Re: Misclassification of packages; "libs" and "doc" sections
Thomas Hood wrote:
>
> I think it would be a bad idea to conflate different "modes"
> of classification together in one graph. There are two
> issues here that have to be kept separate.
>
> (1) The first issue is that there are different modes of
> classification: viz., classification by function, technology,
> license, etc. (See below.)
>
Agreed. This is exactly the issue that there exist
categories of categories, right? In a sense, there are
2nd order predicates in our model. :) In logic terms. Anyway,
I see your point.
> (2) The second issue is that for one kind of classification,
> a package may fall into more than one class. If "editor"
> and "newsreader" are two different function classes then some
> programs with be one, some the other, some neither and
> some both. This is why it would be appropriate to *list* the
> functions of the contents of a certain package. If the items
> listed are defined by their location on a graph then so
> much the better.
>
We already have consensus on this. Almost everybody argued
that there are packages that fall into multiple categories.
> However I do not immediately think that it's a good idea
> to have one big graph embracing all the different modes of
> classification. You have editors, music players, music
> editors that play, and programs that do neither. Okay;
> that's classification by function. But do we want to define
> "nodes" for GPL editors, BSD music players with web interface,
> GPL-BSD-combo music editor players for science, etc., etc.?
> I don't think so. Classification modes are (as I'm using
> the term "mode" here) orthogonal to each other and should
> not be conflated. Perhaps I'm misunderstanding the proposal.
No, you've definitely identified a weakness in my proposal.
Nevertheless, what classification "modes" are right? This is
a point that goes beyond an intuitive organization. That's
why we'll have to delve into some ontology papers! What has
AI people done in such circumstances? A few lessons to learn
I'm sure. I think the papers I referred to will be good starting
points (somewhere in this thread).
I'm not responding fully to your inquiry here, I assume I'm
allowed to defer the discussion until I can. :) Perhaps as a
reply to your other post.
Thanks,
--
Eray (exa) Ozkural
Comp. Sci. Dept., Bilkent University, Ankara
e-mail: erayo@cs.bilkent.edu.tr
www: http://www.cs.bilkent.edu.tr/~erayo
Reply to: