Date: Fri, 21 Jul 1995 13:08:18 -0400 From: Ka-Ping Yee <firstname.lastname@example.org> Subject: Re: color: NCSA Mosaic, Netscape, and HTML3 To: Michael Johnson <email@example.com> Cc: Terry Allen <firstname.lastname@example.org>, HTML discussion list <email@example.com> In-Reply-To: <MICHAELJ.firstname.lastname@example.org> Message-Id: <Pine.3.87.9507211318.A24752email@example.com> On Fri, 21 Jul 1995, Michael Johnson wrote: > Perhaps I should have been more explicit. The following: > > <img alt="> some text <"> > > is, according to , a valid construct. However, my parser will look at > this and assume that it's more likely to be an erroneously unclosed literal > inside of a tag than a literal that happens to contain '<'. > > I deliberately chose to do it this way because people rarely (probably never) > use '>' and '<' in literals, but they often forget to close literals, thanks > to the brain-dead parsers in Mosaic and Netscape. But does that mean your parser should be equally brain-dead? What you have done may well make it impossible for people to use the '<' and '>' characters within attribute values -- a painful and irrational restriction. > I do something similar in a case like this: > > <a href="http://foo.bar.org/" <img alt="some text"></a> > > When I see this, I assume that the author omitted the trailing '>' on the > anchor tag. I flag this as a lexical syntax error and implicitly insert the > missing '>' character. Someone correct me if this is wrong, but i believe this behaviour is actually SGML-compliant. Isn't < one of the ways to implicitly end a tag while starting another one? Ping (Ka-Ping Yee): 2B Computer Engineering, University of Waterloo, Canada firstname.lastname@example.org | 62A Churchill St, Waterloo N2L 2X2, 519 886-3947 CWSF 89, 90, 92; LIYSF 90, 91; Shad Valley 92; DOE 93; IMO 91, 93; ACMIPC 94 :: Skuld :: Tendou Akane :: Belldandy :: Ayukawa Madoka :: Hayakawa Moemi :: New! <http://csclub.uwaterloo.ca/u/kryee/> Yeah, i finally made a home page.