W3C Validation -- Amusing Results

Continuing in my subjecting a few websites to the W3C Markup Validation
Service, here are a few more results, this time for the four sites I check
daily to get caught up with the news....

The New York Times
<http://nytimes.com/> -- 340 errors -- 80 warnings.

USA Today
<http://www.usatoday.com/> -- 328 errors -- 13 warnings

Time
<http://www.time.com/time/> -- 258 errors -- 89 warnings

Newsweek
<http://www.newsweek.com/> -- 30 errors -- 0 warnings


It would be interesting to know what Newsweek is doing that the other three
aren't.

--
James Leo Ryan ..... Austin, Texas ..... taliesinsoft RemoveThis @me.com
 >> Stay informed about: W3C Validation -- Amusing Results 

In article ,
TaliesinSoft wrote:

> I just now subjected four websites to the W3C Markup Validation Service and
> here are the results....
>
> <www.adobe.com> -- 13 errors -- 7 warnings
>
> <www.apple.com> -- 1 error -- 0 warnings
>
> <www.microsoft.com> -- 208 errors -- 37 warnings
>
> <www.softpress.com> -- 0 errors -- 0 warnings
>
> The Microsoft site was actually redirected to a longer URL.

I have found that even when the default first page of a web site is
valid, interior pages are often not. Try some other page at random from
any site that seems valid.

Two of my web sites
http://carlylegardensgnome.com/
http://airliebeachbum.com/
are not only valid XHTML for all their pages, but also Internet Explorer
totally fails to work on them (well, except on the first page, which is
in HTML).

You should check the
http://dev.opera.com/articles/view/mama-markup-validation-report/
for part of a wonderful and extensive study of what (mostly) is not
valid. One interesting sidelight for Macintosh users. Of the 2504 sites
done with Apple iWeb, 81.9% were valid. The next nearest web site editor
was Dreamweaver, with 3.4% valid.

--
http://www.ericlindsay.com
 >> Stay informed about: W3C Validation -- Amusing Results 

In article ,
TaliesinSoft wrote:
> I just now subjected four websites to the W3C Markup Validation Service and
> here are the results....

> <www.adobe.com> -- 13 errors -- 7 warnings

> <www.apple.com> -- 1 error -- 0 warnings

> <www.microsoft.com> -- 208 errors -- 37 warnings

> <www.softpress.com> -- 0 errors -- 0 warnings

> The Microsoft site was actually redirected to a longer URL.


Much of the problem is that no Microsoft software for generating html
seems to be able to produce valid html.

That may be too generalised a statement - but it's clear that a huge
percentage of the non-valid pages are not produced by humans but by
Microsoft software. If Microsoft pay zero attention to validation ...

It's also clear that the validator is rather _too_ pedantic!

--
------------------------------------------------------------------
Richard Torrens. News email address is valid - for a limited time only.
http://www.Torrens.org.uk for genealogy, natural history, wild food, walks, cats
and more!
 >> Stay informed about: W3C Validation -- Amusing Results 

In article ,
"Richard Torrens (News)" wrote:

> It's also clear that the validator is rather _too_ pedantic!

How is that? If you write valid markup, including within a program that
produces valid markup, the validator produces valid results. No? If not,
enlighten me.

leo
 >> Stay informed about: W3C Validation -- Amusing Results 

In article
,
Leonard Blaisdell wrote:
> In article ,
> "Richard Torrens (News)" wrote:

> > It's also clear that the validator is rather _too_ pedantic!

> How is that? If you write valid markup, including within a program that
> produces valid markup, the validator produces valid results. No? If not,
> enlighten me.

There is a huge amount of markup which works fine in just about every
browser, yet is technically incorrect. E.g.

<A NAME="content"><H2>Contents</H2></A>
is wrong but surely doesn't confuse any browser.

<H2><A NAME="content">Contents</A></H2>
is correct.

In general, I approve strongly of valid html. I recently pointed out to
Zen.co.uk that their pages didn't validate.

Their response was that Microsoft's site returned 208 errors...

So we have just requested a MAC from Zen. I'm not paying high charges for
essentially Microsoft only support when I cannot use that support!

--
------------------------------------------------------------------
Richard Torrens. News email address is valid - for a limited time only.
http://www.Torrens.org.uk for genealogy, natural history, wild food, walks, cats
and more!
 >> Stay informed about: W3C Validation -- Amusing Results 

In article ,
"Richard Torrens (News)" wrote:

> In article
> ,
> Leonard Blaisdell wrote:
> > In article ,
> > "Richard Torrens (News)" wrote:
>
> > > It's also clear that the validator is rather _too_ pedantic!
>
> > How is that? If you write valid markup, including within a program that
> > produces valid markup, the validator produces valid results. No? If not,
> > enlighten me.
>
> There is a huge amount of markup which works fine in just about every
> browser, yet is technically incorrect. E.g.
>
> <A NAME="content"><H2>Contents</H2></A>
> is wrong but surely doesn't confuse any browser.

Enclosing a block element like a heading (normally on a new line) within
an inline element like a link seems a reasonable way to start confusing
a browser. You get away with it because browser error handling is so
very good.

