RFC: fix summary text for lots of packages

[Jeff Spaleta]

On Fri, Nov 21, 2008 at 12:37 PM, James Antill <james at fedoraproject.org> wrote:
>  Now sure, I think there are probably some packages that have less than
> optimal summaries ... but I'm not sure length is the most significant
> factor.

I'm coming at this from the point of view that there are different
pressures on the content that needs to be summaries and we aren't
going to find any policy that fits all the packages, or even a
majority of packages.  I don't think we are going to come to a rough
consencous any more concrete than "be aware of summary length"  We can
nitpick individual package summary language to death on a package by
package basis.

Putting a cost on length provides a downward pressure on summary
content without telling any particular packager what they should or
should not provide.

> For instance the yum package summary could probably be better,
> but I'd bet that'll be by _adding_ characters.

So be it.... different pressure balances for different packages.  How
important that the yum package summary really in the scope of end user
oriented UI? Is it ever gonna bubble up in front of a user as part of
a software install selection process?  For something like music player
applications however..where we have multiple peer choices.. multiple
brands on the shelf...position on the shelf can matter..and so can
summary optimization.

We are NOT going to decide on optimal summaries for all packages in a
big group think.  Not going to happen. But if summary length is
important for end-user UI, if "short" really does matter, then the UI
should cost the length and packages which really do matter in the UI
should incorporate that costing as part of jockeying for shelf
position.

>  The fact that a single tool decided that summaries should be used
> instead of names, and so summaries should be roughly the same size of
> names shouldn't make Fedora packages break their summaries for other
> tools ... all IMO.

Who said anything about break their summaries?  I'm talking about
explicitly exposing length of information text as a cost metric. Just
like real world brick and mortar advertising does.  if all 80
characters are of equal cost...there's very little downward pressure
to act as an incentive to make things consice.  Think of it as
character cost as an analog to the dirty socialist hippie notion of a
carbon credits.

-jef