Re: Glossary terms: questions

peter Murray-rust (
Tue, 9 May 1995 17:31:10 +0100 (BST)

Thanks very much,
On Tue, 9 May 1995, Iddo Friedberg wrote:

> Peter,

The first thing to realise is that the new glossary is marked up
in SGML, not HTML. HTML isn't sufficiently powerful for what we want and
so we are building our own language. Essentially HTML and CML are
precise parallels, since each is written using SGML. Therefore we have
to define all our types (ELEMENTs).

Secondarily, markup languages give viewers and servers latitude
in how they render data. Thus <A HREF="junk.pdb">JUNK</A> simply defines
a reference to a file junk.pdb. Until chemical/MIME this would be
treated by servers as a default file type (probaly text/plain). Even if
they treated it as chemical/pdb, the browser can only display it as
'Rasmol' if *it* also has chemical/pdb defined.

We intend that chemical/cml will be an additional type. That
means that it's up to your viewer how it's rendered. Actually we move
into new territory since it's unlikely that a single current viewer can
do all of this, although I'm talking with Roger :-) So, for example, a
file might look like:
<BIBLIOGRAPHY><AUTHOR NAME=petermr REF=http://www.cryst... TYPE=URL ...
<MOLECULE><ATOMS=555 ...>1.234 2.345 ...

This is program and platform-independent. The client has complete
freedom as to what to do. IN practice it might use RasMol to render the
ATOM info, GDE to render the sequence, and send the AUTHOR info for
rendering as HTML with links to the author's home page.

> I hope I haven't misunderstood your software concepts completely... Seems
> like you're out to create a very powerful tool.
It is *sufficently* powerful. The main thing is that it's also
human-friendly. And that the technology works. I think I now have the
glossary entry and editing almost complete under HTML (note the SGML is
rendered into HTML - but it could be rendered into Viual Basic or Linear B).


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