IRC log of sw-meaning on 2003-10-10

Timestamps are in UTC.

14:37:06 [RRSAgent]
RRSAgent has joined #sw-meaning
14:37:10 [sandro]
RRSAgent, pointer
14:37:10 [RRSAgent]
See http://www.w3.org/2003/10/10-sw-meaning-irc#T14-37-10
14:37:21 [sandro]
zakim, this with be sw_meaning
14:37:21 [Zakim]
I don't understand 'this with be sw_meaning', sandro
14:37:27 [sandro]
zakim, this will be sw_meaning
14:37:27 [Zakim]
ok, sandro; I see SW_Meaning()11:30AM scheduled to start in 53 minutes
15:04:48 [gk]
gk has joined #sw-meaning
15:04:57 [gk]
gk has left #sw-meaning
15:22:14 [gk]
gk has joined #sw-meaning
15:26:12 [DanC]
DanC has joined #sw-meaning
15:27:17 [DanC]
DanC has changed the topic to: sw-meaning 10 Oct http://lists.w3.org/Archives/Public/public-sw-meaning/
15:28:45 [JohnBlack]
JohnBlack has joined #sw-meaning
15:29:06 [Zakim]
SW_Meaning()11:30AM has now started
15:29:15 [Zakim]
+Sandro
15:29:17 [Zakim]
+[UMD]
15:29:38 [pfps]
pfps has joined #sw-meaning
15:29:40 [Zakim]
+GrahamKlyne
15:29:40 [sandro]
zakim, [UMD] is Bijan
15:29:41 [Zakim]
+Bijan; got it
15:30:01 [Zakim]
+John_Black
15:30:09 [bijan]
bijan has joined #sw-meaning
15:30:11 [Zakim]
+??P40
15:30:24 [pfps]
zakim, ??P40 is pfps
15:30:24 [Zakim]
+pfps; got it
15:30:45 [sandro]
zakim, who is on the phone?
15:30:45 [Zakim]
On the phone I see Sandro, Bijan, GrahamKlyne, John_Black, pfps
15:31:47 [Stuart]
Stuart has joined #sw-meaning
15:32:30 [Zakim]
+??P42
15:32:59 [sandro]
Zakim, ??P42 is Stuart
15:32:59 [Zakim]
+Stuart; got it
15:33:28 [tim-mit]
tim-mit has joined #sw-meaning
15:33:57 [Zakim]
+TimBL
15:33:59 [DanC]
Zakim, agenda?
15:33:59 [Zakim]
I see nothing on the agenda
15:34:14 [sandro]
zakim, who is here?
15:34:14 [Zakim]
On the phone I see Sandro, Bijan, GrahamKlyne, John_Black, pfps, Stuart, TimBL
15:34:16 [Zakim]
On IRC I see tim-mit, Stuart, bijan, pfps, JohnBlack, DanC, gk-scr, RRSAgent, Zakim, sandro
15:34:24 [DanC]
agenda + A admin http://lists.w3.org/Archives/Public/public-sw-meaning/2003Oct/0042.html
15:34:35 [DanC]
agenda + B. Tim and Bijan discuss "Tim's View"
15:34:44 [Zakim]
+DanC
15:35:28 [DanC]
Zakim, take up agendum 1
15:35:28 [Zakim]
agendum 1. "A admin http://lists.w3.org/Archives/Public/public-sw-meaning/2003Oct/0042.html" taken up [from DanC]
15:35:38 [sandro]
Zakim, who is here?
15:35:38 [Zakim]
On the phone I see Sandro, Bijan, GrahamKlyne, John_Black (muted), pfps, Stuart, TimBL, DanC
15:35:40 [Zakim]
On IRC I see tim-mit, Stuart, bijan, pfps, JohnBlack, DanC, gk-scr, RRSAgent, Zakim, sandro
15:35:55 [gk-scr]
Logictics...
15:36:18 [gk-scr]
(mission impossible?)
15:36:55 [sandro]
(2nd hand regrets for Norm)
15:36:59 [gk-scr]
Present as above. Stuart present regrets for Norm Walsh.
15:37:49 [gk-scr]
Proposed to accept record from last week, per agenda:
15:37:50 [sandro]
PROPOSED: Accept http://www.w3.org/2003/09/26-sw-meaning-irc
15:37:50 [sandro]
as a true record of the last meeting.
15:37:58 [sandro]
TimBL states he hasn't read it
15:38:00 [sandro]
RESOLVED
15:38:02 [gk-scr]
No objections.
