Re: Quick glance through xhtml2 meatainformation module

Hi Mark,

"Mark Birbeck" <mark.birbeck@x-port.net> wrote:

> All that 20.4 says is that these two statements can be 'chained', in that
> the second triple has as its subject the object of the first triple:
>
>   <link rel="a:a" resource="R1">
>     <meta property="a:b">L1</meta>
>   </link>

I think it's just a minor detail, but my remark was that the text seems
to say that only links can be contained within other links. Which as the
example shows isn't true as you can have meta.

>> "Both the rel and rev attributes may be specified simultaneously."
>> 
>> This puzzles me very much. In the case where one writes rel="A" 
>> rev="B" doesn't it imply that A and B are necessarily Functional
>> Inverse Properties?
>
> No. If A has a property a:a, which has a value of R1, there is nothing to
> stop R1 having a property a:b, which has a value of A. Unrequited love seems
> a good example ;)

Good example indeed, I understand much better now...

> It's not entirely necessary, but it does make certain constructions easier
> to write.

It obviously does, and I'll change my parser to take it into account.
One thing makes me uncomfortable when I think of how Im going to
implement it, though: Each <link> (and <meta>) element produces one
triple, except for the case when there's both an rel and rev
attributes on <link> At first glance this exception doesn't seem to
create problems when transforming to RDF, but it's still an exception
to an otherwise clean model.

Cheers,

Max.

Received on Wednesday, 11 August 2004 16:54:23 UTC