Lists Home |
Date Index |
I have suddenly realised that XML-DEV is 5 years old - it could be useful
and fun to review some of what we have achieved and perhaps speculate on
the future. I make a concrete suggestion at the end...
For those who don't know, XML-DEV was set up by Henry Rzepa and myself and
we ran it for ca 3 years before it moved to the current site. It has never
been formally moderated though I put a fair amount of time into "steering"
in the first years. Its fundamental character hasn't changed - it is a
list for anyone, with a friendly welcome and no completely set agenda. It
can be revolutionary, argumentative, cooperative, but we have completely
avoided flame-wars despite many differences.
In the early days it had a fairly public agenda - "to try to make XML
succeed". That may sound strange now, but XML didn't even rate the W3C
front page when it started. The initial emphasis was on creating high
quality, open technology that could be used to prove the language,
stimulate applications, and spread the word. Five years ago the Open ethos
was much weaker than now and one of the many contributions of the list has
been to encourage commercial organisations to make their tools available.
XML succeeded, and in ways that weren't expected - at least not by many.
Originally it was conceived as a document-oriented technology for robust
quality publishing of documents over networks. the original workplan had
three pillars - XML syntax, XML link, and XML stylesheets. Schemas were not
high on the agenda and XML was not seen as an infrastructure for middleware
or glueware. It was expected that at some stage it would be necessary to
manage data but there was little activity in this area in 1997. When
developing Chemical Markup Language (which must be one of the first
published XML applications), I found the lack of datatypes very frustrating!
Well, XML is now a basic infrastructure of much modern information. I doubt
that anyone now designs a protocol, or operating system without including
XML. Although this list sometimes complains that XML isn't as clean as we
would like, it works, and it works pretty well.
But we are up against the hard problems. Five years ago there was very
little desire for one organisation to communicate effectively with another,
especially if they hadn't a prior business agreement. XML is now forcing
organisations to point outwards - even if you don't want to use it, your
clients, customers, collaborators, competitors do. There is a growing
assumption that effective information interchange is straightforward and
I don't think it's easy. Many of the problems are not technical but
cultural. Most of the world has had no experience at devising interoperable
systems and those that have know it's a long and tedious business. Henry
and I are trying to do that in chemistry. It's like many other domains,
full of fuzzy human-only semantics and proprietary systems. The first
problem is extremely hard in itself - to represent a domain *in
machine-processable form*. Chemistry is probably as tough as they come and
so our progress isn't rapid. That is balanced by the relatively slow uptake
of XML in the domain (at least that gives us time!). Chemistry isn't unique
- I do a lot of XML training and many sectors are not yet advanced in XML.
And I think the pace of technical development may start to leave many
people behind. A year ago I had to decide whether to convert CML to XML
Schema. I found the spec very tough going and the tools were difficult to
run. So I decided not to. Then I discovered Roger Costello's **EXCELLENT**
XML Schema tutorials, exercises and best practices and I changed my mind
[Roger pays tribute to XML-DEV for his material.]
I have since converted CML to schema. It's been tough - certainly not
automatic - but during the process I found that the Schema technology
forced me to make the design better. An obvious benefit was modularization
- CML originally subsumed a generic scientific language (alias TecML or XML
(eXperimental Markup Language (sic)). I have now recast this as a separate
language STM-ML and will announce it here shortly.
I've occasionally hinted that I and Henry had a hidden agenda in setting up
XML-DEV. Why should two chemists spend so much time on something
non-chemical? The h.a. was to help XML succeed to the point where it then
became obvious to the chemical community that they would have to adopt it.
We're just about there - last week we had a meeting called by the
international body (Int Union. Pure. Appl. Chem) to discuss XML in
chemistry. So we no longer have to convince people that XML is the way to
go. I sense this has now happened in most sectors, even if the details are
So greetings to all, especially if you are relatively new to the list and
thanks to the OASIS staff for running XML-DEV.
You are invited to contribute thoughtful reviews (to XML-DEV) about XML
that list members may find helpful. If this idea works, it might form a
useful collection - if not, don't worry. Your review should have a title
which would form a separate thread, but could be prefixed by XML-DEV5 (e.g.
"XML-DEV5: On first looking into XML Schemas"
Perhaps it would be possible to avoid loss of coherence if - in the first
instance - a review itself avoids REplying to other postings so that the
threads don't get impossible to follow. REplies should relate to the review
and should be written in the spirit that they are useful and enduring
If anyone has any additional thoughts on the process, REply to this thread!
Reviews should be in ASCII (no embedded markup or binary). NOTHING should
be sent as an attachment! There is an obvious difficulty for very large or
non-textual material so - if that is necessary - it should be on web pages.
Peter Murray-Rust. (CML, VHG and XML-DEV)
Virtual HyperGlossary http://www.vhg.org.uk/ [new website; has XML demos]
Unilever Centre for Molecular Informatics, Cambridge University, UK