15:38:02 [tim-mit]
I have not read it
15:38:21 [Zakim]
+PatH
15:38:34 [gk-scr]
Next meeting October 31, same time
15:38:41 [gk-scr]
(per agenda)
15:39:06 [gk-scr]
At that time, to discuss whether or not to hold future meetings
15:39:12 [Zakim]
+Mike_Dean
15:39:13 [gk-scr]
Bijan proposes agenda change
15:39:51 [sandro]
Bijan: oringinally I thought I'd listen to Tim's explication and interrupt, but now I'm thinking talking points, etc.
15:40:18 [gk-scr]
Proposes an approach to elicit Tims view more clearly: Bijan's test cases posted this morning, Peter patel-Schneider's respnse.
15:40:56 [mdean]
mdean has joined #sw-meaning
15:41:51 [sandro]
Graham: (thought we were doing broader/different issue)
15:41:56 [gk-scr]
GK raises distributed inferencing vs broader issues of social context of meaning
15:42:05 [DanC]
31 Oct is in my calendar. http://calendar.sidekick.dngr.com/event?id=1096 (for my reference; URI is kinda wierd for use by others)
15:42:07 [sandro]
Bijan, Tim: let's focus on the narrower form
15:42:13 [gk-scr]
Response: certain issues in specs need to be addressed first
15:43:09 [gk-scr]
GK is happy to let my commetn ride w.r.t. agenda
15:43:20 [sandro]
zakim, close this agendum
15:43:20 [Zakim]
agendum 1 closed
15:43:21 [Zakim]
I see 1 item remaining on the agenda:
15:43:21 [tim-mit]
http://lists.w3.org/Archives/Public/public-sw-meaning/2003Oct/0059.html
15:43:22 [Zakim]
2. B. Tim and Bijan discuss "Tim's View" [from DanC]
15:43:27 [sandro]
zakim, take up next agendum
15:43:27 [Zakim]
agendum 2. "B. Tim and Bijan discuss "Tim's View"" taken up [from DanC]
15:43:31 [gk-scr]
Talking points are:
15:43:44 [gk-scr]
1. ontollogical commitment not required?
15:44:01 [sandro]
Tim: I don't know what "required" means here
15:44:02 [gk-scr]
TimBL: what does "required" mean here?
15:44:43 [gk-scr]
... e.g. for a compilant parser?
15:45:04 [gk-scr]
Tim: RDF is a language ... There are very many ways to process it.
15:45:18 [gk-scr]
Bijan: but you seem to say we need to nail down some more?
15:45:57 [gk-scr]
Do we need more specification text about ontological commitment?
15:46:22 [gk-scr]
e.g. Is using a URI equivalent to dereferencing it and taking what you get into the graph?
15:46:53 [DanC]
I have *no* expectation that this forum will have any impact on the RDF and OWL documents currently on the REC track.
15:47:10 [gk-scr]
Bijan: If I use a URI "inconsistently", does that make my document inconsistent?
15:47:54 [gk-scr]
TimBL ... consistency of closure over ontologies used in document ... (?)
15:48:57 [gk-scr]
Tim: (a) step that parsed graph is consistent; (b) step that applications MUST check for consistency (?)
15:49:24 [sandro]
Pat: documents can be consistent without actually being merged
15:50:10 [gk-scr]
Bijan: what must I (my software) do operationally?
15:50:31 [gk-scr]
(all:) Only one definition of consistency
15:51:20 [gk-scr]
TimBL: (a) document consistency ... facts are [true?], (b) set of documents, refernced ontologies, etc are true
15:51:36 [gk-scr]
PatH: these are trhe same noytion opf consistency, just over different graphs.
15:51:53 [bijan]
Pat channels me :)
15:51:59 [sandro]
Tim: two different properties of a document
15:52:06 [gk-scr]
TimBL: didn';t say different notions of consistency, just different properties of document
15:52:15 [gk-scr]
PatH: OK
15:52:42 [gk-scr]
Bijan(?): not happy. Not two graphs here. RDF document does not encode the larger graph.
15:52:54 [gk-scr]
TimBL: not encode...
15:53:14 [gk-scr]
... graph G has "ontological closure" H
15:53:35 [tim-mit]
d encodes G
15:53:40 [tim-mit]
G ontologicalClude H
15:53:49 [tim-mit]
G ontologicalClose H
15:53:56 [tim-mit]
q?
