[Lazarus] Why the Java became so strong?

[Marc Santhoff]

Am Samstag, den 25.02.2012, 17:19 +0000 schrieb Frank Church:
> 
> 
> On 25 February 2012 13:52, Marco van de Voort <marcov at stack.nl> wrote:
>         On Mon, Feb 13, 2012 at 02:55:17PM -0200, Everton Vieira
>         wrote:
>         
>         > How to improve things:
>         > - Better container lib.
>         > - Better documentation.
>         > - Easier installation.
>         > - Publicity! This would be the most important now. The other
>         parts are in a decently good condition.
>         >  Successful projects have a public relations side-project,
>         advertising themselves somehow. Product releases are one way
>         to get free advertising. Thus it is very bad for publicity
>         that Lazarus has releases so seldom. A product release is
>         always mentioned in some programming site, read by potential
>         users. Without releases this looks like a dead project.
>         
>         
>          - get billion dollar investments of big IT companies.  Java
>         did too :-)

- anything is (or should be) an object
- no pointers that could get wild
- write once run anywhere

That was one of the starters, because it could run on any OS Sun or
someone else made a VM for. There actually have been many more cpu types
for server hardware.

- it was the new kid on the block

Trying to remember that really was a reason for many people jumping
in ...

- it look easy at first sight (and mostly is IMHO)
- it was available free of charge
- it looks much like C(++)

Much of all this holds true for python as a tool more oriented to
scripting and text processing. There were libraries for mostly anything
in amazingly short time. And it is readable, compared to perl. OK, perl
is unreadable compared to anything, even compared to egyptian
hiroglyphs ...

> There are so many points here is not possible to discuss them all.

I agree, I tried to list some of the more obvious reasons.

> Many Pascal programmers are well versed in the language, and the
> associated tools, having began using it even before the Turbo Pascal
> era. let alone the Dephi/Object Pascal period. They are quite fine
> working as they are and can get to speed quickly. Newcomers and even
> some of old timers are the  ones who ask themselves "If I am starting
> a new project, is Pascal the right one? Of course not!" Newcomers
> problem will be familiarity with the language, old timers issues will
> be the libraries and associated tools.

Yes, but at that time Delphi was nearly dead because of the
Borland/Inprise/whatelse? confusion and many programmers stepped back
from using it. Especially in germany pascal had a very broad acceptance
based on historcal reasons. I remember being a pascal compiler being
comparatively low priced for CP/M2.2 besides pascal was taken for
teaching.

I'm not absolutely sure, but IIRC fpc and lazarus were not as bug free
and mature as they are nowadays. Using fpc 1.9.x/2.0.x and GTK1 for a
program got me some more grey hair. ;)

> How about the reluctance to put documentation in library code? The
> policy with fpdoc is to create documentation separately from code. The
> problem is who has time for that. If documentation goes into the code
> it is better as it forces you to think clearly about what your are
> doing as you are doing that knowing that others depend on it,
> especially if there is a strict community code of ensuring that docs
> in the code are always correct and up to date, and that if they are
> not accurate they should be removed altogether. I am developing a
> FreePascal Lazarus program now, owing to my familiarity with Delphi,
> but at the same time I am keenly aware that in domains I am unfamiliar
> with I would probably better using Java, Python because the tools and
> libraries would be much better.

Documentation is a central issue. As are tutorials and the like for
beginners. I know there are many, many very good example programs, but
at lest they could be made more visible to the newcomer.

Take IBMs developers side as an example. There are tutorials explaning
the most basic thiongs in very short time and amounts of text. First you
think "WTF?" but after reading two or three of those simple texts you
can actually start betting productive.

But you can work that way, mixing code and documentation is what
pasdoc[1] uses. I've tried it, and it is nicely working. Same for fpdoc,
but that was a bit of a rough experience. It took me a while to get it
working at all (generating skeletons IIRC) and the window inside lazarus
never made me happy - got to think about it, dunno why.
In the end I'm writng docs for my own stuff using Openoffice writer and
a template adapted to showing source code snippets. One to threee pages
for tutorials or small unit docs, very readable, very handy writing
support - how did I ever think about writing raw xml files? ;)

> Mind you I am not denigrating Pascal in anyway, but it would be more a
> labour of love than one of economic utility, when economic utility
> would be the better driver.

Choosing the right tool for the job is wise in any case. I do, using C,
Python, Java - but if possible my language still is and will be object
pascal.

But before I forget:
Lazarus and fpc have grown up to a really strong pair of mature tools.
Whenever possible it's my first choice using them. Many thanks and
congrats to anyone making this possible.

And something more:
I can hardly close my mouth reading that lazarus and fpc are generating
java byte code and are having an android widget set. I'll try using that
as soon as possible.

Keep on good work,
Marc

[1] http://pasdoc.sipsolutions.net/

-- 
Marc Santhoff <M.Santhoff at web.de>