We ran a most rewarding workshop and symposium on Semantic Physical Science last Tue/Wed/Thu (2011-01-10/12). The program is below. This is an interim post to keep the momentum going and to let others know that there will be in important continuing activity. There were about 25 for the workshop and another 10 for the symposium.
The event covered a lot of ground, with a lot that was new for many people and many different possible ways of tackling semantics. The immediate outcome is there was very good agreement on:
- The value of semantics
- What to do in the first instance
The areas we covered were:
- Computational chemistry (participants included STFC (DL), PNNL, Cambridge (EarthSci, Chem, Chemeng, Cavendish), QMC, Bristol, Imperial (unfortunately Henry Rzepa had to miss the last day)
- NMR. Concentrating on 1-D NMR spectra (Bruker, Varian, JCAMP), mainly (CSIRO and PNNL). Unfortunately Christoph Steinbeck (NMRShiftDB) couldn't come but we have a very good outline of what needs doing
- Crystal. Main interest in having a repository of published structures independent of any restrictions (e.g. Crystaleye)
The software is now clearly alpha or very close to it. We are particularly grateful to Andrew Walker for FoX and good prototype/proof-of-concept was made with NWChem and AMBER. To be successful we shall need the continuing support of the developers. JUMBOConverters is now distributable and straightforward to add new converters, validators and dictionary-lookup.
Chempound worked well and there is clearly demand for all 3 disciplines
We'll try to get a fuller summary out in the next few days. Basically we have agreed that the immediate phase involves:
- Firming up some conventions, especially NMR
- Pulling the various CML websites and resources together
- Deciding on how to do dictionaries. There are two schools of thought:
- One uber-dictionary for compchem. Anything that is program-specific isn't important
- One dictionary per program and an uber-dictionary for compchem
I am deliberately neutral on dictionaries, but I think the software has to support (2) [That make it able to support (1) as well]. I think during development phase it will be necessary to have a number of dictionaries, but we must be careful they don't get set in concrete. For NMR I suspect we shall have to have instrument-specific dictionaries
- Firming up validator software
- Firming up commandline converters
Many thanks to Charlotte Bolton for organizing all sorts of things. As with last time Charlotte will continue to support with documentation and publication.
I am also working hard on editing the videos. They aren't too hard sequentially in that everyone talked well and in a linear fashion, but I'm still not on top of WMV => ffmpeg => MP4. Anything 1280*960 is likely to drop image frames; half size is fine, but it's hard to read the screen. In any case they need editing. I'm going to post them privately to the participants in case there's anything they want to delete/add. We need to do this as MP4 [ca 100 Mbytes / 30 mins] because WMV is just far too big.
Cameron and I have some political messages about SOPA/RWA etc so those will probably be the first out.
More later. [Next week I have Open Biblio sprint, so not much time for relaxation! But it's fun].