Re: [cclib-devel] next release (was Molpro)

[Karol L. <kar...@kn...>]

On Tuesday 12 June 2007 19:58, Adam Tenderholt wrote:
> > I'm pretty busy at the moment, and haven't had a chance to really
> > catch up with what you have been doing, but we should start thinking
> > about what's going to be in the next release and start focussing on
> > this...
>
> I too have been pretty busy and have just finally caught up on
> reading all of the post you two have made in the last several days.
> Not that I'll have loads of time in the coming weeks as I'm hopefully
> wrapping up another project and going to start writing, but I figured
> I should provide some of my thoughts... ;-)
I am starting new projects now, and I would like to use cclib for them since 
that has proved pretty effective.

> I agree that we should start talking about the next release. Some
> things for discussion (or clarification for me as I'm not quite sure
> where everything stands):
>
> 1) Is the parser refactoring completely finished? It looks to be
> successfully merged into trunk, but I haven't really played with it.
The refactoring is finished to the extent I was willing to work on it. Some 
more things could be done for sure (possibly using a ditionary of functions 
like we discussed), but I have no time for this now. Still, after what was 
done the parsers should be easier to maintain and expand, and the progress 
output is more unified.

> 2) Do we still want to support Numeric with this release? I've
> installed numpy on my computer, so I doubt I'd even notice a Numeric
> bug.
They eigenvalue solver in Numeric (LinearAlgebra, actually) does not work 
correctly on my computer. I'm not sure if this is just my machine, but it 
breaks the Lowdin Analysis method I recently added. I am all for dumping 
Numeric before the next release.

> 3) The molpro parser: It'd be nice to include another parser, but I
> agree with Noel that we shouldn't include it unless it is basically
> finished.
I think I can get it to an acceptable state (usable by GaussSum/PyMOlyze). 

> 4) The XML format that Karol brought up in another thread. What if we
> created another module (say 'format') that allows parser objects to
> be saved in various formats? cclib could support XML, JSON, commented
> checkpoint files, etc.
Noel already got me thinking about formats other than XML. In the long run it 
doesn't matter to me, I just need to have a comfortable way to store (chosen) 
output from various programs. Even better if it takes up less space than the 
output text files.

> 5) DOS plots. This is something I've been meaning to add back to
> PyMOlyze at some point, but since it's something that's also used in
> GaussSum, along with just being generally useful for random scripts,
> what do you think about including some code in the method module for
> this?
I've never used DOS before in my work but they are nice to look at :)

-- 
written by Karol Langner
Tue Jun 12 22:43:53 CEST 2007