15:54:06 [gk-scr]
Bijan: so OWL allows relationship between documents and graphs that is "larger" than RDF(S)
15:54:22 [DanC]
ack danc
15:54:22 [Zakim]
DanC, you wanted to confirm that timbl isn't proposing any changes to the definitions of "conforming RDF document" or to RDF simple entailment
15:54:56 [Zakim]
+??P48
15:55:00 [gk-scr]
DanC: Wishes to clarify no changes to copnforming document or RDF entailment
15:55:27 [sandro]
Zakim, ??P48 is JimH
15:55:27 [Zakim]
+JimH; got it
15:55:32 [sandro]
zakim, who is here?
15:55:32 [Zakim]
On the phone I see Sandro, Bijan, GrahamKlyne, John_Black (muted), pfps, Stuart, TimBL, DanC, PatH, Mike_Dean, JimH
15:55:34 [Zakim]
On IRC I see mdean, tim-mit, Stuart, bijan, pfps, JohnBlack, DanC, gk-scr, RRSAgent, Zakim, sandro
15:55:55 [gk-scr]
Stuart(?): interested in what will be asked of TAG
15:56:23 [sandro]
DanC: Bijan, please try to use terms from current Rec Track documents, when talking about those concepts
15:56:48 [gk-scr]
Do we want to add any notions [to RDF spec] of "conforming RDF++ document", or change existing notions?
15:56:53 [sandro]
Tim: What I want to added in the RDF world is a little hook.....
15:57:04 [gk-scr]
TimBL: wnats a hook to [something?]
15:57:18 [gk-scr]
Bijan: what's missing?
15:57:44 [DanC]
tim, though you may not care what document it goes in, Bijan, among others, has more sensitivity about proposals to change the existing REC-track documents than to future work.
15:57:46 [gk-scr]
Bijan: I think owl:imports is already with us. What more is needed?
15:57:48 [sandro]
Bijan: I think owl:imports might be the hook? Or demonstrates that a random WG can add stuff like that.
15:58:07 [bijan]
It's not a hook, it's an exmample of why the "hook" is already there
15:58:20 [gk-scr]
TimBL: need hook between RDF and other things [not specifically own].
15:58:39 [tim-mit]
:ancestor rdf:type owl:TransitiveProperty.
15:58:48 [gk-scr]
s/own/OWL/
15:59:22 [gk-scr]
TimBL: suppose I send document with above, appropriately labelled...
15:59:40 [gk-scr]
... now what are you supposed to deduce from that?
16:00:00 [gk-scr]
Bijan(?): probably nothing?
16:00:10 [tim-mit]
What doe sit mean?
16:00:17 [sandro]
Bijan: (interjects: be careful of protocol-ness there)
16:00:29 [gk-scr]
... but what *can* an RDF compliant agent, non-application-specific, offer me?
16:01:12 [gk-scr]
Bijan: how do we avoid ending up at cross purposes, due to different ways of "assigning meaning"
16:01:17 [tim-mit]
- things we expected it be consisetnt with
16:01:22 [tim-mit]
- things we can translate it into
16:01:36 [gk-scr]
... if the meaning entails[?] more than the bare triples in the document
16:01:55 [DanC]
EEK
16:02:09 [gk-scr]
TimBL: consider language tranbslation (e.g. English, French ...)
16:02:25 [DanC]
translating RDF to english as an excercise in exploring its meaning? doesn't appeal to me at all.
16:02:54 [gk-scr]
Bijan: first step in understanding the sentence meaning is figuringh out the meaning according to formal conventions of the language ... that ground the languiage ... 1st stop is the model theory.
16:03:13 [sandro]
Bijan: first, before looking at speakers/intended meaning, I look at sentence meaning, and for that I look at the MT.
16:03:18 [sandro]
Tim: Which MT?
16:03:30 [sandro]
Bijan: well, since RDF and OWL don't have differne mime types.....
16:03:30 [tim-mit]
applciation/rdf+xml
16:03:37 [gk-scr]
TimBL: start withj MIME type ... e.g. application/rdf+xml
16:03:55 [gk-scr]
Bijan: I must decide how to interpret that ... RDF, or OWL?
16:03:59 [sandro]
Bijan: I need to decided which semantics to use, RDF or OWL
16:04:23 [sandro]
Bijan: I'd probably use OWL-DL semantics (which is close to OWL-Full but different)
16:04:25 [gk-scr]
Bijan: I would choose (somefuncof(OWL))
16:05:01 [sandro]
Tim: Why should you pick OWL? Maybe I have some semantics wherein carrots are orange.
16:05:20 [sandro]
Bijan: Because that's the semantics I use. I have toolls for it. I like it.
16:05:45 [sandro]
Bijan: if you send me some python code, I wont use OWL semantics.
