W3C home > Mailing lists > Public > www-archive@w3.org > April 2003

#rdfig log (partial?) for 2003-04-20 (logger bot missing)

From: Dan Brickley <danbri@w3.org>
Date: Sun, 20 Apr 2003 14:03:27 -0400
To: www-archive@w3.org
Message-ID: <20030420180327.GF3876@tux.w3.org>

----- Forwarded message from Dan Brickley <danbri@fireball.danbri.org> -----

From: Dan Brickley <danbri@fireball.danbri.org>
Date: Sun, 20 Apr 2003 18:59:12 +0100
To: danbri@w3.org
Message-Id: <E197J5s-0001ZF-00@fireball>

<DanC> what's with all the "These notes do not have any official status" malarky?
<DanC> where are the "official minutes"?
--- q|cook is now known as qmacro
<larsbot> well. this is ISO for you
<DanC> how so?
<larsbot> ISO doesn't do minutes, only recommendations
<larsbot> sorry, resoltions
<DanC> doesn't do minutes? then what the heck does "Decisions of SC34/WG3 are made in accordance with ISO policies and procedures and as such, are recorded solely in the offical minutes of that body." mean?
<larsbot> :-)
<larsbot> subcommittees, being the bodies that make decisions, do resolutions
<larsbot> they also publish the recommendations that each working group make to the SC
<larsbot> in our case that wasn't enough, so we also published unofficial minutes
<larsbot> here's the resolutions from Baltimore, which may help: http://www.y12.doe.gov/sgml/sc34/document/0359res.htm
<DanC> ok, so "Proposed Resolution see solution above." is the record of the agreement among those in Baltimore?
<DanC> s/the record/a record/
<larsbot> yes
<larsbot> ISO itself doesn't make decisions at this level of granularity
<larsbot> only when there's a standard text does ISO itself decide anything
<DanC> sure.
<larsbot> in our case that was months away...
<DanC> so... which of the 5 resolutions in 0359 corresponds to "ditch the topic-naming-constraint"?
<larsbot> the third
<DanC> do you expect the official minutes to cite the unofficial notes, or vice versa?
<larsbot> there is no such thing as "official minutes"
<DanC> sorry, official record of resolutions.
<DanC> oops! typo! "TITLE:  Resolutions of the SC34 Meeting, Baltimore, 7\x{2013}21 December 2002"
<DanC> it's 7-12 Dec, not 7-21 Dec, right?
<larsbot> indeed :)
* DanC thinks he's got a mental model of the decision process in his head now... considers stopping to write it down...
* DanC goes back to technical stuff, the SAM, instead
<DanC> "Date: 2003-04-03"
<DanC> hot off the press!
<larsbot> yeah. one more try and we'll be ready to enter the ISO process towards IS
<DanC> [[ Equality rule: Topic name items are equal if the values of their [value], [type], and [scope] properties are equal and they are contained in the [topic names] property of the same topic item. ]]
<DanC> contained?
<larsbot> why do you ask?
<DanC> what does it mean for ... umm... a value (?) to be contained in a property?
<larsbot> it's not contained in the property, but in the set that is the value of the property
<larsbot> repeating this over and over quickly gets painful, however
* DanC is confused
* DanC has some domestic duties to tend to, meanwhile
<DanC> or rather... domestic opportunities. this is a weekend, after all.
<DanC> ooh... uris for terms... http://www.topicmaps.org/xtm/1.0/core.xtm#class-instance
<DanC> is that URI going to be standardized by ISO?
<DanC> the oid gods didn't notice you or something?
<DanC> I guess the more relevant one is http://psi.topicmaps.org/sam/1.0/#type-instance
<larsbot> bingo
<larsbot> the previous one is obsolete
<larsbot> re gods: we'll see...
<DanC> good luck, then!
* DanC tries http://psi.topicmaps.org/sam/1.0/#type-instance ... bummer... no DNS record for psi.topicmaps.org
<larsbot> not yet. we're still negotiating the transfer of that domain
<DanC> ok, we've reached the bounds of my process curiosity. I don't want to know about domain transfer negotiations.
<DanC> ah... [[ Issue (psi-publishing):
<DanC> Where should the indicators for the subjects published in the new ISO 13250 be published?
<DanC>  ]]
<DanC> any topicmap test suites in development?
