[Disclaimer: although I describe code running under Unix below, I'm assuming that the same principle could be implemented on the Mac by someone (with more experience than I) using a Mac http server. ] At 3:56 PM 1/30/96, eric m. mosher wrote: >... >when they are done with a couple (it takes a while to do one problem) >they usually log out and then come back >maybe the next day, and log in again to complete the problem >set. > >My boss would like it that when >the student logs in the secon, and third and fourth time etc. >that they do not have to scroll through this long html. The way I've handled this kind of thing (albeit on a Unix box; sorry...) is to have the form generated "on the fly" by a perl script that can use previously-stored data to customize the form. It could be done something like this: The student begins with a tiny form that just asks for identification (name, SSN, student ID#, ... whatever). The handler for that form first verifies the student as being in the class, then scans a "results" file corresponding to the exam/exercise set for any previous entries by the student. If none are found, the entire exam/exercise form is sent back. If entries have been made by the student, the handler script matches the items in the results file with a template of the entire form, and sends back a form containing only those items from the template which DO NOT appear in the results file. With only a little effort, one can add a button to the generated form which requests that the entire exam/exercise be sent back (of course, with any previously submitted results from that student filled in). >the entire program is in perl. But, of course! ;-) >current server is httpd on unix system >or http 2.0 on mac server. -------------------------------------------------------------------- -jn- (my two cats have accepted full responsibility for my opinions) --------------------------------------------------------------------