16:05:53 [sandro]
Tim: Right, you'll key off the mime type.
16:06:07 [sandro]
Bijan: No, I'm taking it all as OWL. No RDF.
16:06:24 [DanC]
[scribe notes being lost at this point]
16:06:48 [gk-scr]
Bijan: Our demo uses heuruistics in deciding what to use (?)
16:06:54 [sandro]
Bijan: The MINDLAB demo OWL-reasoner assumes everything is OWL. That's the kind of thing we do.
16:07:22 [sandro]
Tim: In my view, the web arch works the other way around. You should look it up in the Mime-type registry.
16:07:24 [gk-scr]
TimBL: web architecture, see application/rdf+xml, go look up in registry, ...
16:07:40 [gk-scr]
... without prior knowledge, go to the RDF spec ...
16:08:06 [sandro]
Bijan: I've heard this from you many, many, many times.
16:08:11 [gk-scr]
Bijan: I've heard all this... many times ... agree that's your view ...
16:08:20 [gk-scr]
... doesn't help to simply repeat it.
16:08:26 [sandro]
Bijan: I think it's odious.
16:08:42 [gk-scr]
... "I think it's odious" ... it doesn't justify changing my behaviour, or changing any spec
16:09:05 [sandro]
Bijan: It doesnt even remotely justify anything. I need something else
16:09:12 [gk-scr]
DanC: if our target is to present to TAG ... (?)
16:09:35 [tim-mit]
"Odius" is a strong word.
16:09:36 [gk-scr]
ABijan: are we doing something incorrect?
16:09:52 [bijan]
It's not "odious
16:09:53 [gk-scr]
DanC: I think not inconsistent with RDF MIME type.
16:09:56 [bijan]
It's Otiose
16:10:03 [sandro]
"odious: Arousing or meriting strong dislike, aversion, or intense displeasure. See Synonyms at hateful."
16:10:03 [bijan]
I.e., irrelevant, not nasty
16:10:08 [gk-scr]
TimVL: does this depenmd on OWL being consistent with RDF?
16:10:11 [sandro]
Ohhhhhh,
16:10:13 [gk-scr]
DanC: can't answer that
16:10:38 [DanC]
the "spec stack" is encorsed, to some extent, in the TAG finding "Client handling of MIME headers" http://www.w3.org/2001/tag/doc/mime-respect.html
16:10:44 [sandro]
"otiose: 1. Lazy; indolent.
16:10:44 [sandro]
2. Of no use.
16:10:44 [sandro]
3. Ineffective; futile. See Synonyms at vain.
16:10:44 [sandro]
"
16:10:59 [gk-scr]
Bojan... if the architecture requires us put put in spec things that are mistake, I must push back. I'm not getting the connection.
16:11:24 [gk-scr]
TimBL: your commuynity agree to use OWL
16:11:35 [gk-scr]
Bijan: we do more than that!
16:11:44 [sandro]
Bijan: We have a DAML-to-OIL converter. We do lots of stuff with RDF.
16:12:03 [sandro]
Bijan: you're ... making big accusations.
16:12:34 [gk-scr]
TimBL: if we try to allow newcomers to bootstrap, in the archiecture, there are further requirements to be addressed.
16:12:40 [sandro]
Tim: If we want to create a bootstrapping RDF architecture, then we have additional requirements. You may not want that.
16:12:52 [sandro]
Bijan: Is that bootstrapping part of WebArch?
16:13:34 [sandro]
Bijan: Must an application do some stuff with the web (look some stuff up) ... [ not actually download an owl reasoner? ] ....
16:13:41 [gk-scr]
Bijan: programs with no knwoeldge of OWL can use the web ... to do what my OWL reasonber does?
16:13:48 [sandro]
Bijan: Is it a goal to make that not just possible, but easy?
16:13:53 [gk-scr]
... what is the goal?
16:14:06 [gk-scr]
DanC: yes , that's the goal
16:14:22 [sandro]
DanC: Let it be more than a program, let them download code. The goal is for parties to be able to get on the web and get in on the conversation.
16:14:28 [gk-scr]
Bijan: I don't yet know the consequences of those "relaxations"(?) ...
16:14:37 [Stuart]
My view wrt to mime-respect is that the sender of a representation declares what format that representation is in and it is *not* for the recipient to say otherwise.
16:14:40 [sandro]
RRSAgent, pointer?
