[sc34wg3] Analysis of TMRM Use Cases

[Martin Bryan]

Michel hit the nail on the head when he said:

> There is a tension that I know quite well (I am there too)
> between users and implementers.

Wherever there is such a tension it is vital that we ere to the side of the
under-represented user. Unless we use concepts that users can easily grasp
and, more importantly, fit into their existing world view, then any new
concepts are detremental to the ongoing development of the topic map market
that will ensure the sucess of the implementors.

> I consider the topic map standard
> to be useful as purely declarative.

Another vital sentence. Unless users can see from the declarative markup
*exactly* how TM fits into their world view then they are not going to use
it. (This is the major problem I have with XTM - it does not fit my
worldview *exactly*. RM is even worse in this respect!)

But Michel, like many others, is plainly wrong when he states:

> Topic Maps users want their information to be connectable
> with others. Otherwise why would they be doing topic maps for?

There are many, many potential applications of Topic Maps, as Jim has stated
oft enough, where the purpose of applying Topic Maps is strictly not to
allow "information to be connectable with others" - at least if by "others"
you mean "other topic maps". Topic maps allow users to interconnect *their*
information resources, whether they be stored in files, databases, RDF
metadata sets, paper records or, God forbid, in the minds of the human
resources within an organization. They are a tool for knowledge management,
not necessarily knowledge dissemintation. Some information you do not want
to disseminate, as Jim's use case so adequately demonstratates, but as sure
as hell you need to ensure it is properly, adequately and timely managed.
Tell me how many drug companies you know are using topic maps to connect to
topic maps produced by their rivals. The number can be counted on the
fingers of an armless man. Let's not get carried away by the supposed need
of users to merge data. Commercial users need to be able to describe
relationships between information occurrences, not topic maps. Unless we can
describe these simply, adequately and in a way that fits users preceptions
of the relationships between their data resources we will have failed.
Fortunately, so far you have not left the users behind, and have been
successful. Please keep us in mind while you argue over the techincal nitty
gritty.

Arch-fiend (sorry "user") Martin Bryan