Re: Metadata for all and PS for metadata (Was Re: where to hang the metadata?)

Michael Kifer wrote:
>>
>>Just for clarification: what would be the argument against allowing 
>>metadata on all objects, that is, allowing a <meta> role element as a 
>>child of any class (capitalized tag) element (this question is othogonal 
>>to that of identifying rules vs groups)?
> 
> See my answer to Sandro a few mins ago.

But your arguments is that metadata everywhere would bloat the PS beyond 
all recognition: what if we do not specify a PS for metadata, except in 
some specific cases (e.g. groups)?

> Metadata *may* affect the semantics, although I do not know yet how to
> express it in FLD. There is a whole class of useful dialects based on
> prioritized logics (one is courteous LP) where rule labels and other
> non-rule info are used in defining the semantics.

So, FLD would have to specify a PS for metadata on everything. But FLD 
being a framework, that does not mean that any specific dialect based on 
FLD would have a PS for metadata but for a few specific constructs. So, 
no concrete syntax being absurdely bloated, right?

And, anyway, in a dialect where some information, say: priorities, has 
an impact on the semantics, that iformation is part of the rule, and, 
thus, not meta-data. So that even FLD would not have to specify a PS for 
metadata on everything.

(Note to self: seems too easy; I probably missed something)

> Do you mean to standardize some of the attributes of the meta?

Yes, this is what I meant.

> It would be
> useful, although I am afraid we will not be done any time soon with this
> given the amount of heat that even seemingly simple issues tend to
> generate.

Maybe, or maybe not :-)

Maybe we could just have a quick poll on which meta-data everyone would 
like to have standardized, and maybe we would see that there is a 
obvious subset of the answers on which nobody objects?

Christian (always the optimist :-)

Received on Monday, 28 April 2008 16:47:53 UTC