Your post brings back fond memories of what I was doing during the Internet boom five years ago.

But to answer your questions..

How do I handle the output from the plotting process/what is the best method?

I just re-directed the output from the gnuplot command to a temporary file with the .out suffix.

Can I send the image data to the browser?

Sure, but that's probably not what you want to do.

Do I create temporary files?

That's what I did. I just called a routine to cook up a temporary file name and used that in the gnuplot command file. The temporary files went into /var/tmp, and I even wrote a script to clean up that directory every an hour, having no idea that of course there was already a utility to do that (tmpwatch). At least I knew enough to use cron.

The ones referred to in the above referenced post appear to be temporary files. Do I write files to the directory and just href them?

Yup. Works great! Just be sure to clean up after yourself. Were I doing this again, I would use File::Temp to create temporary files. Back then I was a skilled C programmer fumbling my way around Perl, but learning fast.

Good luck!