On Wed, 15 Apr 1998 10:31:59 -0700
Earl Hood<ehood(_at_)hydra(_dot_)acs(_dot_)uci(_dot_)edu> wrote:
On April, 15 1998 at 9:27, J C Lawrence wrote:
If this is referring to the number of messages per page, I
disagree.
Yep, its the number of messages per page on the index pages.
The current number of messages per page is plenty for someone
with a > 28.8 connections.
<shrug> I'll note that I'm offering my MHonArc archives via a 28.8K
dedicated dialup connection. Browsing them here from work is quite
usable.
Depends on patience. Where I work, we are on a T1, and I would not
be happy having to download a 2000 message index.
<nod>
Its definitely a subjective point. I'm on multiple T3's here at SGI
on a 100MB LAN, and the indexes are on a 28K dialup connection (ie the
bottleneck is at the server side), and while its definitely not *fast*
to load a full index, the extra value for me is that I can get a broad
overview of the threads as well as have comparitively few threads
break across index boundaries.
That is roughly a 100K (or more depending on resource settings).
Yup.
Note, on a usability stand-point, smaller pages are generally
better. It is easier for the reader to absorb and navigate. Large
pages can be intimidating/overwhelming and discourage attentiveness
from the reader.
Given the sheer volume of indexed messages (we'll probably pop over
200,000 by next month), there's actually very little index-by-index
browsing. Instead people tend to search for their topic of interest
and then just follow that thread. (Thus also the big demand for
search supports as mentioned in another thread here).
I should note that larger index pages (in multipage indexes) may
cause a little slow-down when MHonArc performs additions to the
archive.
Yeah. I'm not using multi-part indexes for simpler reasons. It was
just easier to seperate the messages out into MH-style folders by
calender range, and then generate a single-page indexes for each
folder.
MHonArc will only update the pages that need to be updated (which is
usually the last page of the archive). Updating a page with 100
messages is faster than one with 2000 messages.
Yup -- but that's what midnight cron jobs are for.
--
J C Lawrence Internet: claw(_at_)null(_dot_)net
(Contractor) Internet: coder(_at_)ibm(_dot_)net
---------(*) Internet:
claw(_at_)under(_dot_)engr(_dot_)sgi(_dot_)com
...Honourary Member of Clan McFud -- Teamer's Avenging Monolith...