16:14:40 [RRSAgent]
See http://www.w3.org/2003/10/10-sw-meaning-irc#T16-14-40
16:14:47 [tim-mit]
q+
16:14:49 [gk-scr]
... how to integrate with ways to specify meaning of documents
16:15:18 [sandro]
Pat: Dan, that sounds like you could enter the web without a web browser. Dont you NEED software to start up?
16:15:18 [gk-scr]
PatH: sounds like jointing the web without a web browser. You must have some software to use it.
16:15:43 [sandro]
DanC: General webarch is to make barrier to entry as low as possible, and let you look stuff up.
16:15:53 [gk-scr]
(?) Tim was asking good question... RDF MIMW type captures many different layers ...
16:15:59 [tim-mit]
q+ to say that the "meaning" stuff coems in beacsue although the we can't make everything work automatically and it takes a mixture of people and machines , we don need to be able to justify what somone ahs done when they have done it.
16:16:06 [sandro]
JimH: At the moment, where the RDF mime type is suppose to capture ....
16:16:27 [DanC]
low barrier to entry is an architectural principle of the web. very nearby is http://esw.w3.org/topic/FollowYourNose
16:16:39 [tim-mit]
why do we want self-describing documents? yes for boostrap, but also for interop.
16:16:55 [gk-scr]
How to decide which level; e.g. (a) all RDFS vocab, (b) some OWL vocal, (c) just HTML (different MIME type)
16:17:05 [sandro]
JimH: 2 rdf/xml documents, one pure RDFS, the other RDFS+OWL, 3rd is HTML. SHould you really assume everything is OWL?
16:17:10 [gk-scr]
... (JimH) how to decide which semantics to apply?
16:17:20 [sandro]
Bijan: Application Dependant Algorithm....
16:17:25 [sandro]
Great question Jim.
16:17:49 [sandro]
Tim: Yes, we're trying to self-describing
16:17:56 [gk-scr]
TimBL: you're right, we're aiming for self describing
16:18:11 [gk-scr]
... with minumum of prior context
16:18:11 [DanC]
hmm... interesting... timbl connects http://esw.w3.org/topic/SelfDescribingWeb to http://esw.w3.org/topic/FollowYourNose
16:18:43 [sandro]
Tim: (1) for bootstrapping, people reading specs, can build system, get on the web, go on from there
16:19:06 [gk-scr]
Bijan: what is dufference betwqeen specifications on web, and bootstrapping?
16:19:19 [sandro]
Bijan: What is the difference between having specs + publishing them on the web, and real self-description?
16:19:20 [gk-scr]
... the former requires intervening person.
16:19:37 [gk-scr]
... publishing human-readable docs makes web self-describing?
16:19:50 [gk-scr]
... that's nothing to me.
16:19:53 [sandro]
Bijan: that's not very interesting.
16:20:14 [gk-scr]
TimBL: Two reasons (a) idea of bootstrapping ....
16:20:40 [gk-scr]
... if you don't know, you can go find out about it. In last few years, people can go and write software ...
16:21:11 [gk-scr]
... algorithm, when seeing new MIME type, to find the spec and maybe find a plug-in, and deal with it.
16:21:15 [sandro]
Tim: auto-download of plugins to handle new mimetypes
16:21:29 [sandro]
Bijan: that almost never works for me.
16:21:39 [gk-scr]
TimBL: it works of a registry.
16:21:40 [sandro]
Bijan: Empirical matter
16:21:48 [gk-scr]
TimBL: second inte4resting reason...
16:22:22 [gk-scr]
... if someone takes the web page, and render it in a particular way, any dispute what it means can trace back ...
16:22:27 [sandro]
TimBL: Any dispute about correct rendering of a web page can be approach via the stack of specifications.
16:22:29 [gk-scr]
... useful for interop
16:22:36 [sandro]
Bijan: Whoa!
16:22:52 [gk-scr]
Jump from "interesting" to "interop" ...
16:23:03 [sandro]
Bijan: you made a huge jump to "this allows for interoperability". there might be lots of other ways.
16:23:31 [sandro]
TimBL: interop with minimum of prior agreement.
16:23:40 [gk-scr]
TimBL: guarantee of interop with *minimum* of prior agreement. Agree to use the web, can exchange URI.
16:24:03 [gk-scr]
... TimBL: most communication has communication out of band.
16:24:38 [gk-scr]
Bijan: but we just use English, (or whatrever)
16:25:13 [gk-scr]
TimBL: yes, lot's of OOB agreement involved with software (e.g. bank statement into Quicken)...
16:25:24 [sandro]
TimBL: protocol of downloading your bank statement to Quicken/Money, ... lots of agreements between parties around that.
