W3C home > Mailing lists > Public > whatwg@whatwg.org > March 2007

[whatwg] Joe Clark's Criticisms of the WHATWG and HTML 5

From: Nicholas Shanks <contact@nickshanks.com>
Date: Fri, 23 Mar 2007 21:27:05 +0000
Message-ID: <733F1C17-8804-4858-ACBD-17D6E2E75669@nickshanks.com>
On 23 Mar 2007, at 18:26, Robert Brodrecht wrote:

>> I welcome <code> to mark up blocks of code, but I don't think HTML
>> should go further than that, if you want to mark up computer  
>> code   that badly, use XHTML + some CodeML equivalent to MathML.
>
> I'd love to, but one of the major browsers doesn't support XHTML. :(

Browsers that don't natively support XHTML aren't that important anyway.
All of the browsers I have access to (that are currently maintained)  
seem to cope with it. This includes Firefox, Opera, Safari, Amaya,  
Lynx, Links, OmniWeb, iCab and many more smaller ones based on the  
Gecko and WebKit engines. Granted they may not understand the nuances  
of XHTML, but I can still read a document's contents.
The browsers that doesn't are w3m 0.5.1 (2004), Netscape Communicator  
4.8 (2002), Internet Explorer 5.2.3 (2001), TurboGopher 2.1 (1995)  
and WWW/Samba 1.0.3 (1993)
I can't test anything else as I don't have it.

I don't see what the problem would be with serving HTML-compatible  
XHTML + CodeML to all of the above. Sure the older browsers wouldn't  
understand the CodeML, but there wouldn't be any great loss, and  
crucially the document authorship could retain it's semantics.
You would also have to send the wrong MIME type to a few archaic  
browsers, but that doesn't really affect the HTML5 specifications.

>> But how can you justify the presence of <kbd> when so few people    
>> write content where keyboard input has to be represented?
>
> <form>
>  <label>
>    Enter your e-mail address (example: <kbd>joe at gmail.com</kbd>
>    <input type="text" name="email">
>  </label>
> </form>

Coming up with usage examples is trivial, justifying why they deserve  
to make the cut into a formal specification is not.

>> Why isn't <tv-show> an  element?
>
> I just use <cite> when I reference a TV show.  Cite is another element
> that I wish had more power...

There has been so much previous discussion on <cite> that I won't  
continue that here, except to say that I think you are using it  
wrongly unless you are actually citing a TV show rather than just  
talking about it and want to italicise the name. (Not sure on your  
meaning of 'reference' above.)

>> One only has to look at the examples given in the HTML5 spec to see
>> how esoteric <samp> is:
>> http://www.whatwg.org/specs/web-apps/current-work/#the-samp
>
> It's used to mark computer output.  That isn't esoteric.

I disagree. It's domain is the root of it's esoteric nature, as its  
domain is esoteric.

> <samp> might not be the best name for the tag, but it does have a  
> specific meaning.

I like a good mix of specific tags and general tags, as long as both  
kinds are applicable to a general audience (meaning most of the  
people and companies that put content online).

>> I strongly agree. It's domain is also not clear enough either. Does
>> morse code count? What about encoded strings?  <code
>> type="rot13">uryyb eboreg</code>
>> People who aren't programmers have a different understanding of the
>> meaning of the word than we do. Confusing elements leads to both
>> decreased and incorrect usage.
>
> It says "computer code."  That would rule out Morse code.  I've always
> looked at "computer code" to be lay-speak  for "programming  
> languages and the like."

Depends on your definition of computer too. The abacus is a computer.
Are punch cards computer code? Does assembly count or are we  
restricted to higher level computer languages like Ada? Does  
scripting count? What about completely visual programming/scripting  
where no code is hand-authored and the <code> block contains a series  
of screenshots or a movie demonstrating how to create the software?

> It may, again, come down to the description of the tag, not
> the tag itself.

But as you have said, most people look at the tag name, draw a set of  
assumptions based on the use of the word in general English, and  
never look at the specifications. They'll then go off and blindly use  
it wrongly.
Even Joe Clark (who's rant this thread originally referred to)  
clearly did not read <meter> and assumed it meant <metre> (he seeming  
cannot spell either :-)

>> No, you missed the point again. <samp> is short for sample. Misguided
>> hair care people of the future will think their product sample counts
>> as a sample and use it for that.
>
> My point was: I can sell you a hunting rifle for hunting, but I  
> can't stop
> you from using it as a walking stick.  All I can do is say, "this is
> intended to be used to kill animals."  It's not the hunting rifle's  
> fault.
>  It's not my fault.  You're the one using it the wrong way.  People
> misused <blockquote> to indent text, for example.

I think misuse is something better handled by error and warning  
messages in the dominant user agents. For instance if every  
<blockquote> required a non-empty child <cite> element, then when  
that was not provided, the UA could say "This page contains 1721  
mistakes.       vv More Info vv
1. Error: This quotation does not cite it's source. All quotations  
are required to cite their source with a <cite> element, even if all  
you can provide is ?Anonymous poet? or similar.
2. Warning: No citation href was provided for this quotation. Authors  
are advised to provide a link where available.
3. ..."
Basically make it frustrating for people to abuse elements, and make  
correct use easy.

> But some people (e.g. your programmer) do care if it is computer  
> output.
> You are telling one group to go to hell while embracing another.

Not at all, I'm saying "your usage case is not general enough for  
explicit support, please use a secondary vocabulary". Whether that  
vocab is Dublin Core, hCard, MathML or whatever is irrelevant.

> I don't
> know why grocers are more important than programmers (and I'm  
> certainly
> not suggesting the other way around).  The problem with your
> interpretation of the word "bloat" is that it is a see-saw.  We can  
> either
> 1) serve one group better and another worse or 2) serve two groups  
> poorly.

I would rather we had
3) Serve all groups equally, and don't favour TimBL/Dan Connolly/Marc  
Andreessen's obsolete, historical views when deciding what is in the  
interests of all. (their views have all since changed, sadly HTML is  
still lumbered with the legacy of 1991 usage?for example, the ISINDEX  
element!)

>  Rather, we could serve all groups well by adding tags in a calculated
> manner (not willy-nilly) so as to avoid cruft while still providing  
> a rich
> set of semantically useful elements.

Yep, this is what we all want, and what I am arguing for. A ?purer?  
reality-based HTML and not a markup language for geeks.

> Just because I like my programming
> tags doesn't mean I don't like your retail tags.  We can coexist.
>
> That said, <var> may be cruft.  I would just use <code>.  But, I think
> computer input (<kbd>), computer output (<samp>), and the stuff that
> processes it (<code>) are important to people who program for a  
> living,
> just as <product> would be important to someone who sells stuff for a
> living.  Adding useful elements wouldn't hurt.  Removing perceived  
> useless
> elements may.

Well I was advocating moving them into another spec, not eliminating  
them entirely. But I have said my piece so I shall shut up now.

- Nicholas.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2157 bytes
Desc: not available
URL: <http://lists.whatwg.org/pipermail/whatwg-whatwg.org/attachments/20070323/9b4a7ff4/attachment.bin>
Received on Friday, 23 March 2007 14:27:05 UTC

This archive was generated by hypermail 2.4.0 : Wednesday, 22 January 2020 16:58:54 UTC