Url Mangling

UrlMangling can actually refer to a number of different things. Here's a few usages I've seen:

JakobNielsen's technique of removing the latter portions of a now non-working URL in an attempt to get a valid link to something that's at least pertinent to what the original link was. This is probably becoming more and more common as people get more familiar with the Web and as more Web pages get generated on-the-fly instead of being statically served. See http://www.useit.com/alertbox/990321.html.

There's an interesting corollary: mappings of an URI space to documents (or dynamic pages) should be constructed such that all 'parent' paths of a given path will point to a meaningful page - precisely because the above strategy is successful for other reasons.

Note that the KFM Web browser does this for you if you press the "up-arrow" button.

Or, for other browsers, try http://www.bookmarklets.com/tools/navigation/index.phtml -> "Up a Directory".

Those of us who like to read up on WorldWideWebConsortium standards late at night will note that they very much recommend leaving off the end of URLs. For example, an image on a server could be image.gif and it could also be image.png. The idea is that the server will negotiate what the browser would prefer. Obviously old browsers will want the image.gif new ones the image.png. This would be achieved by <img src="image"> and the server / browser does the rest.

Neat. Hope it both catches on and people start setting up their servers for it.

-- MatthewTheobalds

In XML/XSL applications, UrlMangling might refer to what either the XML file or stylesheet writer does in attempt to make the output HTML file have legal URLs in any href fields. The characters that XML parsers accept as 'legal' are not always the ones that Web browsers want to see in their HTML for hrefs - particularly when it comes to angle brackets or the space character.

The technique of rewriting URLs on the fly to achieve a result equivalent to cookie-based servlet sessions.

See NameMangling