16:25:36 [gk-scr]
... maybe also the agreement signed when setting up bank account.
16:26:17 [gk-scr]
Bijan: I read that stuff, I recall no agreement (with bank) to use any particular prorocol.
16:26:43 [sandro]
Tim: It's not in the transmitted message that you're using OWL
16:27:17 [DanC]
bijan earlier claimed a right to interrupt at will; tim neither disputed it nor seems to respect it. I find that a significant source of stress.
16:27:28 [sandro]
Tim: I can send you a URI and it's equivalent to sending some other message
16:27:48 [gk-scr]
TimBL: I can send you a URI, and it's (in some sense) equivalent to sending the (message) (what it points to?)
16:28:27 [gk-scr]
.. I can refer to document by URI. And I can be confident you will havce same understanding whjat it means.
16:28:46 [gk-scr]
Bijan: wait...
16:29:17 [sandro]
Bijan: I often see a URI with a wrong mime type, so I download, and load it up with a different mimetype.
16:29:31 [gk-scr]
.. I often see URI on web page, and it has wrong "MIME type", and load it to app that interprets it differently
16:29:43 [gk-scr]
.. I think your confidence is misplaced, or I'm missing soemthing
16:30:01 [sandro]
Bijan: there's arch and descr of arch
16:30:17 [gk-scr]
Bijan: is Web architecture and description of achitecture ...
16:30:31 [sandro]
DanC: We are attempted to describe what's on. There are two things:
16:30:39 [Stuart]
Bijan, when you look at a URI, how would you know the MIME type of the referent or indeed that there is only a single representation format available at that URI
16:30:43 [sandro]
DanC: 1. some robustness in the face of errors is good
16:30:47 [gk-scr]
DanC: rubistm=ness in face of errors is a Good Thing
16:30:59 [tim-mit]
We are being prescriptive. We do also describe some ways in which things breaks.
16:31:05 [sandro]
DanD: 2. a user can apply a different mime type, but then the publisher is no longer liable
16:31:13 [gk-scr]
2. I can't hold sender to conclusions if I change interpretation
16:31:23 [tim-mit]
q+
16:32:04 [gk-scr]
Suppose someone says "there's this RDF on my server" I download it and find it's postscript. What prevails?
16:32:11 [gk-scr]
?: then ther4e's been a bug.
16:32:17 [Stuart]
+1 to what Dan said, representation and reported mime-type may be inconsistent... that's an error.
16:32:18 [sandro]
ack tim
16:32:18 [Zakim]
tim-mit, you wanted to say that the "meaning" stuff coems in beacsue although the we can't make everything work automatically and it takes a mixture of people and machines , we don
16:32:21 [Zakim]
... need to be able to justify what somone ahs done when they have done it. and to
16:33:27 [gk-scr]
TimBL: I'm trying to describe a simple architecture, not one without holes in it.
16:34:04 [gk-scr]
.. and one that can work, in naive mode.
16:34:37 [gk-scr]
.. and the naive architecture can work in many situations, and produce (right?) results
16:34:44 [sandro]
Tim: A better protocol would have to behave like the the naive one would have had it not broken.
16:35:17 [gk-scr]
(architecture/protocol ... seemed to be used slightly interchangeably)
16:35:58 [gk-scr]
Bijan: distinguishes architecture vs architecture description. We may have naive descriptions of the architecture.
16:36:07 [sandro]
Bijan: the architecture != the naive description of the architecture
16:36:25 [sandro]
DanC: please don't redefine architecture
16:36:27 [gk-scr]
DanC: I don't feel bound by this use of "architecture"
16:36:45 [sandro]
Pat: use "idealized" instead of "naive"
16:36:55 [gk-scr]
TimBL: we use it to mean a description of set of constraints ...
16:36:56 [sandro]
DanC: Architecture is an idealization
16:37:56 [sandro]
[ I would love a return to Jim's question. Hrm. ]
16:38:15 [gk-scr]
TimBL: statistically, many systems conform to idealized architecture, and small deviations don't bring down the system.
16:39:16 [sandro]
DanC: Tim was describing properties of good architectures.
16:39:21 [gk-scr]
DanC: Tim describes properties good architectures, which are idealizations.
16:40:16 [gk-scr]
TimBL: architecture is set of constraints; is robust if one small component deviates [violates contraint?] then the system still performs
16:40:35 [sandro]
Tim: So it's useful to talk about architectures, even if there are exceptions and workarounds.
16:40:41 [DanC]
a robust architecture is one in which overall benefits of conformance to it is preserved in the face of deviations from it by small parts
16:40:58 [gk-scr]
Bijan: there are different architectures. DOn't care about robust architecture, do care about robust system.
