Re: text/html for xml extensions of XHTML

On Sun, 24 Jun 2001, Al Gilman wrote:
> At 07:04 AM 2001-06-24 , Ian Hickson wrote:
>> On Sun, 24 Jun 2001, Arjun Ray wrote:
>>> On Sat, 23 Jun 2001, Ian Hickson wrote:

>>>> What's wrong with XHTML sent as text/xml? 

> The Arab of HTML applicability rides on the camel of XML-compliant
> markup in XHTML.

The Arab of XML should think twice before letting HTML the camel stick
its nose inside the tent.

> Arjun's feeling is that the suitability for end use as HTML is
> more important to know than the parseability as XML, for XHTML as
> we know it today.

Well, yes, but that's not how I view the situation.  It is an eternal
frustration of geeks that the world is mainly non-geek.  Put a tool
created by geeks for geeks into the hands of non-geeks, and all the
geeky reasons to have created the tool become irrelevant, because what
happens with the tool depends on what non-geeks make of it.  

Before it went mainstream, the Internet was a very geeky place: the
old saw "be conservative in what you produce, and liberal in what you
accept" is the quintessence of geekery, because the sensibility works
only to the extent that parties *care* about the hows and whys, and
thus are interested enough to know whether things are working by

Non-geeks, in contrast, are typically liberal in what they produce,
and conservative in what they accept.  This is because they have
neither time nor temparament to inquire into the geekiness of things:
they simply want things to work, and are annoyed or nonplussed when
they don't.  The important thing, though, is that this is not a
failing of non-geeks.  They simply have different priorities.

Good engineering saves non-geeks from the trouble of having to work
out whether something is working by design.  It is the height of bad
engineering to *know* that people take things for granted and yet
arrange to falsify the corresponding expectations, especially with
highminded righteous "no, you can't do that" moralizing.

HTML in practice - what goes over the wire as text/html - is very
non-geekish.  XHTML is yet another geeky invention in a long series of
quixotic attempts to retrofit inappropriate formalisms.  That's bad
engineering already, but much worse is the blinkered wishful thinking
that XML will be more important to XHTML than its HTML-ness.

> The message for our type language is that there is a market for
> each kind of information, and that a compact tree of types may
> provide insufficient room for all the information that one would
> seek to know with graceful migration as the business rules of type
> practice change under our feet.

Sorry, you lost me there.


Received on Sunday, 24 June 2001 17:01:49 UTC