Fedora Project, give me 20 Million Euros or Free EDA software

[Matthew Woehlke]

Nicolas Mailhot wrote:
> Le mercredi 04 février 2009 à 17:28 -0600, Matthew Woehlke a écrit :
>> Nicolas Mailhot wrote:
>>> Sure, take the time to make the experiment, remove fonts, remove music,
>>> remove themes, remove images, and see how much stuff is still working or
>>> useful in your nice "software only" repository.
>> First off, I'm not talking about "core material", I'm talking about 
>> things that merely enhance, much of which is currently /not/ in Fedora 
>> repos.
> 
> Who are you to judge what the "merely" point is?
> Who are you to judge that because something does not fall 100% in your
> etricated "software" category it's less useful for Fedora than a random
> sf.net bit of software two people in the world use (the author and the
> packager)?

Oh, give me a break. Are you really going to argue that packaging 
Holst's "The Planets" suite is as worthwhile as packaging inchi*? 
There's a big difference between software that is best built for Fedora 
and best placed in the file system by a package manager, and a 
self-contained .ogg.

(* which, incidentally, violates the new trademark policy)

>>  not only 
>> because I don't think it's an appropriate goal, but because if we're 
>> doing this in Fedora and not in a more general sense, we're keeping 
>> non-Fedora users out.
> 
> I think we have enough Fedora problems without spending a lot of time on
> non*-Fedora-users.

I'm sorry, I must have missed when we became Apple. Ok, let's do 
everything in a way that only benefits Fedora users and excludes anyone 
else.

Since when is that policy? I thought Fedora preferred working with 
upstream. That's what I'm trying to suggest; you're talking about 
packaging things that would in no way be Fedora specific. It would be 
better to create a distro-agnostic repository that Fedora integrates 
well with.

>> "I find that kde-look.org enhances kdebase. We should package 
>> kde-look.org (at least, all appropriately licensed content) and add it 
>> to Fedora's repos."
>>
>> Do you seriously agree with the above?
> 
> If there are people willing to maintain the resulting packages, yes.

That time would be /much/ better spent writing tools to integrate with 
upstream.

> Let's also remove perl packages since you can use them directly via
> "more appropriate" CPAN (and a lot of people do it),

Do (non-pure-perl) Fedora packages depend on these? I think "yes", ergo 
this wouldn't work.

There is also the file management issue. Pure content tends to be 
self-contained and can be placed anywhere. Once you get into things that 
need to live in a certain place in the file system to be useful (without 
unreasonable jumping through hoops), then having such things managed by 
the system package manager is useful. This doesn't apply to video, 
music, pictures, etc (unless integrated with some software, in which 
case, go ahead and package it).

That said, if CPAN and yum played nicely together (i.e. depsolving 
works, tracking installed files works without having to cross-reference 
multiple tools)... then we probably /should/ consider using upstream 
CPAN directly. (I'm pretty sure this is not the case, though.)

> This is something which is going to greatly enhance Fedora, yay!

There is a gaping hole in your argument. You're talking about removing 
things that already exist before an adequate replacement is in place. 
I'm talking about not misdirecting effort into a sub-optimal solution.

>>  Does "A great looking gothic font" give 
>> you any useful information if this is that perfect font you need for 
>> your presentation/art project? 
> 
> It gives enough information for Japanese users that know what a gothic
> style is.

(Hypothetical situation)
There are five of them, all with descriptions similar enough that you 
can't tell anything about what distinguishes each. How do you decide 
which you want short of installing /all/ of them?

We don't have this problem right now because there isn't such choice. 
Try running a font repository with several thousand fonts, with no 
previews, and see if people don't complain.

Different media types need different metadata to be effectively browsable.

Do one thing. Do it well, and do it in a way that it connects well with 
other tools.

-- 
Matthew
Please do not quote my e-mail address unobfuscated in message bodies.
-- 
ELANG: input is in wrong language (please use English on English lists)