Jonas Maebe wrote:
> In article <d8d5b3$qne$1@[EMAIL PROTECTED]
>,
> Scott Moore <samiamsansspam@[EMAIL PROTECTED]
> wrote:
>
>
>>Good, so we agree that FPC is best for Delphi, and Mac users who might
want
>>to move to Delphi.
>
>
> It depends. At least Macintosh Object Pascal is extremely close to
> Delphi-style object Pascal (I was able to add sup****t for it in just
> this weekend, with only 80 lines added/changed in the compiler).
>
>
>>All I been trying to make clear here is that Mac Pascal versions are
already
>>ISO 7185 compilant, and moving their programs to Delphi/FPC amounts to
>>a language change. I don't know why the other members of your team seem
>>to want to gloss over that fact. Lets go forward with honest
discussions.
>
>
> I do not know whether it sup****ts all standard constructs (like put and
> get), but it's indeed quite possible it does. I also do not know which
> features are actually generally used a lot in Mac Pascal (but
> interprocedural goto's and local procedure variables are certainly among
> them).
>
>
>>There's no real reason it need be a separate language mode. The features
>>of ISO and the features of Borland can coexist, and have done so before,
>>in GPC and JPL Pascal.
>
>
> Afaik GPC also has a separate language mode and TP compatibility. I do
> not know JPL. It's possible there are no conflicting features between
> Borland and GPC, but if you combine both modes into one you no longer
> have either Borland or ISO, but a superset of both. Or does standard
> Pascal not require an error for constructs which are not defined in the
> standard?
I see, what you are calling a mode is the error flag that rejects
nonstandard
constructs. I wasn't calling that a mode simply because it does not
change the language accepted, simply the restriction of it.
For my part, I am only concerned with positive testing. Let me explain.
There are two parts to being compliant. The first is to accept all
constructs
of the standard language, the second is to reject all constructs that
are not the standard language. The first I call positive, the second
negative (as in what is NOT in the language). Although both are certainly
part of the standard, the negative check feature isn't technically
required to run standard programs. It also takes a lot of work off the
table for nonstandard implementations. All you must do is sup****t the
standard features, not identify and flag problems.
I'm only concerned with positive features, because I'm primarily
interested
in establi****ng a base language that runs on as wide as possible a
series of installations. If FPC group desires to also implement the
checking flag, thats great, but the goal of simply meeting the positive
standard can be had first.
I suspect that if FPC does get the constructs from the Mac, it will be
either compliant in the positive sense, or very close to it. Then the
fini****ng touch is to be able to run a positive standard suite.
For the purposes of that, the simplest first step is to run Pascals.
It has several features that are ideal for this purpose:
1. It is ISO 7185 compliant (in my version that I place online only:
the original had CDC 6000 features in it).
2. It's nontrivial.
3. Its publically available and do***ented.
4. It was written for the original Pascal by Nicklaus Wirth himself,
and thus is rooted in the time before the standard.
>
>
>>But again, I don't have a problem with FPC, it just needs a cointerpoint
>>to people here claiming that "of course" they should be dropping Mac
>>Pascal for the (incompatible) Delphi variant of the language. Delphi
>>is not Pascal, nor Mac Pascal. Converting to Delphi is converting away
>>from Pascal. If FPC was ISO 7185 (as in original Pascal language)
>>compliant, then it would be simply an extended Pascal, and that would
>>be the end of it.
>
>
> I guess this answers my previous questions about whether the standard
> requires strict (i.e., "negative") adherence or not.
It does. I don't :-)
Seriously, while negatives checking is certainly part of a "very good"
ISO 7185 compliant installation, I think getting a base language that is
widely implemented is the goal. I'm not even ISO 7185 "centered" on this
issue, I have been doing Pascal years before the standard. The reason I
fix on ISO 7185 is that it is a very clear statement of what was and
was not in the original Pascal language as put forth by Wirth.
