<SimonR> Unless we have special knowledge (D-entailment) I don't see that named resources and datatyped literals act differently. Identical names indicate identical references; different names don't prove nonidentity of referents.

Well, identical lexical forms don't indicate indentical references

"001"^^xsd:string, "01"^^xsd:integer

And yes, you need special knowledge

But there is that special knowledge available

Hence my worry that such simplifications complicate the overall picture

<SimonR> I'm considering the `name' in the case of a datatyped literal to comprise both the lexical form and type IRI.

well,t hat's a bit of an arbitrary choice

Why not the triple of lexical form, type URI and value?

<SimonR> Well, I see it as a much simpler picture. You have two types of atomic entitiy, and they're differentied by whether or not you can determine inequality of reference by inequality of name.

Plus, I don't see the simplification. We are going to want those dataype theories

So, there's a saying about simplicity

Simple *with respect to what*

So, we can reduce all propositional first order connective to ~ and v or even just the sheffer stroke

And that is way simpler

One truth table!

But the formula are brutal

So you trade one kind of simplicity for another

So, let me just grant that your variant is simpler in some respect

is taht a useful respect?

What does it help us do?

What does it, at least, clarify

<SimonR> This is how it would be useful -- it's entirely clear to everyone that unicode strings (in Normal Form C, just for pedantry's sake) are equal or different. It's never clear whether any of the four different RDF naming schemes are equal or not. Not *one* of them.

Well, ok, but thus far it gets things wrong, afaict

do bndoes have multiple nodeIDs?

Just one?

<SimonR> How so?

what affect does it have if two different bndoes have the same nodeID?

effect

<SimonR> Yes, the same entity might be labeled with different bnode IDs in various serializations.

But you didn't say, "various serializations"

you made it a property of the node

So is relative to a serialization/ Is that encoded in the graph too?

<SimonR> Right. That property is what you represent the scope with.

How?

It's very unclear

<kendallclark> zakim agenda+ track action items

"""Because there are only two things in the universe now, existential variables whose equality behavior is consistent (known equal, or not known equal) and the set of unicode strings (known equal, or known different)."""

<SimonR> If my database invents a new property for each transaction, that quite accurates indicates when it's okay to share bnode IDs and when it isn't.

This doesn't handle that differenent URIs with the same denotation

<SimonR> accurate+ly

But do these get serialized?

I mean,a re they normal propreties or not?

If so, then you have a bit of regress problem

And I dont' see this is easier than normal lexical approaches

I mean, my prolbem iwth pat's last email is that it means that RDF *is* quite non-standard

I don't find adding even *more* nonstandardness clarifying

I like reading textbooks :)

<SimonR> Yes, regress is a problem. I think there has to be at least one distinguised property to break that regress. It should probably be the property that associates the IRI of the named resource with the string of that IRI.

I don't feel the simplicity joy here

I dread the proofs

Well, you can produce the pudding

That contains the proof

<SimonR> It'd be easier written out.

Rewrite some part of the rdf semantics along these lines

Simple interpretations, for example

heck just do ground graphs

To start

<AndyS> Kendall, the URL spliting issue.

<kendallclark> andy?

<SimonR> It'd end up as a layer below simple interpretation. Then you'd layer simple interpretation with all RDFs different naming schemes, and then RDF interpretation on top of that.

<AndyS> If the emails I receive are anything to go by, it might be because the email sent has long lines + "format=flowed"

<AndyS> The agenda has some long text lines split in strange places. Might be related.

simonr, so, if its' going to simplify, it should simplify those layers

Hack at it!

<kendallclark> ah, yeah, it's so weird how they split differently this time...

<SimonR> Shall do. :)

<kendallclark> and i don't think i have per-header control over Mail.app :(

<AndyS> Something has changed - they didn't used to get split. Upgrade recently?

<kendallclark> no, actually...

Hey, can we put a little time on heartbeat issues/doc status? I take it we're ok becasue we've pubbed *something* yes?

<LeeF> PatH

<patH> it was me, but Im having cell phone trouyble.

<patH> thanx.

<kendallclark> Regrets from FredZ

<AndyS> Regrets

<kendallclark> http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0117.html

<AndyS> for next week from me

<kendallclark> PROPOSED to approve those minutes, 2nd, so resolved

<LeeF> RESOLVED

<kendallclark> ACTION: [NEW] * bijan to show that the "strong" version of DISTNCT [recorded in http://www.w3.org/2006/08/22-dawg-minutes.html#action01]

<kendallclark> doesn't interfere with intermittent algebraic operations [recorded in

<kendallclark> http://www.w3.org/2006/08/15-dawg-minutes.html#action01]

<LeeF> ACTION: bijan to show that the "strong" version of DISTNCT [DONE] [recorded in http://www.w3.org/2006/08/22-dawg-minutes.html#action02]

<kendallclark> ACTION: [NEW] * Bijan to describe reduction algorithm [recorded in [recorded in http://www.w3.org/2006/08/22-dawg-minutes.html#action03]

<kendallclark> http://www.w3.org/2006/08/15-dawg-minutes.html#action02]

track action items

<AndyS> This one? http://lists.w3.org/Archives/Public/public-rdf-dawg/2006JulSep/0151.html

<LeeF> ACTION: Bijan to describe reduction algorithm [DONE] [recorded in http://www.w3.org/2006/08/22-dawg-minutes.html#action04]

<kendallclark> ACTION: [NEW] * EricP to redraft section 11 to support extensible [recorded in http://www.w3.org/2006/08/22-dawg-minutes.html#action05]

<kendallclark> datatypes [recorded in

<kendallclark> http://www.w3.org/2006/08/08-dawg-minutes.html#action08]

Andy, yes

Thx

approve new value tests

<ericP> head of extensibility thread

<ericP> head of extensibility thread

<kendallclark> EricP summarizes the monotonicity and datatypes discussion that's ongoing

<LeeF> ericP: to preserve monotonicity in the face of unknown datatypes, inequality tests involving unknown datatypes should be errors

<AndyS> PatH: can say two things are equal but not conclude two other things are unequal (unknown types).

<kendallclark> did you hear me, andy?

<ericP> higher level write-up of trade-offs in dwag:mon

<AndyS> ACTION: AndyS to Write some tests for value testing (unknown types and extensibility) to add to 2006/JulSep0086 [recorded in http://www.w3.org/2006/08/22-dawg-minutes.html#action06]

LeeF's review of rq24

<AndyS> This is rq23 vs rq24.

<AndyS> That's less than full review

<ericP> Tucanna is not on the list

<AndyS> SimonR is not in the DB records.

<kendallclark> ACTION: SimonR to review rq24 [recorded in http://www.w3.org/2006/08/22-dawg-minutes.html#action07]

<AndyS> rq24 is http://www.w3.org/2001/sw/DataAccess/rq23/rq24.html

<AndyS> rq23 is http://www.w3.org/2001/sw/DataAccess/rq23/

datatype and lang tests

<kendallclark> it's good for you :

<kendallclark> )

<AndyS> example: datatype("plain string")

<AndyS> c.f. RDF MT rules xsd1a and xsd1b

<kendallclark> ADJOURNED

<kendallclark> EricP: can you do the pub minutes magic?