i'm working on a federated searching tool (yes, again), but this time
limiting myself to z39.50 targets, but with intention to include xml
gateways (whenever jstor makes theirs public)....

i'm using php/yaz/dom/xslt, postgresql, and xmlhttprequest objects to
perform these in a threaded/non-blocking manner.....

but i could use some extra input on a coupla things, and my main
"bounce-off guy" is on vacation currently....so, here goes:

    1) the initial search pulls a small number of the total hits, and
    once those have all returned (or timed out), the user is redirected
    to their first result list....

    how do you think i should continue populating their results?  should
    the initial results page use more xmlhttprequests to pull in a
    greater number that match their criteria? if so, what happens when
    they leave the page, etc, before those have been returned?

    yes, i bet doing this in java would solve some problems, but i'd
    prefer to keep this in php for the moment....

    2) ...uh...forgot number two, but if i remember, i'll ask...

anyway, here's a preliminary url:

    http://trash.lib.muohio.edu/federated/

again, this is very beta....getting the fundamentals figured out before
i add any "features"...it currently searches academic search premier and
the ohiolink central catalog....

only 10 results from each database are stored, and i'm currently showing
all 20, sorted by date....again, just working out kinks, and trying to
figure out some of the process involved in populating the results
list....

as for browser compatibility, only really tested on mozilla*...i'm
willing to bet my xmlhttprequest stuff breaks in every other browser,
but again, not too worried about that right now....

sorry so wordy, and thanks in advance for any ideas/suggestions,
rc