Live Fedora start page proposal

seth vidal skvidal at fedoraproject.org
Tue Aug 28 15:00:39 UTC 2007


On Tue, 2007-08-28 at 10:53 -0400, Donald Fischer wrote:

> No ulterior motive here; getting our ducks in a row and the prototype
> page set up took longer than we anticipated but I wanted to push to
> make the deadline to try it in Test2 rather than delaying to Test3.

I wasn't suggesting an ulterior motive. I was more suggesting that last
minute changes like this have gotten us a lot of bad flack before. You
may not remember the 'tracking image' fiasco but it was not much fun.


> Not done yet, but there is minimal text to be translated on the start
> page itself and the third party search services that we'd build on are
> widely internationalized.  We'd use the regular Fedora documentation
> i18n workflow for the necessary translations, I suppose.

okay - not sure how this fits in with the translation freeze, though.

> Red Hat will pay for the incremental server and bandwidth so that it's
> not an additional cost burden competing with existing Fedora project
> expenses.  Open to discussing the governance model the server itself,
> presumably we could handle it like other Fedora infrastructure
> services.  At present it's just a static web page and a few images; in
> the future we anticipate also having a hosted results page that would
> be a simple JSP.

If we handle it like the other fedora infrastructure services then JSPs
are a non-starter. Especially considering we do not, yet, have a free
java on our server installs. Moreover, the consensus of the
infrastructure group has been repeatedly, "no" to all things java.
Mainly b/c of the non-free java but also b/c of the tumbling morass that
is all things java on websites.


> I hear you, definitely think we could look at doing various things
> here and the online desktop team has the resources to do so.  When we
> tried out some workarounds for this issue, the general feeling was
> that we might be trying to be too smart for our own good (e.g. how
> long should you wait before you decide that the user is offline vs.
> has a slow connection?)  Fielding in a test release sooner than later
> would give us some real data about what this issue actually looks like
> in the field.

But it won't give us any real data about our offline users. They don't
download test releases b/c of the limited bandwidth they have.

-sv





More information about the advisory-board mailing list