[MMTK] Re: [PyMOL] improving interoperability between python moleculetoolkits (fwd)

Eugen Leitl eugen@leitl.org
Mon, 16 Dec 2002 11:38:48 +0100 (CET)


On Fri, 13 Dec 2002, Richard Gillilan wrote:

> Even XML is very verbose for molecular data, but I haven't seen any
> better way to write general objects to disk.

Apropos CML and verbose, I'd like to cite a recent post from Peter 
Murray-Rust from svg-developers:

Date: Sun, 15 Dec 2002 09:07:44 +0000
From: Peter Murray-Rust <peter@ursus.demon.co.uk>
Reply-To: svg-developers@yahoogroups.com
To: svg-developers@yahoogroups.com
Cc: h.rzepa@ic.ac.uk
Subject: [svg-developers] CML and SVG (definitive and final)

I am not a regular reader of svg-developers but I have been informed that
a flame war arose on it about CML (Chemical Markup Language). I don't want
to continue this - please treat this as a last posting :-)

Firstly I am a great supporter of SVG - I applauded it on XML-DEV 2-3
years ago as the first killer-app for XML-over-the-wire. I am delighted to
see such strong support for it. I hope that it becomes standard in all
modern browsers and we can rely on this - if anyone has good news on this
front let me know (off-list!).

Secondly public thanks for Michael Bierman and colleagues for the CML/SVG
application at adobe.com. This has helped to evangelise CML.

Thirdly modern CML has been informed by the SVG microsyntax and is
reasonably compact. When I saw the path attribute it convinced me that we
didn't have to mark up every tag. The CML portrayed on your list was the
first version when schemas and XSLT did not exist. The current CML (see
http://www.xml-cml.org) is formulated under XML Schema and uses a rich set
of data types to define and verify the fields. An example:
<molecule>
  <atomArray
   elementType="O C N Cl C C C C..."
   atomId="a1 a2 a3 a78 a23 a34..."
   x3="1.2 3.4 4.5 5.6 6.7 7.8..."
   y3="3.4 4.5 5.6 6.7 7.8 ..."
   z3=" 4.5 8.9 1.2 3.4..."
   />
</molecule>

That represents a complete molecule, with dataTyping (through XML Schema)
and fully explicit semantics (apart from the coordinate units). It is no
more verbose than several current formats and produces SVG of
approximately the same size. Personally I find XSLT works well though it
would be useful to have microparsing in XSLT 2.0. I also really like the
ability to transform SVG to FO - this is really a killer app for chemistry
(although most chemists have a complete phobia about all sorts of XML - as
evidenced by some of the discussion). SVG is a really important tool to
convince chemists of the value of XML.

Peter

BTW It is sad but not uncommon to see uninformed and critical discussion
of XML in chemistry. We are delighted to answer emails and this is always
the best way. We have some useful tools and many supporters :-)

Peter Murray-Rust. (CML, VHG and XML-DEV)
CML http://www.xml-cml.org/
Virtual HyperGlossary http://www.vhg.org.uk/ [new website; has XML demos]
Unilever Centre for Molecular Informatics, Cambridge University, UK
pm286@cam.ac.uk