Msg#: 3035199 posted 4:57 pm on Aug 6, 2006 (gmt 0)
I just did two google searches: allinurl:& and allinurl:&
The first search returns zero results. The second search returns 672 million results, and every one of them has & in the uri, despite the fact that an explicit search generates zero results. So it looks to me that Google is right on top of this, and the two versions of the uri are clearly not seen as duplicates. Rather, everything seems to be mapped to the & version.
Msg#: 3035199 posted 6:53 pm on Aug 6, 2006 (gmt 0)
That might be an artifact of the allinurl: function. I agree that "&" and "&l:" are normally interchangeable, but they are distinct URLs that could theoretically produce different results. Like "www" and non-"www" domain addressing, for example.
Msg#: 3035199 posted 10:25 pm on Aug 6, 2006 (gmt 0)
Links with & are fine on a browser for sure. The browser knows to request a URL with just an & in it from the webserver, and not to try to find a character to draw on the screen called &productid; or &pagenumber; or something.
Look in your raw log files and see that even though the link has & in it, that the browser requests a URL as if it just has a & when the request hits the server.