the Java part of WO 4.5
can't be ported to
OSX, would a WO4.5 "Lite" (ie. sans Java support) for OSX
appease the situation?
This really doesn't work for us either. My main concern is that I
have a desktop solution (currently YellowBox on NT, but could be
MacOSX with Cocoa) and a web solution both which use business logic
in ObjC. I really don't want to have to maintain businesslogic
frameworks for both Java and ObjC.
For me the big question is what's my desktop future which integrates
well with my web deployment (which is the main reason we choose NeXT
in the first place).
Exactly! With an ObjC based WO/EOF common core frameworks can be
developed for desktop, server and web based apps. With pure
Java the bridge will give the performance hit where it's the
most noticable, i.e. on desktop and server apps, where
network bottlenecks are not an issue.
Further, by being able to blindly leverage Java code out there,
Apple's motivation to provide API-consistency over a range of
different fields of application (database, GUI, 3D-graphics, etc.)
diminishes, which means programmer productivity will fall.
Apple's highly touted "10 times faster development" will be
a lie, because a large part of the 10 times increase in productivity
stems from consistent API conventions.
If some code is adopted from SUN, some from former-NeXT, some from
old-Apple, some from IBM, etc. we will end up with an unproductive,
inconsistent API melange that is reminiscent of the MS-DOS
programming days where one ended up buying toolkit libraries
from 5 different tool vendors, and each library got in the way
of the other and used totally different calling/signature
conventions.

Ultimately, what will a pure Java solution have over
the Java Apache project? Between the number of projects
converging there, and with Apaches market and mind share,
if I want a pure Java web solution, I go with the free
Apache stuff, which has a development momentum that
Apple can't keep up with, that's free (in the true
sense i.e. I get source, too), and that already does
certain XML things that WO can only dream of.
A pure Java WO will have nothing over the free stuff
within one to two years time, and that's about as much
as it will take to port the ObjC WO apps to Java anyhow.
So why bother with WO 5.0? I highly recommend that
people have a peek at what's going on at:
http://java.apache.org/
A few keywords:
JServ
ECS
Jetspeed
Turbine
SPFC
Cocoon
Avalon
Put all this stuff together, and ask yourself why you'd bother
paying even $700 for WO in a couple of years.
Now on the other hand, desktop-web integration, the way a
ObjC based WO provided, *that* was something nobody else
could offer and where the true value of WO is now that
thanks to Java more people get the idea that OO frameworks
make a good foundation to write web apps.

But the trouble doesn't end here. With Apple being able to
cheaply scrounge up more or less well designed Java
object frameworks, the question is, when, if ever, will we
see equivalent Cocoa frameworks? Where is XML support for
Cocoa, given that the WO people simply use the IBM stuff?
What will happen to database integration? How can AIAT
be used in common web/desktop frameworks e.g. for groupware
apps? etc.

Lobby Apple for support of an updated version of ObjC.
(Based on ANSI-1999 C, etc.)
People inside Apple have started work on that, but
we can't expect them to do this on their own time next
to having a fulltime+ job at Apple.
An updated version of ObjC with GC, namespaces, etc.
will do away with 99% of what people think makes Java
superior.
Then lobby Apple to actually do what has been talked
about again and again: a unified object runtime system
for Java, ObjC and potentially a few other, similar
languages, such as TOM, SmallTalk and possibly C#.

Enough said...

Ronald
==============================================================================
"The reasonable man adapts himself to the world; the unreasonable one
persists
in trying to adapt the world to himself. Therefore all progress depends on
the
unreasonable man." G.B. Shaw | ***@cubiculum.com | NeXT-mail
welcome
Ronald C.F. Antony
2000-07-13 10:11:48 UTC
Permalink
Post by PhilipM=?iso-8859-1?q?=F6tteli
(Although I have nothing against C#
if it is as good for program designers as ObjC.)
C# has a "virtual" keyword. Need I say more?
(For the language handicapped: that means by default it acts as
a static language, you have to explicitly turn on dynamic features,
and that makes things such as categories pretty much impossible, or
at least useless for 95% of the code...) And that's just the
beginning. C# is what it looks like: another C++ revision with a
few dynamic concepts added by people who really don't quite understand
their importance and relevance.

The only good thing that C# does is to torpedo Java's monopoly on the
web, and hence hopefully wake up some people to the news that
cross-language runtimes are the technology that could make one
vendor have an edge over the rest, and that language choice will
be inevitable, and that hence it's better to stay multi-lingual
now and expand on it, rather than going pure Java now and then having
to back-peddal in a few years when everyone may want C# or some yet
unknown language supported.

Ronald
==============================================================================
"The reasonable man adapts himself to the world; the unreasonable one
persists
in trying to adapt the world to himself. Therefore all progress depends on
the
unreasonable man." G.B. Shaw | ***@cubiculum.com | NeXT-mail
welcome
k***@heimdall.bifrost.com
2000-07-13 18:42:17 UTC
Permalink
Post by P-J Nefkens
Post by Scott Anguish
The place to tell Apple that you want WOF 4.5 on Mac OS X and longer
Apple has shut that mailing address down.
BTW, there might be a way to get WO 4.5 running on OSX DP4.. I've heard some
stuff about renaming, but how many of you guys are ADC Member and have tech
If anybody has 4.5 running under DP4 I would LOVE to hear about it. The
libraries are built for the Mach 2.5 kernel, not the 3.0, and I am unaware
of any tool to provide 2.5 emulation under 3.0.

Continue reading on narkive:
Loading...