[whatwg] Ghosts from the past and the semantic Web

[Henri Sivonen]

On Aug 28, 2008, at 19:46, Ben Adida wrote:

> Henri Sivonen wrote:
>>> Same goes with MySpace widgets. Paste one thing, get the widget.  
>>> Who's
>>> going to go paste two things in two different places? It's really
>>> important to make HTML the carrier of this information.
>>
>> It seems to me that this line of reasoning should lead to using
>> identifiers that are contained in one string instead of splitting
>> identifiers and putting the pieces in two places like CURIEs do.
>
> There's "two places in the same HTML block", and "two places in
> different files or different regions of the HTML page."
>
> Have you seen the output of the CC license generator? Here it is:
>
> =========
> <a rel="license"
>   href="http://creativecommons.org/licenses/by-sa/3.0/us/">
> <img alt="Creative Commons License" style="border-width:0"
>  src="http://i.creativecommons.org/l/by-sa/3.0/us/88x31.png" />
> </a>
> <br />
> <span xmlns:dc="http://purl.org/dc/elements/1.1/"
>      property="dc:title">
>  HTML5 Example
> </span>
> by
> <a xmlns:cc="http://creativecommons.org/ns#" href="http://ben.adida.net 
> "
> property="cc:attributionName" rel="cc:attributionURL">Ben Adida</a> is
> licensed under a
> <a rel="license"
>   href="http://creativecommons.org/licenses/by-sa/3.0/us/">
>  Creative Commons Attribution-Share Alike 3.0 United States License
> </a>.
> =========
>
> One chunk of HTML. Just paste it on your site. Don't worry about the
> details. Done.
>
> It would be a very different story if I had to paste something in the
> HEAD and something in the BODY, or something in a separate file.

Isn't the whole point of splitting URIs into two and introducing  
syntax into later putting the part back together that the length of  
the URI gets amortized when the same prefix is used many times even  
though in the case of a single occurrence, the indirection syntax  
actually makes things *longer*? Putting the prefix declaration and the  
CURIE on the same element to make then travel together when copied and  
pasted seems the defeat the entire point of CURIEs. OTOH, if you  
actually want to make things shorter in the amortized sense, you need  
to put the declarations on (or near) the root of the document, at  
which point the syntax is no longer robust under copy and paste.

This would be more robust under copy and paste:

<a rel="license"
   href="http://creativecommons.org/licenses/by-sa/3.0/us/">
<img alt="Creative Commons License" style="border-width:0"
  src="http://i.creativecommons.org/l/by-sa/3.0/us/88x31.png" />
</a>
<br />
<span
      property="http://purl.org/dc/elements/1.1/title">
  HTML5 Example
</span>
by
<a href="http://ben.adida.net"
property="http://creativecommons.org/ns#attributionName" rel="http://creativecommons.org/ns#attributionURL 
">Ben Adida</a> is
licensed under a
<a rel="license"
   href="http://creativecommons.org/licenses/by-sa/3.0/us/">
  Creative Commons Attribution-Share Alike 3.0 United States License
</a>.

This would be more HTML-like:

<a rel="license"
   href="http://creativecommons.org/licenses/by-sa/3.0/us/">
<img alt="Creative Commons License" style="border-width:0"
  src="http://i.creativecommons.org/l/by-sa/3.0/us/88x31.png" />
</a>
<br>
<span
      property="title">
  HTML5 Example
</span>
by
<a href="http://ben.adida.net"
property="cc-attribution-name" rel="cc-attribution-url">Ben Adida</a> is
licensed under a
<a rel="license"
   href="http://creativecommons.org/licenses/by-sa/3.0/us/">
  Creative Commons Attribution-Share Alike 3.0 (United States) License
</a>.

Well, doing s/property/class/ would be even more HTML-like, but people  
didn't like the idea of a class name registry.

(Aside: The alt values should probably be "Creative Commons  
Attribution-ShareAlike 3.0 US License".)

>> (In fact, the declaration part of CURIE syntax has already been
>> forgotten in examples sent to this mailing list.)
>
> I was just trying to make the examples more concise for the purpose of
> discussion. I assumed that I didn't need to repeat everything in each
> email. I suspect discussions of SQL-in-browser on this list don't
> include complete data models for example SQL queries, either.

SQL in browser is different from RDFa. When at least two out of Gecko,  
WebKit, Opera, IE and Gears teams are determined to ship a feature,  
Hixie can't stop them, but Hixie can step in and steer them to do what  
they are determined to do in such a way that the implementations  
interoperate. In this case, at least two (three?) out of those five  
indicated that they were going to expose SQLite to JavaScript.

-- 
Henri Sivonen
hsivonen at iki.fi
http://hsivonen.iki.fi/