16:40:59 [sandro]
Bijan: I care about robust systems, not robust architecures.
16:41:10 [sandro]
20 minutes, Tim
16:41:15 [sandro]
19
16:41:47 [sandro]
Bijan: If God came down and said "The web doesnt conform to web architecrure" that would be fine, right?
16:42:04 [gk-scr]
Bijan: even if most of web didn't conform to web architecture, but still works as intended, that's OK?
16:42:13 [sandro]
Bijan: two step arguement -- what the TAG has written might be wrong
16:42:43 [gk-scr]
... when Tim says there;s a hole in arch., is not sufficient to change the system; may be cause to change arch description.
16:43:21 [sandro]
Bijan: So Tim's arguments are not compelling to me, because I don't accept the TAG's notion of web archecture. (although I don't reject it either.)
16:43:25 [gk-scr]
TimBL says "if we don't plug this architectural gap, RDF documents are meaningless"
16:43:36 [gk-scr]
^^^ was bijan speaking
16:43:41 [tim-mit]
Timbl OBJECTS to that characterization
16:43:45 [DanC]
pls do study "3 Why MIME headers are authoritative" in http://www.w3.org/2001/tag/doc/mime-respect.html , bijan
16:44:12 [gk-scr]
PatH: could be taken to apply to web architecture generally, or specifically to semantic web?
16:44:18 [sandro]
Pat: are you not-accepting of WebArch as applies to SemWeb, or the WHOLE web?
16:44:53 [sandro]
Bijan: (unintelligble)
16:44:56 [gk-scr]
Bijan: web arch has that saying RDF docs are meaningless is a refutation of that [which?] architecture
16:45:34 [Stuart]
AFAIK the current Web Arch document says nothing about meaning of an RDF document.
16:46:21 [sandro]
I observe that this experiment is failing. Granting Bijan the right to interrupt Tim does not seem to be helping.
16:46:37 [bijan]
It's not respected,yes
16:46:41 [DanC]
this excercise (allowing timbl and bijan to discuss at will) has not been as useful as I had hoped.
16:46:52 [gk-scr]
TimBL: tradition for pieces to fill in architectural links around them.
16:47:03 [DanC]
I don't think you ever asked for or were granted that right, bijan. you claimed it, but not in a useful way.
16:47:13 [gk-scr]
... [try to fit in with broader arch. picture?]
16:47:23 [bijan]
Yes I did
16:47:24 [bijan]
Last time
16:47:29 [DanC]
nope
16:48:05 [gk-scr]
TimBL: (to Bijan) do you now want to not be part of web arch?
16:48:38 [sandro]
Bijan: I said I needed to be able to stop and question on a line-by-line basis.
16:48:43 [gk-scr]
Bijan: I saifd "I don;t understand" ... if I am to understand, I need to be able to interrupt and understand point-by-point.
16:48:59 [sandro]
Bijan: So I still can't follow you.
16:49:09 [sandro]
Tim: So lets try to work through this.
16:49:26 [gk-scr]
DanC... time to adjourn.
16:49:30 [Stuart]
q+
16:49:37 [gk-scr]
Discuss nest meeting
16:49:42 [sandro]
Pat: This was helpful for me.
16:49:58 [sandro]
Tim: This could be toned down.
16:50:12 [Stuart]
q-
16:50:36 [gk-scr]
Discuss alternative modus operandi...
16:50:49 [sandro]
zakim, who is here?
16:50:49 [Zakim]
On the phone I see Sandro, Bijan, GrahamKlyne, John_Black (muted), pfps, Stuart, TimBL, DanC, PatH, Mike_Dean, JimH
16:50:51 [Zakim]
On IRC I see mdean, tim-mit, Stuart, bijan, pfps, JohnBlack, DanC, gk-scr, RRSAgent, Zakim, sandro
16:50:53 [gk-scr]
Could use queue... but long queues not always very effective
16:51:23 [sandro]
ack JohnBlack
16:51:39 [DanC]
I'd like folks to say whether this was useful too.
16:51:41 [pfps]
I found this rather distracting, generally because the statements were too long. I think that this is the same problem that Bijan has, but I have it both with Tim and Bijan.
16:51:45 [gk-scr]
JohnBlack: stay closer to single issues.
16:51:51 [sandro]
John: single issues, closer to talking points.
16:52:08 [Stuart]
+1 wrt pfps
16:52:13 [gk-scr]
.. this was not really useful. Got lost in parenthetical ideas. Stack overflow.
