Lists Home |
Date Index |
2/13/2002 8:04:25 PM, Micah Dubinko <MDubinko@cardiff.com> wrote:
>1. Define a layered system, where the simple cases really are simple
Vote with your feet. Don't use stuff that isn't layered, promote
that which is even if it's not an official "standard". SAX isn't
a "standard" ... but it is simple, modular, and interoperable.
Other informal "organizations" are defining such things in the
absence of official standards or when the the official standard
doesn't meet the needs of a significant group of people. For
example, XMLDB.org fills a signficant gap in the area of
XML database interoperability, the soapbuilders mailing list
has worked on SOAP interoperability, RDDL was worked out on
this list, etc.
>2. Defend the complexity they introduce; explain the phenomenal benefits
>that more than offset the complexity
Almost all standards bodies have public comment lists of some sort. The
W3C process pretty much mandates that every public comment be responded
to before a spec can become a Recommendation. Take advantage of it;
they may not take out you what you don't like but they HAVE to explain
their reasoning if you complain. Or propose a simpler alternative.
And do it early in the process, if possible ... a common lament is
"we put our Requirements document out for public comment two years
ago, nobody complained; we spent immense effort figuring out a
solution to that requirement, and we don't want to throw it away."
>3. Work with vendors to produce useful tools for "hooking stuff together"
>with XML (including "low-tech" methods like XMLchucker or XMLbaton)
Do it yourself, or start an open source group, or collaborate with like-minded
people who are already doing it. Or patronize the little companies
that do it but you have to go looking for.