[Debtags-devel] Debconf5 Debtags paper
Benjamin Mesing
bensmail@gmx.net
Sat, 28 May 2005 17:45:41 +0200
Hello Enrico
Cool, the article is great! You have managed to make it pleasent to read
and keep it informativ. I also strongly praise, that finally there is a
comprehensive introduction which can be handed to people interested in
Debtags.
Here are some issues were I stumbled (use your search function to find
the line numbers :-):
In the abstract:
[sections] which are not useful anymore due to
the large and always increasing size of the Debian archive.
Shouldn't you say that a little "softer"? They are by no means _not_
useful, but they are no longer sufficient to be a good aid when
searching.
In order for people to share the knowledge encoded in facets and
tags, we need
to collect them into a shared and commonly understood
vocabulary.
Besides the usefulness in the tagging work, the vocabulary can
also be used to
present a list of facets and tags in a user interface, to store
meta-information about facets and tags (such as a description),
and to validate
tag data.
I find this first (short) section about the vocabulary hard to
understand. I am really not sure what you are saying. Perhaps you should
hint to the following section which explains in more detail. Such a
hint, might ease people who struggle with this section (like me).
``autodebtags``
Experimental tool to perform some automatic and euristic
tagging tasks.
Shouldn't it be heuristic here?
Change
Uitoolit and not implemented-in
to Uitoolkit :-)
Perhaps you should mention the carfully-tagged tag when speaking about
how people can edit the tags?
Thanks for mentioning the AI-Tagger and packagesearch. Would you mind me
sending you a nifty KDE screenshot instead of this ...umm... ugly
motif(?) one. Also should I select some tags for the screenshot (perhaps
the ones you mentioned in your paper)? This would make it more
expressive.
So far my thoughts
Greetings Ben