Re: http://g95.sourceforge.net or http://gcc-g95.sourceforge.net

[Steven Bosscher <s dot bosscher at student dot tudelft dot nl>]

Mike Stump wrote:

> On Sunday, June 1, 2003, at 05:29 AM, Steven Bosscher wrote:
>
> > The link was not updated because forks are generally not a good thing 
> > in the Free software business.  The G95 project is still the official 
> > GNU Fortran 95 project in the GNU software database, but gcc-g95 is 
> > the project that will be integrated in GCC.
>
>
> The gcc SC can probably help make gcc-g95 the GNU Fortran 95 project.  
> :-) 

We have already tries that, but RMS would have to be involved, and as 
Toon put it: RMS should not have to do this kind of micro-management.  
Let's just get our tree in, and people will know.

> > The gcc-g95 project forked from the original g95 
> > <http://g95.sourceforge.net> project because the maintainer of the 
> > original project, Andy Vaught, is obviously a good programmer (he 
> > wrote most of the parser), but he has shown over and over again that 
> > he is a difficult maintainer to work together with on "his" project.
>
>
> I like your solution to the problem.  :-) 

Frankly, I don't.  In fact we have tried real hard to hold on to a 
single project, and only forked when things all of a sudden seriously 
changed for the worse (ref. earlier in this thread: 5 different g95 
projects on sourceforge :-\)

> > That does not make this whole fork issue any better, and we would 
> > still prefer to work with Andy.  He has been making steady progress 
> > in his own tree just like we have, but he has decided to only 
> > distribute binaries but to keep the source code to himself, 
> > unfortunately.  We are trying to work towards a solution for this.
>
>
> :-(
>
> Pretend that is doesn't exist.  The smart people catch up in a day, 
> the slow ones might take 3 years, eventually, they will come to 
> understand the truth. 


Well, we could do that and both shoot ourselves in the foot.  Andy has 
some pretty significant changes in his tree that we would really like to 
have in our tree as well.  Also, since he wrote so much of the parser, 
he knows the bottlenecks better than we do, so we could very much use 
his help to finish this compiler sooner.
OTOH Paul and I wrote the interface between the parser and the 
middle-end, as well as most of the runtime library.  We have made some 
big pushes in these areas.  We don't know if he takes our improvements 
from our tree, probably not.  But at least he _can_ do if he wishes to.

Executive summary: It is not in anyone's interest to just ignore each other.
(Though I am under the impression that he filters all mail from me to 
/dev/null...)

Note that we (you, I, anyone) have the _right_ to get the source code of 
his tree, by the GPL.  Just pretending that his tree does not exist, is 
an action that in IMHO undermines the GPL.

But you also have to remember that this guy wrote a parser for a 
language that is lexically and syntactically close to insane; he did so 
by hand, on his own, in his own time with nobody paying for it.  So he 
should be a pretty smart person.  Andy deserves a bit more credit than 
he would get if we pretend that he doesn't exist.

Gr.
Steven