Giovanni Gentili wrote:
James Graham:
The issue when trying to abstract problems is that you can end up doing
"architecture astronautics"; you concentrate on making generic ways to build
solutions to weakly constrained problems without any attention to the
details of those problems that make them unique.
I think the right level, like in my proposal,
is greatly under "astronautics"
but no so low as "single vocabularies".
I rather disagree. How we interact with information depends
fundamentally on the type of information. If the information is a set of
geographical coordinates, for example, the set of useful interactions
are rather different to those for a bibliographic entry. Trying to
pretend that the two problems are just interchangeable instances of the
same "semantically structured information" problem is likely to hide the
important distinctions between the two problem domains.