<larsbot> not yet. we're still working on the canonical TM syntax
<larsbot> well, that's half-true. Ontopia has one
* DanC wanders off, hoping to catch larsbot again soon re esw:TopicMaps
* larsbot goes off in the direction of beer
* danbri plays jabber/irc gatewaying w/ qmacro
--- qmacro is now known as q|zzz
<mortenf> hmmm, jabber is that boring?
--> mariyo (mnishikawa@h219-110-113-153.catv01.itscom.jp) has joined #rdfig
<mortenf> does HTTP PUT make sense in a tripled world?
--> MarkB (~mbaker@ has joined #rdfig
--> Radical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- Radical_Edward has quit ("Always knew it would come to this Things have never been so swell And I have never felt so well")
--> Radical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- orphik has quit (Remote closed the connection)
<-- pixel has quit (Read error: 104 (Connection reset by peer))
<-- MarkB has quit (Read error: 110 (Connection timed out))
--> DanC_jam (~connolly@151-30-252.cmts03.ksle.everestkc.net) has joined #rdfig
--> niq (~nick@walkham.demon.co.uk) has joined #rdfig
<-- mortenf has quit ("Client Exiting")
<-- niq has quit ("[x]chat")
--> R4dical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- Radical_Edward has quit (Read error: 110 (Connection timed out))
--> Radical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- R4dical_Edward has quit (Read error: 110 (Connection timed out))
<-- Radical_Edward has quit (Read error: 104 (Connection reset by peer))
--> Radical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- mariyo has quit (Remote closed the connection)
<-- Radical_Edward has quit (Read error: 104 (Connection reset by peer))
--> Radical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- Radical_Edward has quit (Read error: 110 (Connection timed out))
<-- Oblomov has quit (Read error: 104 (Connection reset by peer))
--> orphik (orphic@chicks.dig.h4x0rs.net) has joined #rdfig
--> Radical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- DanC_jam has quit (Read error: 110 (Connection timed out))
--> Radical_Edw4rd (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- orphik (orphic@chicks.dig.h4x0rs.net) has left #rdfig
--> call (call@nrjo15-1162.dial.rjo.embratel.net.br) has joined #rdfig
<-- Radical_Edward has quit (Read error: 110 (Connection timed out))
<-- call (call@nrjo15-1162.dial.rjo.embratel.net.br) has left #rdfig
--> dnm (~dnm@cleanwhisker.420.am) has joined #rdfig
--> Radical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- Radical_Edw4rd has quit (Read error: 104 (Connection reset by peer))
--> Esaj (~jason@jdd30.joh.cam.ac.uk) has joined #rdfig
<-- Esaj has quit ("Client exiting")
--- q|zzz is now known as qmacro
--> mortenf (~mortenf@ has joined #rdfig
<qmacro> morning mortenf ; liked your comment about zzz last night :)
<mortenf> hi there.
<mortenf> oh yeah, that one. ;)
<qmacro> one big plus point for IRC: there are lots of text-mode clients. I'm currently sitting outside at a DEC VT330 ...
<mortenf> heh.
--> JibberJim (~none@ has joined #rdfig
--> jordan (jordan@ has joined #rdfig
--> MarkB (~mbaker@ has joined #rdfig
--> Esaj (~jason@jdd30.joh.cam.ac.uk) has joined #rdfig
<-- JibberJim has quit ("Bye")
<-- MarkB has quit (Read error: 110 (Connection timed out))
* mortenf is away: family time...
--> libby (~libby@pc-62-30-66-25-az.blueyonder.co.uk) has joined #rdfig
<-- libby has quit ("Client exiting")
* danbri relearns about ruby mixins...
<danbri> [
<danbri> class SimpleSQLGraph
<danbri>   include RDF4R::Pastiche::Cwm
<danbri> end
<danbri> module RDF4R
<danbri>   module Pastiche
<danbri>     class Cwm
<danbri>       def initialize
<danbri>         puts "Cwm pastiche initializing"
<danbri>       end
<danbri>       def Cwm.load(uri)
<danbri>         return Cwm.new(Loader.get_rdf_from_uri(uri))
<danbri>       end
<danbri>     end
<danbri>   end
<danbri> end
<danbri> ]
<danbri> not quite right... basically i'm trying to wrap my stuff in a Cwm-alike
<danbri> for the basic RDF api stuff, no fancy thinkin' and deducin' in my code ;)
<danbri> I might offer RDF4R::Pastiche::* "tributes" to some other APIs too... once/if i get this working.
