Re: [cclib-devel] Update

[Adam T. <a-t...@st...>]

> I still need to look at the ADF SCF criteria...but I will.

I already added code for ADF. Hopefully you'll agree that it does  
what we need it to do. I updated the wiki to reflect what's implemented.

> I'm going to take a lesson in Jaguar either tomorrow or early next  
> week,
> as I need to tweak the settings on the test files. E.g. the geometry
> optimisation converged immediately (given the G03 output coordinates),
> so I don't have any information on the output of a geoopt. Also, I  
> found
> that virtual orbitals aren't printed unless a particular keyword is
> given. And I need to figure out how to print the overlap matrix. So,
> lots to be done there - but the test files need to be created first.

Ok. You're using Jaguar 4.2, right? I think someone in my lab uses  
Jaguar 5, so we should probably get some of those files because I  
believe the output is different between the two versions...but I may  
be wrong.

> We also need to start adding unittests for some of the other
> calculations, e.g. the unrestricted calculation.

Yeah. I'm not as familiar with the unittest framework, but I suppose  
I can learn. ;o)

> I'm thinking of making a TODO page on the wiki, which will focus us a
> bit more on what we need to do before we make a release. How do you  
> feel
> things are going? Obviously, it's taking longer that we hoped at  
> first,
> but I think we don't really have any technical difficulties left -  
> it's
> just a question of ironing out bugs and the unittests should do a
> reasonable job at that.

I think a TODO is a great idea. I feel like things are going a bit  
slower than we anticipated, but what project doesn't have that  
problem. From the way the table on the wiki looks, it appears we are  
making some progress.

> In fact, I'm also thinking about making a preliminary release as  
> soon as
> possible - to get all the release stuff working (e.g. uploading to  
> PyPi
> and Sourceforge). Also, we can start incorporating cclib into our own
> projects.

I think we should make a release fairly soon. I'd be completely happy  
with an initial release if it just parsed aonames/fonames, mocoeffs,  
and aooverlap/fooverlap since that's all I need to incorporate it  
into PyMOlyze.

> I'm very annoyed with Google as it still does not pick up the wiki.  
> This
> seems to be a permanent problem although I've tried to do some magic
> with Google Sitemaps and a php file (off the web) that generates a
> sitemap for MediaWiki. I think I will contact Geoff at OpenBabel as  
> they
> also use a wiki but they don't have this problem.

I have no idea in this area. I remember it took a while for Google to  
pick up my other projects. I think it wasn't until after you had  
linked to it from linux4chemistry. I dunno for sure tho.

Adam