Hello I'm preparing to start an ecommerce project that will require the use of sessions throughout the entire site for referrer and affiliate tracking purposes. I also plan to use mod_rewrite to make links to dynamic content search spider friendly. example: instead of /somescript.php?var1=value1&var2=value2 I'll have /somescript/var1/value1/var2/value2
The obstacle I'm facing is the automatic appending of the session id to every link upon first entering the site and to every link on every page of the site if session cookies are disabled. This will make search spiders either not index the linked pages, or worse index them with an old session id! I am hoping to find two solutions to dealing with this inherent problem with sessions, spiders, trans sid, and url rewriting. 1) I am thinking of obtaining and using a list of known spiders to determine if the visitor is a spider. If they are a spider I simply won't call session_start() thus eliminating the auto appending of the session id to every link. URL rewriting will take care of the ugly query strings and I'm in business to get spidered. How ever, this could mean a lot of overhead if the site gets busy and a lot of work keeping an updated spider list. Does anyone have a solution for this? I want some real world experience here please. 2) If I rewrite query strings how will I deal with trans sid? For instance, I have a rewritten url that looks like this: /somescript/var1/value1/var2/value2 and the session id is going to get appended leaving me with this: /somescript/var1/value1/var2/value2&PHPSESSID=some_session_id. Is there a way to make php add the session id in the slash delimited format? I am aware of arg_separator.output directive but that does not fix the = in the query string. I appreciate any help, tips, and/or pointers. The problem is clear - it's the solution that's a little fuzzy! Thanks for reading, Jim
