== Browsers are the wrong place to report HTML validation errors

A popular idea for dealing with 'malformed' HTML is to have the browsers
warn users about it (the most recent example I've run across is in
comments [[here http://zaitcev.livejournal.com/116670.html]]), on
the theory that this will cause authors to make their HTML validate.
Unfortunately, doing this is about as useful as showing error pukes to
website visitors, and for the same reason: it is reporting the problem
to the wrong person.

Almost everyone visiting your site is a visitor, not the site's author.
It follows that almost every time this hypothetical 'page is malformed'
error would go off it would go off to a visitor, who can't do anything
about the problem, instead of to the site's author (who can).

The usual retort is that the site's author can visit the page as the
final step in publishing and see the warning and do something about it.
This is a marvelous theory, but (I argue) incorrect in fact, in part
because it assumes that site authors actually bother to check their
work, and in part because it assumes that site authors are going to
notice a little status notice any more than they notice any of the other
little broken things that they let slip by now.

(And if site authors do care about validated HTML they are probably
already using one of the validation tools to check their pages, and this
feature would not be a particularly big bonus to them.)

This is also a terrible feature from a pragmatic user interface point of
view: on today's Internet, it would be the boy who screams wolf all the
time, because a rather large number of the pages out there do not pass
validation. Such a warning notice would be on a lot; if it is intrusive
it gets in your face almost all the time (about something you can't do
anything about), and if it's not intrusive it's pretty much a noisy
waste of space. This is not a winning user interface element.

(But if you really want it, you can get Firefox extensions that do
this.)