Re: RDFa Last Call Comment: garbage collecting "useless" triples doesn't seem necessary and could hurt authors as they write RDFa

Hi Ben,

I agree that it is not necessary, although I'm not so sure that it's harmful.

Either way, this part of the processing doesn't actually add much
complexity to the parsing rules. To achieve this 'fix', all I did was
to move the step that completes the hanging triples, so that it
occurred _after_ the recursion, and then I added a flag to indicate
whether triples had been added or not during the recursion. The
incomplete triples were then only added if the flag was set.

There _were_ some changes in the last couple of months that added
complexity, but they originated in trying to solve genuine flaws that
were spotted by eagle-eyed reviewers, and implementers, and not an
attempt to address this question.

Which is not to say that we couldn't change this if the group decides
to do so. But since this would be a substantive change to the parsing
rules, I would guess that it would probably bounce us back through
last call again. If that is the case I really hope that the group
seeks really strong arguments for why it would be worthwhile to do
that, before accepting the change.

Regards,

Mark

Ben wrote:
>  I struggled for a while regarding whether to bring up this issue as a
>  Last Call Comment, but I do think it's worth considering.
>
>  I don't believe it is necessary, and it may be harmful, to garbage
>  collect so-called "useless" triples in cases like:
>
>  <div rel="foaf:knows">
>    <div rel="foaf:knows">
>      <div rel="foaf:knows">
>        ...
>      </div>
>    </div>
>  </div>
>
>  So what if someone writes useless triples? That's their intent, so let
>  it be :) This step in the processing adds notable complexity to the
>  parser rules. It also may make life difficult for authors who are
>  progressively writing RDFa, building up a skeleton of bnodes before
>  hanging properties onto them: they wouldn't see any triples at all until
>  they add a @property, at which point a slew of triples would appear.
>
>  There's a notable cost here, and questionable value.
>
>
>  -Ben
>
>


-- 
  Mark Birbeck

  mark.birbeck@x-port.net | +44 (0) 20 7689 9232
  http://www.x-port.net | http://internet-apps.blogspot.com

  x-port.net Ltd. is registered in England and Wales, number 03730711
  The registered office is at:

    2nd Floor
    Titchfield House
    69-85 Tabernacle Street
    London
    EC2A 4RR

Received on Thursday, 20 March 2008 10:37:22 UTC