cesar martin.iworkwithyou ([EMAIL PROTECTED]; Friday, October 25, 2002 2:07 PM):
> sometimes the ability to count the "number of users" looks more like a > political decision than a "technical decision". > I mean, is easy to count users and looks like analog can not do it just because > "we" don't want to be able to do it... > but why? > and don't tell me is the way internet works... It's some of both. Tehcnical and political. The only mechanism that HTTP provides for this is authentication and most web sites don't require any log in. In addition, HTTP and Web browsers support mechanisms that make is complicated to get full details on distinct users (caching and proxying). This is covered in http://analog.cx/docs/webworks.html, which it sounds like you have read. If you do have a system in your web server that manages state (for example an e-commerce system or a client-side solution) then you could use that information to track users. You can re-use one of the "extra" variables, like %u or %v, in a custom log format and get a report on users from that. But unfortunately, there are no standards for this and in reality, the majority of Analog users don't have any state-generating systems on their web sites (and yes, I realize this is self-fulfilling). Estimating visits, sessions or unique users is prone to error. It may be "easy" but that doesn't mean it's right. It's possible to get a very close estimate of sessions with some fancy heuristics, but even that has at least a 10% margin of error. Even provided that you could generate a state variable that was countable, the best Analog would do is provide a total of the number of "visits" or "users" and list each (with some meaningless key, no doubt) and the number of total requests, pages, and bytes each visit or user requested. See http://analog.cx/docs/faq.html#faq128. (Yes, this is also a technical/political decision.) That's the technical end of the argument. For the political one, I suppose will have to speak up. :-) -- Jeremy Wadsack Wadsack-Allen Digital Group +------------------------------------------------------------------------ | This is the analog-help mailing list. To unsubscribe from this | mailing list, go to | http://lists.isite.net/listgate/analog-help/unsubscribe.html | | List archives are available at | http://www.mail-archive.com/analog-help@;lists.isite.net/ | http://lists.isite.net/listgate/analog-help/archives/ | http://www.tallylist.com/archives/index.cfm/mlist.7 +------------------------------------------------------------------------
