[fpc-pascal] OT a bit - Planet Object Pascal

ik idokan at gmail.com
Sun Feb 17 21:30:53 CET 2013


On Sun, Feb 17, 2013 at 9:31 PM, Frank Church <vfclists at gmail.com> wrote:
> On 17 February 2013 17:10, ik <idokan at gmail.com> wrote:
>> On Sun, Feb 17, 2013 at 5:05 PM, Florian Klämpfl <florian at freepascal.org> wrote:
>>>
>>> Am 17.02.2013 15:36, schrieb ik:
>>> >
>>> >
>>> > On Sun, Feb 17, 2013 <tel:2013> at 1:58 PM, Florian Klämpfl
>>> > Why is it that FPC/K now exists for 20 years as an open source project,
>>> > and you
>>> > can't earn money for developing it,
>>>
>>> Actually, I'am not interested in earning money with it.
>>>
>>> >  while Linus that have an OS that
>>> > exists for 21
>>> > years can ?
>>> >
>>> > This type of things are important more then you think.
>>>
>>> For what? For the fun? FPC is probably by far the most successful non
>>> company supported oss compiler so I'am sure I know what is important.
>>
>>
>> Yes to solve everything but keep it for yourself :P
>>
>>>
>>>
>>> >
>>> > But not only this, take a look at the freepascal.org
>>> > <http://freepascal.org> web site: animated gif (from the 90's),
>>>
>>> Ever looked at gcc.gnu.org, kernel.org, llvm.org etc.?
>>
>>
>> Nice, Pascal is at the same level of usage and exposure as gcc, Linux
>> kernel and llvm,
>> nothing more to do, 0xPascal maybe, just to show that 201x can have
>> something new ?
>>
>>>
>>>
>>> > the whole way of thinking there is not designer oriented - It does not
>>> > invite new people
>>> > to the project,
>>> > does not provide a proper place to be etc...
>>>
>>> Why should I care? They might use the tools they like, I use the tools I
>>> like.
>>>
>>> >
>>> > It's not easy to find documentation for things, some are at the wiki,
>>>
>>> A compiler is not a toy, and I know that serious programmers find the
>>> needed docs.
>>
>>
>> Sure that compile is a toy, it can't do much, just parse something and
>> move it to be something else :P
>>
>> And documentation should be at one place, including new features,
>> workarounds etc... for old versions.
>> Look at:
>>
>> http://ruby-doc.org/
>> http://www.python.org/doc/
>> http://perldoc.perl.org/
>> http://golang.org/doc/
>> http://docs.oracle.com/javase/7/docs/api/
>> http://docs.embarcadero.com/products/rad_studio/
>> https://developer.mozilla.org/en/docs/JavaScript
>> http://msdn.microsoft.com/en-us/library/
>> https://developers.google.com/
>>
>> Should I continue, or do you get my point ?
>>
>>>
>>>
>>> > some at the /doc-html/ path
>>> > and many does not exists.
>>> >
>>> > So you are working on all this cool features and is the fastest compiler
>>> > in the world, and you can
>>> > do everything you wanted with pascal and many things you don't (and it's
>>> > not a c++ like
>>> > technology garbage can). But no new blood see it, use it or can be
>>> > attracted to it, so what's
>>> > the point ?
>>>
>>> FPC is still growing so what is your point?
>>
>> If a tree fall in the forest and no one hear about it, did it made a sound ?
>>
>> My point is that features are good, but when on one knows about it, then
>> what's the point of it all ?
>>
>>>
>>> _______________________________________________
>>> fpc-pascal maillist  -  fpc-pascal at lists.freepascal.org
>>> http://lists.freepascal.org/mailman/listinfo/fpc-pascal
>> _______________________________________________
>> fpc-pascal maillist  -  fpc-pascal at lists.freepascal.org
>> http://lists.freepascal.org/mailman/listinfo/fpc-pascal
>
> To Id in particular and others like minds I think there is something
> which isn't quite understood here. The compiler and the language ie
> FPC is quite a separate thing from the development tools and

FPC implement things. It does not implement all of Pascal ISO, and it
add a lot of things on it's own


> libraries, toolchains and the areas where it is applied. For instance
> if you read fpc-devel you will see that it is targetted at a wide
> range of CPUs and embedded devices, and that is not an area where the
> latest language developments are of much interest. Rightly or wrongly
> the compiler developers have become the focus where all matters FPC
> related are centred and this shouldn't be the case.
>
> I had my issues with this a long time ago and I even blogged about it
> - https://devblog.brahmancreations.com/content/observations-on-freepascal-and-lazarus-development,
> and it also resulted in
> https://devblog.brahmancreations.com/content/build-scripts-for-installing-freepascal-and-lazarus-from-source
> (which has resulted in almost 60.000 hits to date, LOL) but I have
> come to understand why things are the way they are.
>
> Some of your comparisons with other language projects are not quite
> fair. You mentioned Ruby, Python, Java but this comparison is not
> appropriate. Leaving aside the quality of documentation, something
> must be noted here. The languages you mentioned are more or less owned
> by a few individuals/corporations/committees who determine EXACTLY how
> the language is defined, the primary libraries and how they are
> implemented. There is only one Matz, one Guido, and one Oracle.

