[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Permanent page reference for packages



On 9/25/06, Daniel P. Berrange <berrange redhat com> wrote:
And link to the various other resources related to the package, eg the upstream
website / download URL, the Bugzilla component, the appropriate path in CVS web
and Fedora maintainer's name & email address.

Where exactly do you pull that other resource information from without
a lot of manual love from a human or dolphin page content editor?

A single overview page can not rely on summarizing package header
information across all available versions of the package, because the
information for any header field beyond the package name maybe
different from release to release.  You are going to quickly spiral
into a death spiral of complexity and miscommunication if you try to
make room in your automated script to summarize resource information
that cannot be garunteed to be the same for all releases of the same
package.

For example, lets talk about Bugzilla component.. if a package moves
from Core or Extras between a two fedora releases.. doesn't it's
bugzilla component change?  Doesn't its maintainer change?

I simply don't see how this summary information can be auto-created by
a repoview script and be accurate for all releases.  I think the best
you can do for an auto-generated summary page is to create a simple
table for each package that leads to release specific pages with
resource details, as previously described.  If you try to include all
release specific info for all releases, it will just be a big gross
messy table of conflicting information that will confuse people.

And quite honestly, the most useful information a package specific
page could provide is exactly the information that is release specific
which is not encoded in the package headers and does not represent a
re-telling of systematic Fedora infrastructure information.

The best information this page can offer to users is maintainer or
developer experience based knowledge per release which would require
active editting.  Generating yet another web interface to be a link
farm to fedora's bugzilla or to fedora's cvs is just wasteful
interface re-engineering and will make it more difficult in the future
to 'fix' such 'deep links' if the infrastructure tools such as the cvs
location or the bug tracking system are updated as Fedora evolves.

-jef"For you very clever readers who want to argue semantics, the
package name is by definition the same for every release encoded in
the page summary as described in this thread since you are creating
the webpage based on the package name unless you go all super-crazy
and also want to try to keep up with obsolete relationships on the
summary page to."spaleta


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]