PureBasic Forums - English

skywalk wrote:Yes, given enough frustration, I have considered writing my own. But I cannot justify the opportunity cost given the current IDE is really more than adequate. The reason I code is to save time!
The tease is freak's hint at a plug-in. But, after a year, I could have made dozens of small changes to an open source IDE that would have saved ME many keystrokes and clicks.

Or extend an existing IDE to handle PureBasic

I actually played around with adding PureBasic support to visual studio as a plug-in some years ago, but the PB IDE isnt that bad to justify that amount of work IMHO. Another obvious way would be to add PureBasic support to Ecplise. That way you could have both a great cross-platform IDE right away - plus you can extend it as you wish.

And about VD - why would anybody use VD (or care about it's state), when you have the free PureForm replacement and the cheap commercial PureVisionXP?

Open source ...heh funny thinking.
I read once on one topic that Fred don't agree if you wrapp purebasic functions in
your own programming language(read interpreter).
But he also think that is ok when he wrapp C or C++ functions developing pure basic compiler
if is written in C/C++ (i guess) ...
so talking about open sourcing is pointless..

To all the people who say no to open sourcing PB
because of the potential chaos that may follow:
It is true that some OS projects fail because of that,
but as already mentioned: All the successful OS
projects actually have a single maintainer (which
gets elected once in a while). There aren't multiple
different versions of the same thing, there are
branches/forks which you can check out and if they
are good enough, they normally get merged back
into mainline. Sometimes the mainline is to restrictive
and a fork gains all the community and therefore
becomes the new "mainline". I see only advantages
to that (since a non-mainline-compatible fork would
probably have no chance of surviving anyway). This
is like evolution, the best one survives.

To the people that mention the old OS IDE example:
I think the license for the IDE was GPL. This allowed
jaPBe to be developed and after GPI went WoW,
Gnozal could take over because the source was
freely available. If it was not wanted that the IDE
could be modified by anyone and published
for other languages, Fred & Co. could have chosen
another license. OS does not mean Free Software.
That means Fred could make the source available to
us but also disallow it from being distributed. That way
we could create patch-sets for the IDE with our own
features. Those patch-sets could be maintained by
somebody and maybe even incorporated to the
original IDE from Fred (maybe requiring signing of
a contributor's agreement).
There are many ways to open source a project.

This should be applicable to the IDE, the compiler
and all the libraries.

All in all I'd be in favour of open sourcing especially
the IDE. The compiler is probably too complex to just
jump in and add features and the libraries are not
essential, as we can just write our own libraries or
wrap other well known libraries.

But that's just me. If you're afraid of OS you might
want to read up on it a bit more and get some
experience yourself.

Cheers
remi

I agree, good post, Remi!

JaPBe has been violating the GPL since its existence:
- the GPL is included neither in the source nor the binary distribution (both is a must)
- there are no copyright notices from the original IDE in neither the source nor binary (also both a must). The only reference is the phrase "based on the original IDE".
- in fact, the GPL is mentioned nowhere in the source or binary distribution. The manual even includes its own two sentence license that is in no way similar to the GPL.

Other "forks" of the GPL'ed IDE sources were no different, and the contributions back to the IDE itself were minimal. All in all, making the source of the old IDE available was a bad decision from our standpoint.

That may all be true. So jaPBe violates the GPL
(which I vaguely remember we had this discussion
a long time ago). It is still open source and allows
people to modify it though. So you can't really be
angry about jaPBe leaving out the license?
I don't want to justify the violation of the GPL, but
when you release something under the GPL you
made a conscious decision that people are allowed
to copy the source, modify it and release it as some
different project. So adding the license notices to
jaPBe and everything should be fine.

Other "forks" of the GPL'ed IDE sources were no different, and the contributions back to the IDE itself were minimal. All in all, making the source of the old IDE available was a bad decision from our standpoint.

Understandably so, as you seemingly didn't want
to allow everybody to copy the code. You should
have chosen a different license.
About the contributions: That might be true and it
might not change. I can't tell you. What I can tell you
is, that if the IDE source would have been a "git checkout"
away, I would have been able to track down many of
the crash bugs to the line and possibly include a patch
in the bug-report (I don't know the complexity of
those bug-fixes but at least I'd have tried).

Well, nobody will force you. This is just some impatient
chap talking

Good heavens. It's up to fred and freak. Can't we just respect their decision? These topics seem to go on over and over and forever.

Maybe I add my $0.02...

In short:
Making PureBasic open source is IMHO one of the worst ideas ever.

The following is just the opinion of an old fart, but if you want to read it go ahead...

Not so short:
IMHO PureBasic is not in a position to benefit in becoming open source because Fred&Freak want PureBasic to stay the way it is now: a cool language for beginners and a decent language for experts that need to hammer some code together - fast!

As they already mentioned (several times) there is no space for OOP or the stuff that some users think is invaluable; a must have thing.
And some users would try to add this stuff into PureBasic if it would become open source.
(BTW: there is a nice add-on that integrates well into the IDE to add OOP functionality already)

Also fixing bugs in a compiler is not as easy as some might think, as it depends how the compiler is written:
1.) Decent
after certain coding aspects (divide code into sections tokenizer/lexer/parser/ast/emitter and make it modular with some intermediate lists/trees/IR or whatever)
or
2.) Intertwined
one big spaghetti-mix without the sauce (tokenizer/lexer/parser/emitter - all in one but... hey it works!).

