RE: [topicmaps-comment] RE: OASIS vs W3C

["Tsao, Scott" <scott.tsao@Boeing.com>]

> From: Tony.Coates@reuters.com

> [...] that technology guys like Scott and I have to sell to business people (i.e. the ones who
control the budgets), there is no such clear separation about when you might use one or the
other.  For many problems, you can see how you could use either (and believe me, we have done
this exercise).  Nobody wants to tool up and train to use two different technologies if they can
avoid it.  Nobody can afford to use every different technology in every different niche in which
it may be the "best of breed".  Often the best practical solutions are build around using a not
quite-best-of-breed technology that you already have experience with.  I see (or believe) that
you are trying to suggest that XTM is "just" a serialisation syntax for topic maps, one of many
which are possible, and so what is the big issue, it can't be harmonised itself, but some other
serialisation might provide the harmonisation.  However, those without your clarity see XTM and
topics maps as rather synonymous, which is hardly surprising.

I have been sort of "selling" TM here for the last 2 years, but everytime I had to come up reason
why not to use RDF, which came from a standards body that we have been spending big money on
(~$50K/year).

> Really, the issue that (I believe) Scott is referring to is not XTM per se, as tightly focussed
as you like to define it, but rather the problem that large enterprises don't want to deal with
their information assets as topic maps if RDF is going to become the dominant standard, and vice
versa.  If the tools support both somewhat equally, then it doesn't become an issue, but if they
don't, then it is a *big* issue.  Both RDF & topic maps have useful and desirable features, and
neither has a set of available tools which is convincingly better than the other's.  This makes
it a hard choice, which means the choice is often not being made, just being left on the
backburner.  The more topic maps and RDF can converge, so that the available tools can converge
and support both, the easier it will be to get business buy-in, because the situation won't
appear so fragmented (and hence immature).

From my knothole, there are plenty of other (more) urgent XML deployement issues for us to
handle.  We (really) have no reason to jump the gun and deploy TM and/or RDF in a big way, as
long as they are different animals.  I like the way Michel described XTM/RDF (analogically) as
XML syntax, because there should be plenty of lessons to learn from since XML syntax became a
standard.

> I haven't even mentioned DAML+OIL, but similar comments apply.

This is why I think we need to do architectural analysis one layer at a time.

Thanks,

Scott