16:52:19 [DanC]
fair enough, bijan
16:52:39 [pfps]
I agree that staying with single issues is going to be important.
16:52:41 [bijan]
+1!
16:52:46 [gk-scr]
Peter: too much wandering over the known universe. Like to stick to issues. (My issues!)
16:53:16 [gk-scr]
Stuart: someone sit down with Bijan and Tim, and summarize points of disagreement.
16:53:21 [sandro]
Stuart: maybe someone can mediate, go-between Bijan and Tim.
16:53:47 [sandro]
Stuart: as they interrupt each other, we don't get full thoughts.
16:53:47 [gk-scr]
.. this has been hard to follow. Don't like sence of having words put in their mouths.
16:54:41 [gk-scr]
TimBL: agree that "depth first" not working... too much divergence down side topics like what is an architecture.
16:54:56 [bijan]
q+
16:55:00 [sandro]
Tim: main fault was depth-first, ... arch/spec, what and arch is, whether there should be one, if it's wrong, ... didnt really get to techincal issue. Bad tendency to take apart first word in sentence. This was more frustration.
16:55:00 [gk-scr]
In the past, we made more progress.
16:55:21 [sandro]
Dan: Neither Tim nor Bijan is guilty of anything -- we asked them to do this.
16:55:26 [gk-scr]
DanC: don't think Tim or Bijan are "guilty" of anything.
16:55:42 [bijan]
I thank dan for that
16:55:54 [gk-scr]
.. this had to be tried. Was frustrating. Would be interested if Pat or Peter woulc chair a meeting.
16:55:54 [bijan]
q-
16:56:22 [sandro]
Peter: if you give me a gavel...
16:56:32 [gk-scr]
Peter: the deal... equip me with a gavel.
16:56:41 [sandro]
Peter: but I agree with Bijan abotu when to apply gavel.
16:56:48 [sandro]
Pat: +1 DanC
16:56:54 [sandro]
Pat: frustrating but productive
16:57:14 [sandro]
Pat: have to float high to reveal disconnects
16:57:16 [gk-scr]
Pat: what Dan said. Productive, if frustrating. Need to find background assumptions causing disconnects.
16:57:39 [sandro]
Pat: Yeah, I could chair.
16:57:40 [bijan]
Earlyand aoften :)
16:57:50 [gk-scr]
I could chair, but time is pressing for next few weeks. Next meeting 31-Oct is fine.
16:57:58 [pfps]
Early, often, and loud!
16:58:02 [gk-scr]
MikeD?
16:58:11 [Stuart]
:-)
16:58:12 [bijan]
I note that the early folks didn't get a chance to comment on the process
16:58:12 [gk-scr]
MikeD: nothing to add.
16:58:27 [sandro]
Jim: Useful in terms of establishing vocabulary
16:58:44 [gk-scr]
Jim: useful in establishing the vocabulary. Too much debating for debating.
16:58:53 [sandro]
Jim: Today got too much like debating for debating. Too much deconstruction.
16:58:57 [gk-scr]
... better to grant some hypotheses
16:59:41 [sandro]
Jim: I like Pat/Peter try chairing -- picking agenda, ... Issues like "If you're right, it causes this problem."
16:59:50 [gk-scr]
.. like idea trying Pat/Peter as chair. Would like when possible to identify cases of "if your right then this problem".
17:00:17 [Zakim]
-JimH
17:00:20 [sandro]
ADJOURN
17:00:24 [Zakim]
-Stuart
17:00:25 [Zakim]
-Bijan
17:00:26 [Zakim]
-TimBL
17:00:27 [Zakim]
-Mike_Dean
17:00:27 [Zakim]
-PatH
17:00:28 [Zakim]
-John_Black
17:00:30 [gk-scr]
phew!
17:00:32 [Zakim]
-DanC
17:00:33 [Zakim]
-pfps
17:00:38 [Zakim]
-Sandro
17:00:43 [Zakim]
-GrahamKlyne
17:00:43 [sandro]
Great job, GK!
17:00:44 [Zakim]
SW_Meaning()11:30AM has ended
17:01:16 [gk-scr]
(Times like this, I wish I learned to touch-type)
17:02:42 [tim-mit]
thanks, graham!
19:42:48 [tim-mit]
tim-mit has joined #sw-meaning
20:43:18 [tim-mit]
tim-mit has left #sw-meaning
21:24:25 [Zakim]
Zakim has left #sw-meaning
21:25:59 [sandro]
bye, zakim
21:26:23 [DanC]
DanC has left #sw-meaning