IMHO the guys/gals that contemplate to help to fix compiler issues should first pick a compiler from sourceforge/github/etc. try to compile it and then try to add modifications to it.
At one point you will realize there is a lot of time and skills needed to get familiar with this stuff and really "own it". (you master the code and not the other way around... )

@Fred&Freak
Kudos for creating PureBasic as it is today.

@All the others
Even if PureBasic would be open source only 0.1% (or less) of the users would be in a position to help out the team. (without adding some additional burdens to the dev team)

For whatever it's worth:
If you have problems to understand my long explanation go to the top and read again the short version.

PEACE

rsts wrote:Good heavens. It's up to fred and freak. Can't we just respect their decision? These topics seem to go on over and over and forever.

Why do you want to stifle people asking questions? I've not seen a topic before about open sourcing PB. Even so an explanation of why it's good/bad depending on what you think would be better than ranting.

moogle wrote:I've not seen a topic before about open sourcing PB.

True, but it's very closely related to the hundreds of "add OOP or PureBasic will die" topics. In other words, it's just another post of people thinking they know best, rather than the owners and creators. But the team is obviously doing everything right, as the language has flourished and survived for over 10 years now, with only 2 lead programmers.

My thoughts on open-sourcing it? "Too many cooks spoil the broth".

@all: You may want to jump over this post, it's just my
opinion.

rsts wrote:Good heavens. It's up to fred and freak. Can't we just respect their decision? These topics seem to go on over and over and forever.

I like discussing such things, even if they prove to be utopian.
You're not adding to this discussion.

@fsw:
About OOP: I don't think anyone capable of adding this to the
compiler would actually bother, as they are capable of using
languages which already have this feature.
In any case, as I have said above, this OOP-fork would have
to prove itself before it would go into mainline. So why worry
about something that people obviously do not want here?

About the compiler: A compiler is nothing magical or complex,
it's just a huge piece of code. Until now, I've written several
small compilers (in both of your categories ) but I would
probably prefer well defined interfaces for plugins to the compiler
instead of the full code.

About the 0.1%: This might actually be a good estimate of
the ratio, but I'd range it more in the 1% range . Anyway,
if you look at any OS project out there, like e.g. the Linux
kernel - the ratio is around the same. A small group of people
can make a lot of changes (Fred+Freak=2).

@MachineCode
With a few exceptions, most people actually just voice their
opinions and don't tell Fred & Freak to implement anything.
If we cannot even discuss the pros and cons of something,
that would be pretty boring. So next time someone asks
for OOP (in a polite way), just give them a link to the old
discussions and ask them if they can add anything new.
If you are able to do that in a polite manner, I'm sure any
polite person will refrain from pushing without arguments.
You can rant at impolite people as much as you want as
far as I'm concerned

MachineCode wrote:

moogle wrote:I've not seen a topic before about open sourcing PB.

True, but it's very closely related to the hundreds of "add OOP or PureBasic will die" topics. In other words, it's just another post of people thinking they know best, rather than the owners and creators. But the team is obviously doing everything right, as the language has flourished and survived for over 10 years now, with only 2 lead programmers.

My thoughts on open-sourcing it? "Too many cooks spoil the broth".

Well I can't see where the OP has said about PureBasic dying if it doesn't go open source

However I do agree about too many cooks spoiling the "broth".

remi puts it very nicely though, I think maybe some sticky posts and "New members read this" announcements would help so that any posts like these would have something different instead of the same old stuff that has been discussed many times before.

remi_meier wrote: I like discussing such things, even if they prove to be utopian.

remi_meier wrote: About OOP: I don't think anyone capable of adding this to the
compiler would actually bother, as they are capable of using
languages which already have this feature.

Good point.
I like D v2 a lot. DMD has a small footprint, the GDC version is still developed.
AFAIK if everything goes well D v2 will be included in stock GCC 4.8.

The only things bothering me are:
1.) file size of compiled executable
I'm in the process to write a lexer/parser in D myself (just for fun) and the code size is around 45k (scanner/tokenizer/lexer already done) but the exe file size is over 1MB. The next thing I have to do is the parser and ast. (will be a 2 pass parser...)
I like to keep my brain moving

2.) no decent/easy/mature GUI
QT4D and DFL are stalled, GTK might be the only choice for now but never tested it myself...

remi_meier wrote: In any case, as I have said above, this OOP-fork would have
to prove itself before it would go into mainline. So why worry
about something that people obviously do not want here?

I concur.

remi_meier wrote: About the compiler: A compiler is nothing magical or complex,
it's just a huge piece of code. Until now, I've written several
small compilers (in both of your categories ) but I would
probably prefer well defined interfaces for plugins to the compiler
instead of the full code.

I'm with you on this one as well.

When I talked about the Intertwined compiler code - one big spaghetti-mix without the sauce (tokenizer/lexer/parser/emitter/linker - all in one) I talked from experience. Many years ago I worked on such a small compiler (coded in good old PureBasic) and it was able to generate 32bit Windows exe but compiling was slow (compared to the processed code size) and adding more stuff to it was cumbersome, due to the missing spaghetti sauce I guess; but... hey it worked!

remi_meier wrote: About the 0.1%: This might actually be a good estimate of
the ratio, but I'd range it more in the 1% range . Anyway,
if you look at any OS project out there, like e.g. the Linux
kernel - the ratio is around the same. A small group of people
can make a lot of changes (Fred+Freak=2).

Well April 1st is approaching and Fred promised some amazing news

Can't wait

Where has Fred promised amazing news?

The blog just says "maybe some big surprises".

well if he's gone to the trouble of supporting cocoa maybe he's bit the bullet
and had a go at targeting llvm.