Michael Crawford wrote: >Happy as I am that people find me through google, I also want to find >out who's taken the trouble to give me a real link. I would like to >get a list of static pages that have actual links to my pages.
Stephen Turner suggested REFEXCLUDE REGEXP:\?, but I didn't like this because it's side effect was to kill the Search Query report. James Herrmann, in an earlier message, suggested "You might try grouping as many search engines together into 1 line item using REFREPALIAS. This will free up your report." Except that doesn't work, because REFREPALIAS is an output alias, of which the docs say "they never combine lines, even if two lines end up with the same name." Thanks to the ideas of Michael, Stephen, James, I came up with this: REFALIAS REGEXP:google.*\?(.*)$ "search-engines?$1" REFALIAS REGEXP:search.yahoo.*\?(.*)$ "search-engines?$1" REFALIAS REGEXP:search.msn.*\?(.*)$ "search-engines?$1" ... REFREPEXCLUDE REGEXP:search-engines\? SEARCHENGINE search-engines* q,terms,search,query,keyword,request,MT,qu,p This converts every search engine name from http://www.google.com/... or search.yahoo.com... or whatever into just "search-engines?..." This makes it easy to exclude from the Referrer Report, which now will list only static links as desired. And it shows how many search engine referrers there were, without overloading the referrer report. And the Search Query report is left unchanged. QUESTION: There is one final unsatisfactory thing about this, so I'd be grateful for advice. Sites like Winamp do everything dynamic, so a referrer like this http://www.winamp.com/details.jhtml?componentId=126605 should, "morally speaking", be considered more like http://www.winamp.com/details.jhtml/componentId=126605 That's to say, for the thing after the question mark in this case, it makes more sense to think of it as a directory rather than an argument. In particular, in the referrer report, I don't want to see (pages)100 http://www.winamp.com/details.jhtml (pages)100 http://www.winamp.com/details.jthml?componentId=126605 Is there any way to tell Analog to do this? for a given url, to treat the arguments as part of the filename? I thought the ARGSINCLUDE and ARGSEXCLUDE might help, but they don't seem to. I tried to REFALIAS the ? into a /, and then output-alias it back again, but this didn't work because Analog's hyperref for the item appears to use the thing before the output alias has been applied. ----- As a related thing, I also wanted to clean out the "attacks" from the failure report, so I could see the real failures. James Herrmann again suggested an output-alias, FAILALIAS, but that has the same problem as before. So instead I used this solution: FILEALIAS */_vti_bin* attack FILEALIAS */_vti_inf* attack FILEALIAS */msoffice/cltreq.asp* attack FILEALIAS */cgi-bin/formmail.cgi* attack FILEALIAS */system32/cmd.exe* attack -- Lucian +------------------------------------------------------------------------ | TO UNSUBSCRIBE from this list: | http://lists.isite.net/listgate/analog-help/unsubscribe.html | | Digest version: http://lists.isite.net/listgate/analog-help-digest/ | Usenet version: news://news.gmane.org/gmane.comp.web.analog.general | List archives: http://www.analog.cx/docs/mailing.html#listarchives +------------------------------------------------------------------------
