Lachlan Hunt wrote:
Semantics only become useful when there are tools that make use of
them in a useful way.

On 2/8/07, Geoff Pack <[EMAIL PROTECTED]> wrote:
So <name>Joe Blogs</name> is meaningless with out a spec to tell me that
'name' means a name, while <j79hfd98y28>[EMAIL PROTECTED]&*</j79hfd98y28> is
meaningful if a spec says so?

Read Lachlan's words carefully, Geoff. Using XML, you're free to
define your own language with whatever tags you'd like. That doesn't
mean there's any tool (e.g., web browser) that's going to be able to
parse whatever tag in whatever language and apply semantic value to
it.

Just because I feel <headerone> or
<mysupercoolheadinglevela1abeachfrontavenue> is a better way to
specify a heading than <h1>, is it reasonable to expect a browser
maker to cater to my linguistic whim? And by extension, to anyone's
linguistic whim? Browsers don't handle any random tags, browsers work
with a previously defined subset. That's just how they work. Atom
feeds don't accept any old tags, either. OpenOffice.org documents,
though they be XML, handle only a specific set of tags. I'm sensing a
pattern.

What if I write spec that says simply: "The meanings of all my tags
names are the same as the meanings defined in the Standard Oxford
English Dictionary"? What if I claim my spec to be the English language?

I could then further claim my document is more widely understood (and
implemented?) than HTML, simply because more people understand plain
English than HTML.

You might be able to get away with handling the presentation issues
with CSS, but that's not going to help the semantics of your document.
Human readability, unfortunately, does not translate to machine
readability. If machines were completely capable of parsing natural
language, we wouldn't be programming in Ruby, we'd be programming in
Japanese.

(I'm playing devil's advocate here, but only to show how absurd this
is.)

Yep.

Love,

Dan Dorman


*******************************************************************
List Guidelines: http://webstandardsgroup.org/mail/guidelines.cfm
Unsubscribe: http://webstandardsgroup.org/join/unsubscribe.cfm
Help: [EMAIL PROTECTED]
*******************************************************************

Reply via email to