Yet, Google was able to create Dalvik
(http://en.wikipedia.org/wiki/Dalvik_%28software%29)
by using Sun/Oracle API, there are numerous implementation of
Ruby (http://jruby.org/, http://rubini.us/ for example) and Python
(http://www.jython.org/, http://pypy.org/), a lot more people to
contribute to it,
because it can offer things.


> FreePascal aims to support and retain compatibility with many
> different dialects which sprung up over the years and that isn't easy.

FPC is mostly attempting to be compatible with Delphi this days afaik.

> Graeme's complaints about what he considers an unwarranted desire to
> maintain compatibility with Delphi are legion, yet (ex) Delphi
> developers are the ones who can do the most to help FreePascal evolve
> if they don't buy into promises by Embarcadero to have Delphi working
> fine and dandy on Linux, which really means the Mac, as there is one
> too many variants of Linux consumer-wise.
>
> The main hindrance in the FPC world over the years, I believe, has
> been the reluctance to allow a modicum of documentation comments in
> the source code that would allow documentation to be built directly
> from the IDE with some CodeTools, ie in Lazarus. I think that is where
> it has been hurt most because it is not a corporate funded language
> and hobbyists have enough problems going through an edit-compile-run
> cycle to finish their projects, let alone document them in a separate
> files using plain text editors. There is more flexibility now and I
> think the upcoming Lazarus versions will make it easier.

I don't need an IDE to develop Pascal, unlike Java for example, I can use VIM
(and actually sometimes do), to develop. A language that must be with IDE
sucks big time.
I love Lazarus, but there are other IDE's out there as well, including
text editors
such as sublime and notepad+ etc...

I need documentation and understanding of what is implemented, why,
differences between versions, what to expect from one version to
another with my code etc...

It's not about the IDE/Editor, it's about the project, and FPC
developers know better what is
in plan, what is the direction, why things works as they do etc...

I'm an expert in the IT industry. I have a lot of knowledge in
information security, VoIP, databases
working with system tools (firewalls etc) and more. Building
applications that can use such technologies,
i do not have knowledge in compilers. I have written several parsers
in my life, but it's not the same.

Don't expect me to understand something that is not part of what I do
because then I need to enter to
an new set of field and learn it.

>
> One more thing lest I forget. The official FPC documentation is very
> good, especially for documents created by volunteers and hobbyists.
> That it is not accompanied by examples and demos is quite another
> matter. The solution to that is would be some kind of Pascal
> interpreter with a REPL that would enable FreePascal developers to
> have their Web 2.0 cake and eat it, so they could develop their IDEs
> and some web-based documentation CMSs speedily, but most of the
> efforts in that area, like ElevateSoft Web Buillder, Raudus,
> SmartMobileStudio etc are closed source.
>
> What the Open Source Pascal world needs most now is a good interpreter
> to start developing their IDE and documentation in, which is compiled
> to machine code after, not because Lazarus is not good enough, but
> turnaround times need to be fast for a hobbyist and volunteer driven
> project.

I'm working on making msgpack in native Pascal. I needed to understand better
how strings works in FPC (that are different then Delphi btw), and I
could not find
good documentation on each individual string type, I could not find a
lot of needed
information about the path the string is going to have in FPC. Did you
know that 2.7.1
have new way of working with locales of strings ?

When I create a library, I need to know such things to prepare my
code, or understand
that it is something I will need to do, only recently there where few
wiki entries about it.

With generics, can you give me good documentation of FPC about it ?
The wiki has old information, some of it not relevant anymore.

There are bugs in the way it is implemented with 2.6.0, and fixes for
it on 2.7.1, but the workarounds
of the bugs are not documented, I found them in the bug tracker by accident.
I could not find any reference for such issues on the wiki, or the
official documentation.

So no, the documentation is enough, you need more information, you
need different way of creating it.
For example the Ruby and Python way of documentation works a bit
different then the state of mind of FPC's.
They also explain the thinking behind it.


But you missed understand my whole claim here.
Pascal is an underdog, that many people are unwilling to hear about.
I'm doing a lot of evangelism in that area. I moved from talks about
"stop talking about Pascal, it is useless" to
talks such as "but FPC has static linking, and gcc has dynamic
linking, and dynamic linking is the
answer to life universe and everything".
That's a whole new world you see ? Instead of not caring, now it's "we
checked it a bit, but will not use", and that's
after almost 10 years of my own work on this.

Two months ago, I had a meeting at one of the biggest portals in my
country for something they want me to build them.
The whole room of the place I had the meeting, either been in one of
the lecture I gave, or seen something in open source
that I've done.
Everyone knew who I was when I entered the room. It's not simple or
easy to do such things, and when you do the right noise,
people at the end will go with what you say, just to prove you wrong,
and then when they can't they know something, and even
might start using it themselves.

That's what I do for FPC and Lazarus for the past 10 years +-, besides
some contributions.

>
> --
> Frank Church

Ido

>
> =======================
> http://devblog.brahmancreations.com
> _______________________________________________
> fpc-pascal maillist  -  fpc-pascal at lists.freepascal.org
> http://lists.freepascal.org/mailman/listinfo/fpc-pascal



More information about the fpc-pascal mailing list