On Sat, Mar 11, 2000 at 04:15:32AM -0700, Bruce Sass wrote:
> Hello Jay,
I'm not the only webmaster, but I'll answer :)
> Snooping around the Python site I came across
> www.python.org/download/download_linux.html, and thought it would be
> nice to have a link to the python package pages.
>
> I soon realized that it would be a maintenance headache for a webmaster
> to keep up with the releases and package name changes, which led to two
> questions:
>
> Does Debian have a policy about getting linked to?
>
Yes. It's whatever I like. :)
Actually, we add links to things that would be of general interest to a
Debian user (especially newbies) and stable (in the sense that the link
will not disappear). The information should also be of a general nature.
While I firmly believe that the web's greatest purpose is in distributing
information, you must be careful not to put too much on a page or the
reader won't be able to find anything. In fact, The site has been reorganized
twice since I've been involved with it, mostly to make information
more accesable.
While it would be nice if the python page gave better instructions, i.e.
apt-get install task-python (but only starting with potato), a user shouldn't
have to go to our related links page. By this, I mean that the tools on the
machine should provide enough information for a user to figure out what
packages they are interested in and how to intall them. I'm confident that
there are a lot of Debian users out there who have never used apt-cache and
apt-get. We need to both provide better information (more does not
necessarily equal more) for newbies and a better front end for apt.
> What would it take to implement a (uhm) virtual or task package tree?
>
They can be generated reasonably easily from the Packages file.
This is information that is be available on the local machine.
('apt-cache search task-' accomplishes the second task). The
list of virtual packages could easily be printed by apt-cache, but
for what purpose? The list, by itself, is of little use. People
only need information about virtual packages when they are looking
for a list of packages to satisfy a dependency. apt-cache already
does that using 'apt-cache depends <package_name>'. As already
stated, we need a better front end for apt(*).
Sorry for the long winded answer. As with most developers, I have a lot
of opinions on the way things should be done. :) Also, this may
prompt discussion on debian-www about improvements to the site.
(*)I haven't forgotten the front ends being worked on. It's just that
they aren't finished yet.
--
James (Jay) Treacy
[EMAIL PROTECTED]
