[uf-new] PROPOSAL: Replace hAudio FN with TITLE

[Manu Sporny]

Walter Logeman wrote:
> The question concerning FN & TITLE (and FN as well) has an underlying
> more general question.
> 
> Are Microformats broken when terms are repeated across different formats?

"broken" isn't the right word in this case, "not scalable" is closer to
the mark. They are broken if we don't know what we are doing and this
happened accidentally. The reality is quite the contrary, this trade-off
was made knowing that there would be situations like this.

It prevents Microformats from scaling... but those in the community know
that as well. We're trying to create solutions that work for most people
now... not in some hypothetical future that is going to mark up
everything on a web page with various semantic vocabularies (that's what
RDFa is for).

> How is this answered?  By a rule?  Is there one?  Should there be one?

In general, this community is averse to rules that are rigid... they
don't allow the community to grow with the rapid changes of the web.
That's why we have so much debate here - we aren't afraid to question
what we're doing. That is the sign of a healthy community - can it be
critical of itself and not devolve into name calling and
pseudo-political in-fighting?

We try to focus on: "Here's the semantically right way to do things (if
you're interested in doing them the right way)". Some just aren't in
some cases.

> The main reason to allow repeating terms is proximity to natural use
> of language, human friendly - that is a Microformat principle.

Instead of using the word "repeating", I think you mean "reusing"?

> The main reason for not repeating terms is that it breaks
> Microformats. Machine friendly.

Again, not "repeating", "reusing". It doesn't always "break"
Microformats... it causes greater difficulty when marking data up that
overlaps.

This may be more fair:

As the number of Microformats increases, and as the rate of re-use of
terms increases, the likely hood that two Microformats will clash with
each other on a single page increases.

WHEN this clash occurs is a known unknown. Brian Suda is concerned
(please correct me if I'm wrong, Brian) that hAudio TITLE may be the
first step towards a clash-heavy future.

We have several principles that are at odds with one another (which is
not a bad thing, it forces us to make hard decisions):

1. Microformats should re-use terms from other Microformats.
2. Microformats should not use fully qualified namespaces.
3. Microformats should use emulated namespaces as a last resort.
4. Microformats have one scope, either you're in a Microformatted block,
   or you're not. (For example, TITLE is inside an HAUDIO or it is not).

#1 is at odds with #2 and #3. #2 is what allows vocabularies to scale
(and is what RDFa allows). #3 is a stop-gap measure when you don't have
the option of #2. #4 makes scaling more difficult as well.

> However the question here is not one of priorities, it is more that
> Microformats would not work at all then they would not be people
> friendly either!

While true, Microformats also has a principle to try and address 80% of
publishing practices out there. We will never meet everybody's
publishing goals - we're not trying to do that. You'll see people
talking about "boiling the oceans" from time to time on here and this is
what they mean by that - we can't solve everybody's publishing problems.

Because we are using markup that is limited, there will always be corner
cases that won't be addressable using Microformats. The reason we use
simple markup is to help people publish semantic information in a much
easier fashion.

> In either approach regarding the repetition (or not) of terms, the
> guidelines for developing parsers need to make it clear how the format
> should be parsed.  Are there such guidelines?  

They are here and there, sprinkled throughout the wiki and buried deep
inside the heads of people that have been with this community from the
beginning.

Really, Mike Kaply (the guy that wrote and maintains Operator) is one of
the few that really knows what all of the rules are... Tantek being the
other one. Is there anyone else?

It would be nice if we could create a "Microformats Processing Rules"
document or a "Microformats Syntax Specification" document. RDFa has one
and while it is ridiculously verbose, it's hard to misinterpret:

http://www.w3.org/MarkUp/2008/ED-rdfa-syntax-20080125/

> Any term used in Microformats is unique and can be correctly rendered
> into a specific use.  For example:  The term "title" always refers to
> the job title of a person.

This is what I'm currently arguing against.

> Terms used in Microformats can be correctly rendered into a specific
> use by reference to the class they are nested in.  For example:  The
> term "title" refers to the job title of a person in vCard, and to the
> audio recording title in hAudio.

This is what I'm currently arguing for, albeit - we still need to look
at these things on a case-by-case basis. But the general guideline that
you wrote above should hold. Ideally, we would note that the Oxford
English dictionary would be the document determining the possible
definitions of a word and precise meaning would be acquired by examining
the context of the word (hCard or hAudio).

> Does my summary make sense?

It does... not bad at all for somebody that just joined the community
and claims to have no formal background in these linguistics/computer
sciencey things :)

... or are you a spy sent by a rival semantic community?* =P

-- manu

* as if there were such a thing... :)