> <H2><A NAME="content">Contents</A></H2>
> is correct.

However if a web page is both valid and uses a Strict Doctype, that puts
a browser in Standards mode, you get to take advantage of improvements
in browsers. The folks who write the browsers seem to have hit their
limits in handling old bugs in Transitional web pages. The improvements
are mostly in Standards mode.

At least with a valid page in Standards mode, you (mostly) know what the
web browser is supposed to do with it, according to the specifications.
With an invalid page, you have to guess how the browser error handling
will cope with it. There is no specification for error handling in
browsers. This makes it even harder for Macintosh users, because Safari
and Firefox and Opera all have different error handling to Internet
Explorer.

--
http://www.ericlindsay.com
 >> Stay informed about: W3C Validation -- Amusing Results 

In article ,
Eric Lindsay wrote:
> >
> > <A NAME="content"><H2>Contents</H2></A>
> > is wrong but surely doesn't confuse any browser.

> Enclosing a block element like a heading (normally on a new line) within
> an inline element like a link seems a reasonable way to start confusing
> a browser. You get away with it because browser error handling is so
> very good.

I cannot disagree. However - browser parsing _is_ that good. Maybe the
validation should accept that.

But it's no big issue: the issue is that sites should at least try and
validate.

> > <H2><A NAME="content">Contents</A></H2>
> > is correct.

> However if a web page is both valid and uses a Strict Doctype, that puts
> a browser in Standards mode, you get to take advantage of improvements
> in browsers. The folks who write the browsers seem to have hit their
> limits in handling old bugs in Transitional web pages. The improvements
> are mostly in Standards mode.
>
> At least with a valid page in Standards mode, you (mostly) know what the
> web browser is supposed to do with it, according to the specifications.
> With an invalid page, you have to guess how the browser error handling
> will cope with it. There is no specification for error handling in
> browsers.

Apparently there indeed are standards for how browsers should handle
broken pages.

> This makes it even harder for Macintosh users, because Safari
> and Firefox and Opera all have different error handling to Internet

RISC OS is in an even worse state, so I fully understand you. The only
browser being developed on RISC OS is www.netsurf-browser.org

also for unices, and Amiga.

--
------------------------------------------------------------------
Richard Torrens. News email address is valid - for a limited time only.
http://www.Torrens.org.uk for genealogy, natural history, wild food, walks, cats
and more!
 >> Stay informed about: W3C Validation -- Amusing Results 

In article ,
"Richard Torrens (News)" wrote:

> In article ,
> Eric Lindsay wrote:
> > >
> > > <A NAME="content"><H2>Contents</H2></A>
> > > is wrong but surely doesn't confuse any browser.
>
> > You get away with it because browser error handling is so
> > very good.
>
> I cannot disagree. However - browser parsing _is_ that good. Maybe the
> validation should accept that.

However, ever since HTML2, the HTML markup language has been declared as
being an SGML derived product. This in turn means that the Doctype
points to a machine readable DTD which actually specifies exactly how
you need to parse the HTML. That is technically what the validator does.
You could write a browser that did the same, except most pages would
break. So all the browsers ignore the DTD, in favour of using their own
internal representation of how a web page is constructed.

In Internet Explorer for Mac 5, Tantek Celik first started using Doctype
sniffing to decide whether a web site was an old site (probably broken),
or a new site that should be written to W3C specifications (indicated by
a Doctype). IE6 for Windows did the same. Unfortunately, people doing
web sites started putting Doctypes in them, mostly because other people
were putting Doctypes in them. And people writing web editors started
adding Doctypes automatically.

> But it's no big issue: the issue is that sites should at least try and
> validate.

I could not agree more. That is why several of my web sites are actually
written in XHTML rather than HTML. This is not something any commercial
site could do. As a result, Macintosh OS X users can view my web sites,
since all standard OS X web browsers can handle XHTML correctly.

However Windows users with Internet Explorer can not view my web sites.
Internet Explorer is totally unable to use XHTML. As a result, you
basically can not use XHTML on the web. So, no-one does.

> > At least with a valid page in Standards mode, you (mostly) know what the
> > web browser is supposed to do with it, according to the specifications.
> > With an invalid page, you have to guess how the browser error handling
> > will cope with it. There is no specification for error handling in
> > browsers.
>
> Apparently there indeed are standards for how browsers should handle
> broken pages.

No, there are not. There are W3C specifications for how to write web
pages, and these have the authority of a standard (according to the RFC
that passed control to W3C). All the major browser makers (including
Microsoft, which were first to use CSS, for example) are members of
W3C. All have agreed to make their web browsers use W3C specifications.

For handling errors, each web browser tries to imitate what previous
versions of browsers (mostly Netscape Navigator and IE3 through 5) did.
Then IE6 tried to imitate IE5. Then IE7 tried to imitate IE6. Then IE8
got a spevial little switch that would let it imitate IE7 imitating IE6,
and so on. And every other browser tried to imitate most of this.

--
http://www.ericlindsay.com
 >> Stay informed about: W3C Validation -- Amusing Results