Refactoring the vocabulary

Erich Schubert erich.schubert at gmail.com
Thu Mar 25 14:40:01 UTC 2010


Hi all,
When we started DebTags, we were considering doing a full tree or even
graph for tags.
You could easily model this using RDF, I guess.

However, a key driver for DebTags has been the KISS principle: if we
don't make both editing and usage as easy as possible, we won't get
anywhere in getting to a critical mass of data.
We can however add this on later, easily.

It should be obvious that doing such a reorganization is just a matter
of renaming tags. We could have just used numbers, technicall anyway.
So there isn't a technical reason for doing such a renaming.

What remains is largely an UI issue:
- will the user find the "second level" facets easier when they are
organized differently?
- will the UIs work better then?

I'd suggest adding some local mapping to get the desired changes,
while leaving the actual data as is for now.
Then building an UI using the new mapping. If that is clearly more
useful, I'm all in favor of doing the change in the database.

On the other hand, it might even make sense to have different UIs use
different mappings. Like one being the tech users UI, that offers all
the "implemented in", "uitoolkit" and similar facets, whereas the
"user" view uses suite, use, ... as top levels.

best regards,
Erich Schubert



More information about the Debtags-devel mailing list