Meaning of GET/favicon.ico absent any other GET

Foreign host requests favicon but nothing else

[jec2002]

How can a machine request a favicon, which means a page is being bookmarked, without first requesting the page itself. I find this very puzzling.

While reading my access_log file, I noticed that a machine requested my favicon, but the same machine made no other requests. How can that be? And what does it mean? I'd appreciate your comments. Thanks.

[jdMorgan]

jec2002,

If the user-agent was Gecko, then it means the user reopened a browser tab into which he had previously loaded your page from his cache. For some reason, the Gecko-based Mozilla and Netscape browsers won't use the cached icon when a tab is reopened...

Jim

[bobriggs]

in addition to jdMorgan's post, I've seen spiders/robots requesting favicon.ico without robots.txt/or any other page.

Requests seem to come from universities, maybe they're teaching how to write bots there, or count % of sites using the favicon.ico file. Seems harmless, unless you rely on that statistic and you're getting hammered with them.

[jec2002]

Thanks, guys. I think the machine is at a university. Here's the request:

dsc11-waf-dc-1-228.rasserver.net - - [27/Nov/2002:12:09:10 -0800] "GET /favicon.ico HTTP/1.1" 200 21630 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows 98; Win 9x 4.90; H010818; MSN 8.0; MSNbMSNI; MSNmen-us; MSNcIA)"

My concern is that my work, which is literature, is being served by another machine. There's a good demand for my writing, but I give it away for free, so I can't understand why someone would want to serve it up somewhere else. But, I have seen knowledge banks spider it. I've got a NOARCHIVE meta tag on every page, and I just have to hope that the tag is respected.

Any additional comments on this query would be appreciated. Thanks again.

[bull]

not a university, IP belongs to icg.net

you have <link rel="SHORTCUT ICON" href="http://www.you.net/favicon.ico"> set in your html file(s)?

bull

[jec2002]

Bull, the circumstance you cite was the one I feared because it would mean that my work is being served up somewhere and they're bookmarking it, calling the favicon, BUT, I don't use the shortcut link, I just put the favicon in the root. I don't see how they could be calling my favicon on my server without the link relation, do you? Please elaborate. Thanks much.

[edited by: jec2002 at 11:42 pm (utc) on Nov. 30, 2002]

[jec2002]

0

[jdMorgan]

Bull,

IIRC, the default behaviour if no favicon path is specified is to attempt to fetch it from the site's root.

If you set up your cache-control server headers and tag your index.html (home) page with "no-cache, must-revalidate", then at least you see requests for that page.

Jim