The primary reason to not right a generic one is the suspicion that someone else has already written one that is good enough, which is why I'm here asking about it.

I also did an analysis of using our custom solutions as a base to create a generic solution. The result would need to be installed in under an hour and be flexible in other ways: it would need to easily support different designs, different shipping options, and different payment processor options.

I estimated this would take about 75 developer hours to create that software in-house.

When OScommerce takes 15 minutes to install a free, feature-laden stable store which generally meets my criteria, I think it merits heavy consideration versus spending a lot more time to have /exactly/ what I want.

The OSCommerce data model looks solid. I suspect that I wanted to extend it, it would be fairly easy to use Perl and talk to the same database directly.