Browser Beatings:
Sidebars

"User-Agent:"/"HTTP header"
A quick minimally-technical introduction to how a Web client "gets" a page: the client opens a connection to the server, it then sends a request for a page by sending an appropriate command and the URL minus the "http://machinename" part (so "http://www.doh.org/blah.html" becomes "/blah.html" in the request). Attached to the plain request can be bunch of information about what kind of information the server can send it. One of the things the client can say, for example, is what kind of stuff it can display, so if the server is smart enough to be able to selectively serve JPEGs instead of GIFs, it knows which to send in response to this request.

"User-Agent:" is another line which it can send, identifying itself. So, for example, my Mac's Netscape Navigator identifies itself as "Mozilla/1.1b3 (Macintosh; I; 68K)," whereas my Sun's Lynx says "Lynx/2.3 BETA libwww/2.14."
After the server gets all of this information, it can process it any way it wants (most don't really do anything with it) and serve the page.

CERN provides more information about User-Agent:, the HTTP header and HTTP in general.

"usable by any browser"
OK, so this is a really strong statement. Maybe too strong. After all, in the end there are only two browser types that really matter: that which your client uses that which your client's target customer uses. However, if the intended target is "the Internet" then it's very important that your stuff work with as many browsers as possible. You never know, after all, when Bill Clinton may be surfing with some outdated $400,000 DoD browser looking for a new Secretary of Web Design or something. ;-)

mikek@presence.com
Last Modified: 4/16/95