I (as in I personally) am not that interested in pedantic mode flags,
and keeping your language strictly ISO so that it is ****table. Certainly
there are people who are, and I sup****t the negative checking flag
(the /standard flag) for IP Pascal. However, the real world is programs
that use extended features. No, to me having the base language in common
brings other benefits that I think you are discovering. For example,
Mac Pascal is standard Pascal+objects, and the objects are pretty close
to Borland. Thats probally not an accident, the story goes that Wirth
created the object extentions specifically for Apple, and Borland
copied a lot of that for their installation. So standard Pascal+
Borland objects is not far from a description of what Mac Pascal is
(at some point, I am going to do a better job of researching the
history of Apple Pascal in general).
>
>
>>Thats Borland. The question is why an open source group such as FPC
>>would want to keep this annoying and pointless divergence between
original
>>Pascal and the Delphi variant going. It serves no purpose.
>
>
> The only purpose is that
>
> a) we are only a limited number of people with a limited amount of time
> b) none of the people writing code for the compiler up till now had any
> interest in sup****ting standard Pascal, because they do not use it (and
> all worked with TP and/or Delphi from the start)
>
> We get tons and tons of feature requests and most are turned down due to
> lack of time (unless people are willing to write the sup****t
> themselves). Apart from people who think FPC should sup****t standard
> Pascal out of principle, I have not even seen any previous requests for
> sup****ting the standard (probably for the simple reason that there are
> already a perfectly working ISO-sup****ting alternatives).
>
I realize that. I think there are two basic points I would like to address
that with:
1. The remaining features of Wirth Pascal that Borland didn't implement
are fairly small. The amout of effort for say, the object oriented
extentions,
is way more than just to fix up the remaining ISO gaps.
2. The gaps are going to be the same as what is in Mac Pascal, so that
work is going to need doing in any case. So ISO 7185 compliance (in the
positive sense) can be seen simply as the fini****ng test for Mac Pascal
compliance.
>
>>Ie., I am suggesting FPC stop fighting over names, stop fighting
Borland's
>>pointless proprietary war, and call it a day. Then we can all move on.
>
>
> And I would suggest you stop claiming that we are fighting some war
> which on top of that is pointless. That would go a long way in avoiding
> irritated responses from people on the FPC team. It has nothing to do
> with a belief, malevolence or an attempt to exterminate the ISO standard
> Pascal world.
>
> It's just about time and personal interest, as is the case with many
> open source projects.
>
>
> Jonas
I think you have misstated my stance a bit here. What I have been saying
is that the standard (and original Pascal language) is relivant, and
some in the FPC camp have been claiming that it is not. To me, its a
word game. "ISO 7185 Pascal is irrelivant, we just need to worry
about Mac Pascal" is pointless. If you implement ISO 7185 Pascal,
you have Mac Pascal. Mac Pascal implemented Wirths original Pascal
better than Borland, thats all. The standard, and Wirths original
Pascal, are not pointless and irrelivant in this case. In fact,
its very much the center issue.
Is there a "war" ? Well, Borland is a big company (and used to be
bigger). They have done some amazing things with their Pascal. With
all of their resources, it would have been trivial for them to add
the remaining features of original Pascal they left out, and meet
the standard. That they didn't do, despite years of pressure to
do so. They have certainly demonstrated that they could do it. Borland
had a incompatible, proprietary version of C (as many compiler makers
did, including Microsoft), and they eventually caved in and produced
an ANSI compliant C. Borland, like many software makers, wanted to
have a proprietary language so that their customers were locked in.
With C they couldn't maintain that, there was too much pressure to
meet the standard. With Pascal, they did achieve it.
I'm not blaming FPC for this, just wondering why an open source
community would continue it.
And back to he original point (we are here in a Macintosh forum),
Mac Pascal users don't use this proprietary language, and I don't think
its in any Mac users interest to change their source code to meet
the proprietary Borland language.
--
Samiam is Scott A. Moore
Personal web site: http:/www.moorecad.com/scott
My electronics engineering consulting site: http://www.moorecad.com
ISO 7185 Standard Pascal web site: http://www.moorecad.com/standardpascal
Classic Basic Games web site: http://www.moorecad.com/classicbasic
The IP Pascal web site, a high performance, highly ****table ISO 7185
Pascal
compiler system: http://www.moorecad.com/ippas
Good does not always win. But good is more patient.
|
