xml-dev - Re: OASIS/SAX - looking around

[Peter Murray-Rust <peter@ursus.demon.co.uk>]

At 04:55 PM 2/11/00 -0500, Simon St.Laurent wrote:
>At 01:09 PM 2/11/00 -0800, Jon Bosak wrote:
>>Despite its success so far, the process you've got now for SAX is
>>not democratic.  It puts all decision-making authority in the
>>hands of a single person.  The fact that the person in question
>>happens to be above reproach does not change the fact that this is
>>a benevolent dictatorship.  I don't buy the benevolent dictator
>>model of standards development no matter how highly I regard the
>>dictator of the moment.

Although I have probably confused this issue already, here are some more
thoughts...

XML-DEV is a "bottom-up" process and must remain so. It must remain
independent of influence from any organisations. I won't repeat *why* Henry
and I offered the list's home to OASIS [the reasons are in the XML-Deviant
interview] but emphasise that it was in part due to being approached by
Jon. [H and I have had approaches from for-profit organisations which we
have declined.] I have admired the way that Jon has taken XML V1.0 through
the potential quagmire of committees and vested interests. I am convinced
that he - and fellow OASIS members whom I have met - wish to preserve the
essence of XML-DEV as part of the development of "XML". If there were ever
signs that XML-DEV sat unhappily in the OASIS fold then it would move
elsewhere - in spirit if not necessarily in name. XML-DEV is a successful
meme and can mutate if required to defend itself! [And that is unstoppable
- regardless of what I, Henry, Jon or OASIS. OASIS' role is to maintain a
stable *technical* infrastructure for the list which Henry and I cannot.
[Our only choices were: hand it over to some *organisation*, or set up
"xml-dev.org" and try to raise support.]


One of the several people who wrote to thank Henry and me compared XML-DEV
to a "mini-IETF" and this appeals to me. The primary difference is that
IETF has a real-life existence and XML-DEV does not (Henry and I are
actually humans, but it ends there). The particular arrangements at
Imperial could not have continued indefinitely. The IETF seems to us to be
a valuable way of getting "bottom-up" approaches accepted by the community
and turned into "standards" that can be accepted by those who worry about
them.

One of my main activities is trying to get XML accepted as a standard for
the submission of new drugs. [I am evangelising XML next week in
Philadelphia at the Drug Information Association]. Authorities such as the
Food and Drug Administration need to "validate" the submission systems used
by pharma companies and so, if XML is to be used, it will have to be part
of that process. Although any home-grown system can be validated (and
unfortunately almost all systems *are* currently partly home-grown) it
helps to be able to point to "standards" and "standards organisations" in
the process. [There are certainly some companies who are extremely loath to
accept anything less than ISO or equivalent, so a "Recommendation" is of
lesser weight.] Paradoxically, in IT systems, manufacturer dominance also
acts as a "standard" in some domains. Last year I went to a major pharma
[no names...]. "What software do you use?" "No Java, No C++, No XML...",
"What *do* you use?" "XYZZY, "PLUGH" and "FOO". [3 large companies]. So,
"authority" is critical. Incidentally at one of my evangelical sessions
last year, the response to XML was "we have decided on PDF as the document
format for the next 5 years[sic] and we may move to XML after that...".
There are signs that this is becoming untenable and I am more optimistic
about next week. So XML needs not only innovation, but marketing in some
areas.

How, then, does XML-DEV act as a "bottom-up" process that can be taken
seriously by cautious "late adopters"? Henry and I face the same challenge
with Chemical Markup Language. Our strategy there is to get the early
adopters to tackle the bottom-up approach by building OpenSource systems
that show it works. We then get the more innovative companies to be
medium-adopters and show there is *commercial viability". [SAX has reached
the medium-adopter stage and I am *publicly extremely grateful to all the
IT vendors who have adopted SAX*. Without them it could have been very
difficult to move further.] However we are now moving to late-adopters.
Some of them may, by default, be happy because SAX is a component of any
system they buy, but others might wish to be reassured that it has been
implemented *consistently* in any product. That is why I suggest
conformance tests. [Example: In running a test file on parser-SAX demos
from various implementers, a newline is sometimes output as "\n" and
sometimes as "
". While I believe that this is simply the demo wrapper
that has been used, a customer might query whether the systems were
compatible.]

The challenge, therefore is how we support the unfettered innovation on
XML-DEV, move the results to proof-of-concept, early-adoption and then
migrate it by some means to a certifiable protocol. I am sure there are
successful models that we can borrow from - probably better than what I
have suggested.

On a related issue it is clear that the unwritten "constitution" of XML-DEV
needs to be preserved and abstracted. These discussions are extremely
valuable and act as a historical record to guide us in the future if we
ever lose track. I have been very grateful to Leigh Dodds for giving me the
chance to recall the ideals and history of the list, and he has been doing
an extremely important job of capturing our history *as it happens*. If
this is not done, e-history often decays. Leigh, Henry and I have exchanged
e-mails and I strongly encourage his efforts in this area. If he and other
volunteers can pull together key points from the current "political"
discussions [which have been of very high quality] - I think many of us
would be extremely grateful. Otherwise in 6 months time they get lost in
the sheer volume of list traffic!

	P.