<danbri> I have just written glue so I can use a statements-matching api against SQL-backed (and scutter-populated) persistent store. Before it was Squish-only, which is a pain as squish doesn't allow for defaults, so you have to make a half dozen separate query strings if your data is gappy and unpredictable
<danbri> (which scutter'd data always is)
<danbri> BLURB:for RDF Query requirements
-dc_rdfig/#rdfig- A: for RDF Query requirements from danbri
<danbri> A:My RDF harvesting-based apps, where data is partial, fragmented and merged from multiple sketchy sources, suggest that an RDF query language SHOULD/MAY/??? allow for specification of optional matches in graph templates...
-dc_rdfig/#rdfig- Added comment A1.
<danbri> A:In practice, this manifests itself in my building apps through repeated, speculative API calls, instead of via more rigid graph-match specifications via Squish/RQDL
-dc_rdfig/#rdfig- Added comment A2.
<danbri> A:Comments anyone? Am I alone in finding this?
-dc_rdfig/#rdfig- Added comment A3.
<-- jordan has quit (Read error: 110 (Connection timed out))
--> R4dical_Edward (mea_culpa@12-212-98-134.client.attbi.com) has joined #rdfig
<-- Radical_Edward has quit (Read error: 104 (Connection reset by peer))
<-- xover has quit (Excess Flood)
--> xover (~link@mail.tj.unn.no) has joined #rdfig
<danbri> http://membled.com/work/apps/xmltv/presentation/07history.html
-dc_rdfig/#rdfig- B: http://membled.com/work/apps/xmltv/presentation/07history.html from danbri
<danbri> B:|XMLTV project history
-dc_rdfig/#rdfig- Titled item B.
<danbri> B:with [http://www.ananova.com/tv_listings/_tv_raw.php|xml feed from Ananova] of tv listings. Candidate for scraping into ESW:RdfCalendar ?
-dc_rdfig/#rdfig- Added comment B1.
--> AndyS (~AndyS@82-32-58-37.cable.ubr08.azte.blueyonder.co.uk) has joined #rdfig
<AndyS> DanBri?  "Yes" to your chump on partial data.
* danbri nods, remembers chatting about this with you at some point
<danbri> mozilla, dql have such constructs, and i think even SQL-backed ql rewriting systems can support it...
<danbri> (but... 'scope creep' ;)
<danbri> I want to get more experience with using v dump QL + go back into the DB via API for rest of graph, find out quite how painful that is
<danbri> lots of what i need to do use provenance etc info too, not sure a lightweight api could support that anyhow
<danbri> s/api/ql/
<-- dnm has quit ()
<AndyS> I think there are two functions here - locating data and extracting data
<AndyS> Locating is OK with the rigid query but extracting on partial data is not a good fit
<danbri> yup
<AndyS> I'm assuming you are not querying/constraining on the partial stuff (but you could be)
--> jordan (jordan@ has joined #rdfig
<danbri> not in these cases, though thats a good q how often that's needed
<-- jordan has quit (Client Quit)
<AndyS> I'm not sure what it means to test on partial data - it is like missing keys in a DB
* danbri ponders, one might want the dc:rights to be either 'free' or unspecfied, but not 'not for circulation'
<AndyS> I do think it is a requirement for an RDF QL
<danbri> (off top of my head... not clear about this at all)
<AndyS> Hmm - that's default values - yet another case/requirment?
<danbri> it could be done w/ defaulting, but that'd spook the OWL folks probably...
<AndyS> RDF is open world so the meaning of "not present" isn't default but I guess applications often add this rule
<danbri> Talking of requirements, we could find ourselves in a tug-of-love between pull of XML Query and OWL... Was wondering how familiar you were with the XML Query work
<AndyS> If so, core query shouldn't do it, application i/f might
<AndyS> Moderately familiar.
* danbri nods re openworld; apps need a good excuse for adding defaulting assumptions (eg. extra knowledge of the dataset...)
<AndyS> (Also throw premises into the mix)
<danbri> urk :)
<AndyS> Was there something specific on XML Query?
<danbri> was asking in general, though the 1st point of interaction probably the datatype-oriented operators they listed, and scenarios for doing xpathy matches against literal XML data within rdf graphs
<AndyS> Operators look like a good starting point - it is in value-space, not pure RDF, but we can't do without it.
