ports/www is too full

[Tillman Hodgson]

On Wed, Oct 27, 2004 at 04:12:50AM +0200, Benjamin Lutz wrote:
> > Nice tool. But how does that help with quick window-shopping?
> 
> http://www.maxlor.com/freebsd/files/portsbrowse.py
> 
> Usage is simple: "./portsbrowse.py www" will display all www ports. 
> "./portsbrowse.py www net" will display both www and net ports (logical 
> OR), while "./portsbrowse.py -and www linux" will display only ports that 
> are in both categories (logical AND).
> 
> How does this work for you, as far as "window-shopping" is concerned? 
> Sure, you don't have autocomplete, but you do get to grep the scripts 
> output instead.

I haven't looked at the script, but I suspect that you might have missed
the point of what I was getting at.

Let's say I'm looking for a Apache modules. I'm not looking for anything
in particular, I just want to see what's been ported. Window-shopping.

How does looking at the www directory at the shell, via a python script,
or via a KDE GUI help me?

There's 792 ports in the www tree according to `ls | wc -l`. Of those,
113 are _possibly_ Apache modules (as determined by `ls -d *mod* | wc
-l`). Which of those are really Apache modules, and of those, which are
Apache 1.3.x modules is impossible to easily tell from the output.
`grep`ing for "mod" in the output of some utility would have the same
problem. The fallacy I've fallen into with this example, and the fallacy
that searching tools fall into, is the idea that port names are always
going to be representative of what the port contains (or at least that
the port comment will magically have the right keywords). That's not the
same thing as meta-information like fined-grained categories.

Perhaps virtual categories can do what I'm thinking of, if they were
widely used.

-T


-- 
[Re : quantum physics]
I can't say I care one way or another. -- Kai Henningsen
That's just because nobody's measured you yet. -- Christian Bauernfeind