here is an example
http://www.amazon.com/exec/obidos/ASIN/073571228x/qid%3D1099770 587/sr%3D11-1/ref%3Dsr%5F11%5F1/102-8947619-9086523
http://www.amazon.com/exec/obidos/ASIN/1861004494/qid%3D1099770 864/sr%3D11-1/ref%3Dsr%5F11%5F1/102-8947619-9086523
the isbn# folows the /ASIN/ in the url note that the next pointer /qid%xxx... has changed.
simply changing the field after /ASIN/ to the isbn will not give you the correct result.
If I understand you correctly, the root of your problem is that all the stuff *after* the ISBN doesn't matter at all. It's all session stuff. All you need is this: http://www.amazon.com/exec/obidos/ASIN/073571228x/ ...in order to pull up an Amazon record. Amazon adds all the rest. So we make the link in the template like this: http://www.amazon.com/exec/obidos/ASIN/<!-- TMPL_VAR NAME="isbn" --> It should work fine.
3. a computer driven enter or cr.
This is the part that's not possible. You can't simulate user input like that on a separate site. And if we're just talking about pulling up an Amazon record from an ISBN, we don't have to. One concern, however, is how far to trust Amazon to always have details on the item in your catalog. That's where a solution like Chris' has advantages--you're sucking down the information into your own database, so you know whether there's extended information to display. -- Owen