Re: some thoughts

[Peter Wainwright]

Marko,

I think it's healthy to have an uninhibited discussion of future
directions for DDD. I think that in if DDD is to survive as an active
project, then there needs to be some prospect of real functional changes
to attract new developers. A agree that a major priority has to be
debugging the debugger - I still encounter far too many random crashes
as I use it in my day-to-day work. However, I would rate a new GUI as
fairly important if DDD is to retain developer mindshare against younger
projects such as Nemiver, for example. The gtkmm version is fairly well
advanced in its own branch of the svn repository.

Looking forward, one thing which would make my life much easier would be
support for multi-thread/multi-process debugging such as I believe is
offered by the commercial product TotalView. I've recently had to debug
MPI programs which I have done with multiple instances of DDD - It's not
pretty. A multi-process DDD would be a useful and ambitious project.

Peter Wainwright

On Wed, 2008-10-08 at 14:28 -0400, Marko Schütz wrote:
> Dear All,
> 
> I just subscribed to the list, but have been a user of DDD for quite
> some time.
> 
> The reason I subscribed is that I found some postings on the web that
> I wanted to comment on.
> 
> "rewrite DDD in C++"
> IMHO that will tie up quite some effort. But will it bring substantial
> improvement? I doubt that very much. I have not looked at the source
> code of DDD yet, but I assume/hope that it's fairly clean. DDD has not
> seen substantial development in quite some time and I seriously doubt
> that rewriting it in C++ would be a good way to invest the little
> volunteer effort that goes into it right now. I also doubt that this
> is a direction that would attract many new volunteers.
> 
> "change to Qt, GTKmm" While I agree that the GUI feels antiquated,
> again I do not think the effort required by a project like this would
> be well invested. Pretty much for the same reasons as above. I
> consider forking DDD into two projects, one using Qt the other using
> GTKmm a very bad decision as this will further divert any
> contributions to the project.
> 
> "complete control by keystrokes"
> That I actually think would be a good way forward.
> 
> "debugging"
> This would be my number one priority. Fix all currently outstanding
> bugs, collect user feedback on bugs not yet reported, fix those
> too. In general make DDD more solid.
> 
> Best regards,
> 
> Marko
> _______________________________________________ ddd mailing list 
> address@hidden http://lists.gnu.org/mailman/listinfo/ddd