* danbri agrees
<AndyS> The xpath stuff (c.f. versa) pulls two ways for me; 
<danbri> i'd like to see some datatype-happy rdfq implementations, haven't really looked at what folk have built on that front yet
<danbri> some folks seem to want to do the whole thing in a path-y way...
<AndyS> good point is that some group has put some ground work in but it is very XML-syntax ...
<AndyS> .. attributes, siblings etc and I have not convinced myself it is better than graph match.
<AndyS> Also, the mapping to JOINs is getting beyond my familiarity with the details
<danbri> It would be good to know how much actual XML content (eg. XHTML/weblog in RSS) gets shoveled around inside RDF...
<AndyS> There are different aspects - the more RDF/XML way (paths) and "RDF as a database" (more declarative)
<AndyS> Paths are OK if the query is still a declarative specification and not a recipe for execution without heavy duty implementations
<danbri> the other thing re xml query is it seems to be a full programming/scripting solution wrapped around a data-oriented view of XML. Should we anticipate their being people who think of themselves as 'xml query programmers'... if so, how might they view RDF plugging into their world?
<danbri> agreed re impl impositions
<AndyS> An "RDF programming language" (heavier than just scripting I hope) is interesting, if large
<danbri> yes, though for now i'm happy with calling RDF stuff from perl/ruby/python, the mismatch isn't too painful at moment
<AndyS> I wonder if RDF changes one or two base assumptions requiring reworking the entire execution paradigm (AKA model!)
<AndyS> (The "scripting" remark was that ease of use is significant over efficiency.  I like scripting (in bash, perl, emacs))
<danbri> yes, definitely re use of use
<AndyS> How can we encourage some part of the XML query community to the conversation, along with rules people
<AndyS> Do you think the OWL folks see query being on "OWL data" or on "RDF data"?
<AndyS> Back on partial data - do you think a separate clause is needed in Squish? 
<danbri> I think OWL folks would hope OWL semantics might be honoured, in whole or part, by query answering systems. Re encouraging others along, I think sending them a 'heads up' ping soon would be good, and maybe seeking out a few RDF sympathisers who could take a more detailed involvement
<AndyS> So WHERE is conjunctive exact match, and a second EXTRACT describes stuff to pull out?
<danbri> that could do it, i've wondered about syntax... not sure what's best
<danbri> perhaps get the functionality API-visible first then worry about its textual representation?
<AndyS> Did you see SeRQL with its CONSTRUCT form?
<danbri> I saw the announce, didn't look at the spec yet. Is that what CONSTRUCT does?
* danbri heard it could compose an RDF graph (for recursive query of some kind?)
<AndyS> I think APIs are prog language specific: my appraoch is its abstract model then lang syntax 
<AndyS> I think CONSTRUCT is a RDF template like a cwm RHS.
<AndyS> But saying that, what API ops would you include?
<AndyS> Extra variables in the binding set?  Additional calls into the graph anchored on the current result?
<danbri> yes, prog lang specific... but if there were a couple of different rdfq impls that had their own apis for optional 'extract's, that'd help motivate getting it into a common textual language, for testbedding across implementations
* danbri not sure how how to do it at all yet, sorry!
* AndyS asks that danbri adds to Alberto's use cases if it is not there yet
<danbri> you mean write up a scenario where we pull out data including optional properties?
<AndyS> That - or an ideal query solution and write a desciptive para (less work?)
* danbri looks for the XUL example that covers this
<danbri> dc_rdfig:view
-dc_rdfig/#rdfig- A: for RDF Query requirements (blurb)
-dc_rdfig/#rdfig- B: XMLTV project history (http://membled.com/work/apps/xmltv/presentation/07history.html)
<danbri> logger_1, pointer?
<danbri> eek no logger :(
<AndyS> No logger and no rdfig/latest :-(
* danbri stashes a TODO at botttom of http://esw.w3.org/topic/RDFQueryTestcasesRequirements w/ XUL link
<danbri> ie http://www.mozilla.org/docs/xul/xulnotes/template-bindings.html
<danbri> no logger/etc... it's the Internet's way of telling us to get offline at the weekend ;)
* danbri heads off, have a good evening!
* AndyS notes there is no record of this conversation .... :-)
<-- AndyS has quit ("Bye")

----- End forwarded message -----
Received on Sunday, 20 April 2003 14:03:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 7 November 2012 14:17:29 GMT