On 11/3/99 at 10:44 PM, ehood(_at_)hydra(_dot_)acs(_dot_)uci(_dot_)edu (Earl
I think this is a limitation when using reverse ordering with
multi-page indexes. I've advocated in the past that reverse ordering
is not needed for multi-page since getting to the last page is easy (
and reverse thread indexing has been continually bothersome in the
I tend to think that the most current information should load first, and
be at the top of the page. This is a separate issue from the peculiar
rudeness of asking people to download a 500K+ HTML document (are you
listening, omni development?)
Small pages with new material first are important to people who are
using the archives as a primary source and looking for very recent
material, as opposed to people who see them as irregularly-checked
backups to their mail clients.
Even with multipage enabled, many of my date indexes are 80K -- that's a
whole lot to ask someone with a modem to download just to find out if
there's anything interesting and new. I'm disinclined to lower the
separation threshold (currently it's 300), because some of them already
split as far as six or eight times.
Thanks for your time; I'll keep my eyes open while poking through the
codebase. Mostly, I'd like to know why the failure is so unpredictable.