RFC: Keywords instead of Section
I am very much in favor of replacing "Section:" with
a more useful classification system.
However, just replacing our current single-keyword system
with a multiple-keyword system isn't the best solution.
One problem is that keywords can be ambiguous. Keyword
'X' can mean either that the package is part of an X server,
or that it is a program that uses X as a user interface.
But this isn't the most serious problem. The most serious
problem is that classification systems have a tendency to
become ontologically disorderly. (I won't get into what
that means here.)
A good solution is explicitly to _type_ the keywords. Dividing
keywords up into groups of the same type is very a useful way
of keeping the system orderly.
Helpfully, Erich Schubert lists his suggested keywords by
type. He lists: user interface, programming language,
user app vs daemon vs web app, "group", "others"
In my opinion, the types should not be omitted from
the system but integrated into it. (However, we must absolutely
not have a type called 'other' !!!)
SourceForge and Freshmeat do this. Their keyword types are:
* Development Status
* Environment (i.e., UI)
* Intended Audience
* Operating System
* Programming Language
I would suggest that instead of re-inventing the wheel, we
simply adapt the SourceForge/Freshmeat system for our own use.
The simplest thing to do is replace "Section:" with several tags
Environment: X, console
Programming-Language: C, Python
Topic: cooking, sex
allowing any number of keywords to be given on each line.
An alternative is to do it this way:
Class: Development-Status:stable, Environment:X,
Environment:console, Intended-Audience:user, License:GPL,
Programming-Language:C, Programming-Language:Python, Topic:cooking,