[Savannah-hackers] Re: GNU Speech status

[David R. Hill]

On Wed, 2 Jul 2003, Richard Stallman wrote:

>     I had asked if there was any reason why I should not put the sources on 
>     the Peanuts and Peak sites for the benefit of the NeXTSTEP/NeXT community 
>     as part of making it public.  I think it was you who demurred.
> 
> Now I remember that discussion.  The idea is that it makes no sense to
> release something as a GNU package that only runs on some other
> system.

My thought was that, since the code is already available under the terms 
of a GPL, it might make sense to make it available at several places to 
establish that the GPL really does cover it.  Your argument is also 
reasonable, and I went that route.  However, in one of your recent 
emails, you say that the sources should be available along with other 
information on:

        www.gnu.org/software/gnuspeech

which is what led me into the current round of (probably incorrectly) 
resubmitting the project.  By "sources" you may mean the ported sources of 
a newly working GNU/LINUX system, or you may mean the sources we started 
with (the latter are, in principle at least, issued already under a GPL).

I can access the CVS repository and copy the software from there myself as 
a developer, but there is an implication that the URL access above would 
allow others (i.e. "the public" to access some version of the software.  I 
also got/get the impression that there should be other stuff available for 
public inspection to encourage interest and helpers.  I think this would 
be good, but having accessed some of the other projects, I don't see much 
that would be of interest to other than developers writing code in the
equivalent places in those projects.

In fact there's a whole bunch of information on the CVS site for gnuspeech
that could be moved over into a public access area, or I can create web
pages that link to my university site, or both.

I am working on these issues at present and can appreciate that this is 
not a normal situation, which undoubtedly (a) provides a basis for 
confusion, and (b) means the standard documentation doesn't directly 
apply.  

> 
> I'd like to know how far you have come in changing it so it can run on
> GNU.  If you have come far enough, maybe now it makes sense to release
> the code.

We have got stuck into it and things are happening.  There's a work and
reporting schedule.  I have real-time speech on my system, but there are a
number of integration and re-write issues still to address.  It will
likely help that FSF now appears to endorse free Java use (we were using
Blackdown anyway as an intermediate step so, assuming it is OK for the
final software package, this could speed things up quite a bit).

> 
>       The 
>     savannah site was by definition not public, while under construction, and 
>     I missed the idea that there should be web pages including the sources at 
>     some other fsf/gnu site.
> 
> We ought to have some sort of web page about it, so people know what's
> being done and so we can recruit helpers.
> 

Yes.  I'd like some specific guidance on the mechanics of where the "some 
sort of web page" would fit into the general scheme of savannah/gnu.org.  
I presume it would hang off:

        www.gnu.org/software/gnuspeech

in some way and still be publicly accessable using some other URL.  If I 
know where to put it, and have access, I can set something up pretty 
quickly (though there may be style issues to cope with -- I wonder what 
tools/libraries you may have to support this).

>       Maybe you 
>     need a Texinfo manual for those people responsible for starting/running 
>     a Gnu-project.
> 
> We have two of them: standards.texi and maintain.texi.  They should
> have been sent to you--were they?  But they treat only common
> situations, because those are the ones that seem worth writing up.
> This situation, contribution of an existing program which does not
> already run on GNU, is quite rare.
> 
No, I didn't get either of those manuals, but you did send me a copy of 
the manual on Texinfo itself.  As I said before, I realise the gnuspeech 
situation is not normal, so maybe those manuals weren't sent because they 
did not include stuff that really applied.  Be interesting to see copies 
though.