Re: your mail

peter Murray-rust (
Sat, 10 Feb 1996 13:28:19 +0000 (GMT)

On Sat, 10 Feb 1996, Kurt Giles wrote:

> > The basis for the
> > divisions are in the broad categories that I as a biochemist think about
> > proteins.
> >
Maybe I was a bit too dogmatic yesterday because I wanted to avoid the
'Life Universe and everything' syndrome.

Here are two approaches:
- imagine PPS as a 'book' with a glossary in the back. That
glossary might well have 'Golgi', etc.
- take this as *the* world resource on protein structure so that
there are precise defintions of phi, labelling of aminoacid residues, etc.

I think these can co-exist - the former would be short, textbook-like
with pointers to other glossaries when they arise. The latter would be a
primary reference site.

*Don't feel bound by the terms we used*. I simply went through Alan's
backbone and extracted everything that was a technical term.

> Here's what I think the list should be based on Darren's original list and
> taking into account Peter's comments. I guess it doesn't matter too much since
> we should be able to move things around.

Yes. Forget the technology. (Lesley, I and others will try to work out
whether something is a 'related term' of a crossreference :-). Just go
ahead and write a list of terms, with definitions. Add URLs
(consistently if possible). Highlight all terms in the definitions which
are potentially in the glossary. (These are sometimes called 'entailed
terms'.) Then decide whether they are best included in the glossary or

Of course the crossreferences may not yet exist and could be replaced
temporarily by a brief description.

> --classification systems ****I think it would be a lot of unnessesary work
> ---SCOP ****to list all examples here, if links could just
> ---Enzyme commission ****be put into e.g. SCOP

agreed. Wherever possible use existing resources.
> -quaternary structure

> One important point is how basic do we start? (the PPS95 glossary contains
> terms like 'acid') This could produce a very large list of terms for

I think you have to assume a basic knowledge of chemistry and biology
(e.g. acid, cell, etc should not need referencing. It's always possible
to link them in later

> Here's how I envisage the structure of each entry:
> Definition
> [three to four lines max.]
> Related words
> [intra and inter VHG links, i.e. at least a list of the sub-directories or
> terms leading from the 'concept']
> Related information
> [other WWW links]

Agreed. Where there are likely to be several terms using the same URL,
use a term to describe it, e.g. Not:

This is in the <A HREF="">PDB</A>

This is in the <glossary>PDB</glossary>

(We'll address the namespace later - how to link to other terms.)

This means that maintenance is MUCH easier - if the PDB ever changed its
address then it would be much easier to change a single link.

` P.

Peter Murray-Rust, Glaxo Research & Dev. (; (BioMOO: PeterMR)
Birkbeck College,, CBMT/Daresbury,