Re: Chemical markup language

peter Murray-rust (p.murray-rust@mail.cryst.bbk.ac.uk)
Thu, 4 May 1995 21:22:33 +0100 (BST)

On Thu, 4 May 1995, Lesley West wrote:
>
> The prototype looks exciting. It almost seems a shame to develop entry
> creation when you can use feeder files like mopaq :-) !
MOPAC. This is a *data* file, and will need glossary entries for
THEOCHEM. I think it may be practical to generate the DTD from the
glossary - ambitious, but possibly correct. Alternatively we may need to
validate the DTD and the glossary against each other.

>
> > I also haven't quite mapped out the name space yet for cross
> references.
>
> I had a look at your cross reference file. I think it's nearly there. My questions may help clarify this section:
> 1. In your example, namespace is REF = "glossary:alt.uu/gna".
I rethought this on the train today. References to the
hyeprglossary will have to be through cgi-bin. I think we shall also
have to develop our own namespace, perhaps following the virtual library.
So:
http://vhg.bbk.ac.uk/cgi-bin/vhg?
glossary=bio:/protein/structure&term=alphahelix
(it's all one line, but I've broken it at ? for the mailer).
This goes to BBK (say). It checks the glossary table. BBK has
bio:/protein/structure locally. It retrieves the entry. Let's assume we
then try
http://vhg.bbk.ac.uk/cgi-bin/vhg?
glossary=art:/impressionism&term=monet
It will fail to find the glossary at BBK, but will find that
art:/impressionism is located (say) at vhg.anu.edu.au . It sends a document
back to the client saying:

Entry not found at BBK, try
<a href=http://vhg.anu.edu.au/cgi-bin/vhg?art:/impressionism&term=monet

The beauty of this is that until we get, say, > 10**4 glossaries :-), the
various VHG sites only have to update quite infrequently.

> 2. Is "glossary:" always going to repeat?
> 3. Are subjects at the end of the REF? e.g.glossary:alt.uu/gna/vsns-pps?
See above -
> 4. Or are subjects an <INDEXTERM>?
Thanks. The simplest way is then to either use uniqid=atp or
indexterm=adenosine+triphosphate. Obvious, but I hadn't thought of it
till your question.

> 5. Are terms with the same names & different meanings etc in this file?
The question will always (at present) specify a glossary and a
term/uniquid. It will be fairly straighforward to write:
glossary=bio:/protein/*&indexterm=*enzym*
but obviously we can only allow that one server. (I don't want people
dumping the whole hyperglossary through:
glossary=*:/*&indexterm=*
!!!

>
> Lesley West for PPS-glossary team

Thanks Lesley,
This is going to FLY!

Peter Murray-Rust, Glaxo Research & Dev. (pmr1716@ggr.co.uk); (BioMOO: PeterMR)
Birkbeck College, ubcg09q@cryst.bbk.ac.uk, CBMT/Daresbury mbglx@seqnet.dl.ac.uk
http://www.cryst.bbk.ac.uk/PPS/index.html, http://www.dl.ac.uk/CBMT/HOME.html