Re: MarkUp Validator's new clothes

Bjoern Hoehrmann wrote:

> * it groups identical error messages

Test this with some _real_ pages.  Often errors are only a
side-effect of another error starting before / ending after
the shown error, and then it's essential to see errors in
their context (i.e. by line number).

> * it reintroduces the ^ marker which helps to spot certain
>   errors

Nice.

> * no right-hand navigation bar
> I think it just distracts and takes space,

Yes, get rid of it.  It's ugly in browsers without CSS.
Fragments without name= don't work with my browser, so
I'm very happy if you eliminate this useless stuff.

The validator can handle HTML 3.2, therefore its output
should be also visible with a HTML 3.2 browser.  My HTML
3.2 browser is quite happy with XHTML 1.0 (after all XML
is simpler than SGML), but it insists on a name=

> The same applies to most of the links, a separate page
> maybe even the homepage, is a better place for these.

True.

> already sufficiently clear about whether the document
> validates or not.

I like the big "This page is valid XHTML 1.0 transitional".
Only transitional, because I found no better "visible with
any browser" DTD, although some 1.1 ideas make sense... :-(

> * it does not include page source

Excellent.  For big documents it's annoying if the validator
echoes what I sent (or what it found at the URL).  The line
numbers were _always_ correct so far, therefore showing the
complete source unconditionally (for any error) was a pain.

> It's sometimes useful, sometimes not

Then it should be an option, respected even after errors.

> * it does not include the navigation elements on the top,
>   it certainly should, but again, that's been out of scope

All links not working with "any browser" are useless.

> * it does not include all this legal stuff in the footer,
>  whether it must I do not know; I would not put it there

Legal stuff is always important, and you can collect more
irrelevant nonsense in the footer.  Insert some <del> tags,
they have no effect with legacy browsers.

> * it looks much much better, IMHO, and the styles are
> cross-browser

_Much_ better than the original output (your [1] link) with
my browser, which doesn't support any CSS.

> it could use scripting to hide the source code fragments
> and show them on user request

In another version designed for browsers supporting scripts.
Please be careful with the test, sometimes I'm confused and
enable JavaScript 1.1 in my browser.

                      Bye, Frank

Received on Sunday, 18 April 2004 15:00:51 UTC