Re: HTML Validation Form [Was: Nesting of HTML elements ]

Marc VanHeyningen (mvanheyn@cs.indiana.edu)
Sun, 30 Oct 1994 21:54:54 -0500


[Hmmm... looks like www-talk is working again, at least as well as it
ever does.]

Thus wrote:
>/*
> * "Re: HTML Validation Form [Was: Nesting of HTML elements ]" by Rick Troth
> * written Sun, 30 Oct 1994 22:40:23 +0100
> *
> ** PS Dan, why do you always say %7E where other folks use ~ is this
> ** something else we are all doing wrong?
> *
> * He just wants attention.
> */
>
>Heh. Actually, everyone is doing something wrong, and it's all my
>fault. When I added user public information directories to NCSA httpd,
>I used /~ as the magic prefix. But ~ is a character which, according
>to the URL spec, is not supposed to appear in URLs unless % escaped. I
>didn't realize this until either Dan or Marc V. pointed it out to me
>some weeks later.

I think it was me, I'm pedantic like that. :-) Actually the spec has
varied a bit over time, but both RFC 1630 describing WWW URLs and the
latest standards-track URL definition from the URI WG declare that the
tilde may not appear in URLs unencoded. As far as "parts of
specifications for WWW stuff that lots of people ignore in an
indignant fashion" it's far from the worst offender, of course. I
think NCSA's httpd will work with the proper %7E encoding anyway, so I
wouldn't exactly blame Rob for it. I suppose the documentation could
mention it.

What I really don't understand is the reaction some people have to
this; it's often something like "That's stupid and annoying" or "Who
made the people who write the spec god?" or "He just wants attention"
rather than rational reactions like "Hmmm... why is that?"

--
Marc VanHeyningen  <http://www.cs.indiana.edu/hyplan/mvanheyn.html>