- From: Dave J Woolley <DJW@bts.co.uk>
- Date: Tue, 4 Jan 2000 14:29:55 -0000
- To: www-amaya@w3.org
> From: Richard Webber [SMTP:rwebber@bigpond.com] > > Behalf Of Dave J Woolley > > > From: Patrick T. Rourke [SMTP:ptrourke@mediaone.net] > > > <a > > > > > > href="http://www.perseus.tufts.edu/cgi-bin/morphindex?&lookup=%29abde%2Flu > > > kt > > > > > > a&display=Latin+transliteration&lang=Greek&corpus=2.0&formentry=1&embed=2" > > > >A > > > bdelukta</a> <br> > > > > > This is invalid; all the &'s should be replaced with & > > > > No, the &'s are delimiting the arguments of a GET-style CGI query. > However, the & before lookup is unnecessary, and may be confusing things. > > They *should* be &, although the one before lookup is also wrong. These &s occur in a CDATA parameter context. In such a context, SGML entitities are replaced. That means that any paramter name that spoofs a known entity will get substituted by that entity and any that doesn't will be at the mercy of the parser's error recovery. The HTML 4 specification goes as far as to advise people creating form handlers to accept ; as a synonym for & when the server is likely to be used with GET type form URLs coded in anchors rather than generated by forms. It does this to avoid the need to use &
Received on Tuesday, 4 January 2000 09:36:41 UTC