Hi Terry, This is the sort of problem that is almost impossible to diagnose blind. We can't see your C-program and CGI scripts source, don't know what server your running them on, nor can we easily see a reproduceable example of it going wrong. Macs clients are renowned for producing CGI-type problems, especially from forms. It may be of some help to those on this list to have this info. on which to base a diagnosis, but it's probably easier in the long run for you and/or your colleagues to hack the parser to return the results of its parsing, and test the output with different clients with input data that you know exhibits the symptoms. Then you should be able to home in on the problem fairly precisely. . . . trying to be helpful :-) Rgds., Alan at Venus .sig below On Thu, 18 Jul 1996, Terry M. Gray wrote:

For two years we have been developing a web interface to some quantum chemistry programs developed at Calvin. We're preparing a manuscript for J. Chem. Ed. describing these.

The approach that we took was to keep the original FORTRAN programs intact but to develop a web-form based interface calling a CGI script that runs the progroams and prepares numerical and graphical output that is returned to the user. We also have a quantum chem primer and excerises on on-line to support these.

The URL's for the forms are given below. The main URL for the site is http://www.calvin.edu/chemistry/computational/computationalTOC.html You are welcome to check these out and comment if you like.

My main reason for writing, however, is not to promote the site, but rather to get some help with a problem that we are having. I recently submitted this problem to the comp.infosystems.www.authoring.cgi newsgroup, but have not received any help yet. So, since this was a chemistry based web-group, I thought I might get some help here.

The CGI scripts worked under Netscape 1.12. They still work with Netscape 2.0.2 on a UNIX box (Sun and SGI). BUT they don't work with Netscape 2.x and up (even 3.0b4) on the Mac. I think that they work okay under Windows and Windows 95, but I haven't tested them extensively there.

Has anybody heard of such weirdness?

They are fairly complicated CGI's that take some form data and then run some computational chemistry programs on the data.

Here are some of the form URL's for anyone who wants to try them out:

http://www.calvin.edu/chemistry/computational/begatom/begform.html http://www.calvin.edu/chemistry/computational/atom/atomform.html http://www.calvin.edu/chemistry/computational/atomplus/atomplusform.html http://www.calvin.edu/chemistry/computational/ion/ionform.html http://www.calvin.edu/chemistry/computational/webgauss/webform.html http://www.calvin.edu/chemistry/computational/webadvgauss/webadvform.html

The problem seems to be in the parsing step, which is a C program. The rest of the CGI script seems to work fine. In other words, it takes the default values from the form or old input files from a previous run and works on them. Of course, this is disastrous because the user thinks that the data returned is for the data input in the original form submission, but is not.

Thanks for anyone who can help.

Terry M. Gray ----------------8<----------snip Rgds.,

Alan Mills at Venus Internet Ltd, 24 Denmark Street, London WC2H 8NJ, UK email:alan@venus.co.uk web:http://www.venus.co.uk Tel: +44 (0)171 240 5858 Fax: +44 (0)171 240 5859 Providing Web and Internet services and solutions ----- chemweb: A list for Chemical Applications of the Internet. Archived as: http://www.ch.ic.ac.uk/hypermail/chemweb/ To unsubscribe, send to listserver@ic.ac.uk the following message; unsubscribe chemweb List coordinator, Henry Rzepa (rzepa@ic.ac.uk)