Web server

[Tim]

Scott van Looy:
>>> It's a better structure. Makes one more diciplined. It's just not the best
>>> thing to learn from scratch.

Tim:
>> It's the same structure.  The main difference being that you do write in
>> the previously optional closing tags.  And that's always been a
>> recommendation with HTML, anyway.  There's been quite a few browsers
>> which make mistakes when you didn't explictly type in the closing tags.

Scott van Looy:
> It's not. You can mix tags up in HTML4, even though it's not recommended, 
> whereas in XHTML they have to nest correctly.

Utter rubbish.  You cannot nest elements inproperly in HTML, it's not
allowed.  That's "not allowed", not "not recommended".  The same HTML
construction rules apply.  Just try validating HTML written improperly,
like that, and you'll soon see it's flagged as being an "error".

The behaviour of browsers is *another* matter.  Some will try to honour
your intentions, many will make a pigs breakfast.  all of them will be
operating under error conditions (the author, and the browser's
programmer), so all bets are off.

And whether you can badly type HTML or XHTML, is *another* issue, yet
again.

>>> Not at all. If you add a doctype then your document is rendered in
>>> standards compliance mode.

>> There's more to it than that.  Depending on how you typed in the
>> doctype, browsers would behave variously, in manners that they shouldn't
>> do.

> Not at all. If you type in the correct doctype they switch from quirks 
> mode to standards compliance mode. IE6 is fussy and so will stay in quirks 
> mode if anything appears above the doctype (spaces, carriage returns, the 
> XML declaration, anything) so you need to watch for that, but that's all

Now you're contradicting yourself.  First it was if you type in a
doctype, it's standards compliance.  I countered that, now you're
agreeing with me (it depends how it was typed).

If you type a doctype, depending on whether you include a URI, or not,
triggers silly behaviours in some clients.  Depending on whether you
used strict or transitional triggers silly behaviours.  Depending on
where it's typed in a file, also triggers silly behaviours.  That's
definitely not simply, "If you add a doctype then your document is
rendered in standards compliance mode" (your words).

IE is a particurarly good "bad" example.  If you give it XHTML,
properly, according to everything that you're supposed to do with XHTML,
it doesn't work.  It cannot, and will not, display it.


>> Both <br/> or <br /> are actually <br>> in "HTML".  And HTML client that
>> renders is like that is behaving correctly, any that doesn't, isn't.

> In which case not a single modern browser behaves correctly, which I find 
> unlikely ;)

I don't find it unlikely.  They're all loaded with masses of bugs.
Always have been, probably always will.

Besides it doesn't matter, unless you're playing silly browser
customisation games.

      * You serve HTML as text/html to anything, and you get HTML.
      * You *improperly* serve XHTML as text/html to anything, and you
        get varying results.  Which includes HTML improperly written
        with XHTML constructs, *and* XHTML properly typed as XHTML, but
        improperly served as if it were HTML.
      * You serve XHTML as application/xhtml+xml and you get proper
        results in browsers that can handle XHTML.

There are more browsers about, old and new, than any web author can
possibly know about.  It's a really stupid idea to send something to
them that isn't what it's described as.

>> Well formed HTML does *exactly* the same job as well formed XHTML does.
>> XHTML is not the magic fix people vaunt it to be.  Nor will it ever be
>> until Microsoft gets their head kicked in.

> You can write perfectly valid HTML4 that renders entirely differently in 
> FF IE and Safari. It's only in standards compliance mode using XHTML that 
> they start to actually correctly render pages.

They, i.e. all of them, still will not produce the same results if you
use XHTML.

> FF is broken in bits, IE is broken in other bits and so is Safari. Not
> a single one of them actually adheres correctly to the standard.

That much is true.  They've all got faults, some of them deliberate.

>> Conversely, correctly written and served XHTML can't even be read by
>> over 60% of the browsers.

> 84% (IE)

I was being generous.  More than 60% of all browsers are MSIE (80% is
"more").  MSIE is down to just 60% of the traffic to my website, and I
don't have a Mozilla magnet on it (hides something under the table).

Stats for this month, so far:

Browsers              Hits
--------              ----
MS Internet Explorer  3859  59.6 %
Firefox               1930  29.8 %
Safari                 221   3.4 %
Opera                  181   2.7 %
Mozilla                136   2.1 %
Unknown                 72   1.1 %
Konqueror               47   0.7 %
Netscape                20   0.3 %
WebCollage               2   0 %
Links                    1   0 %

Operating systems  Hits
-----------------  ----
Windows            5347  82.6 %
Linux               671  10.3 %
Macintosh           339   5.2 %
Unknown             104   1.6 %
Sun Solaris           9   0.1 %

And that's about the average sort of results, for the last year, or so.
I'm amused that, ocassionally, I see a Commodore 64 in there.  ;-)

There's a fair chance that MSIE actually has less hits than it's
tabulated for.  As quite a lot of browsers masquerade as it.  There's
little chance of the opposite, as I'm yet to find anybody other than the
odd screwball making MSIE masquerade as something else.

-- 
(This box runs FC6, my others run FC4 & FC5, in case that's
 important to the thread.)

Don't send private replies to my address, the mailbox is ignored.
I read messages from the public lists.