Semantic Web Interest Group IRC Chat Logs for 2004-03-29

00:00:10 <Arnia> Moggi has McCarthy pinned to the ground and is forcing a monad down his throat

00:00:22 * chaalsMEL reminds Timbl that he didn't even pay for the argument yet!

00:01:03 <timbl> yes i did

00:01:05 <chaalsMEL> Arnia, the point was that I was wondering if you could get adenosine to run as n3 anyway...

00:02:06 <Arnia> chaalsMEL: You can compile adenosine to n3 (in much the same way as Fabl is compiled, but with a richer semantics)

00:02:37 <Arnia> chaalsMEL: And you can use any current n3 file as if it were a behaviour-less adenosine file

00:03:38 <Arnia> (at least in theory, I'm only just about to start building a compiler for this language :)

00:04:47 <chaalsMEL> timbl, you haven't produced any evidence. Therefore you are just making declarative assertions

00:05:29 <IsoosI> I much prefer something with nice, trivial semantics, such as n-triples. if you want something pretty, use an editor.

00:05:34 <chaalsMEL> ... how can you expect anyone to construct an argument on that basis? Are there any Rules?

00:05:56 <Arnia> n-triples has the same semantics as RDF/XML, N3 or anything else, surely :)

00:06:15 <Arnia> chaalsMEL: But are the rules forward or backward chaining

00:06:19 <chaalsMEL> Issosl: Right. I use an editor. I just find that when people send me data in 5 formats it is difficult to find an editor that can keep up with the proliferation.

00:06:54 <sh1mmer> chaalsMEL I am trying to get damian to add n3 to RDF Author since jena 2 supports it

00:07:08 * chaalsMEL wonders about the implications of forward chaining timbl into the arguments room, or backward chaining him here...

00:07:15 <deltab> IsoosI: those have different syntax, not semantics, afaik

00:07:26 <sh1mmer> libby said that he is going to start a stint at hp tho, so who knows if he will have time

00:07:32 <IsoosI> deltab: yep.

00:08:03 <IsoosI> I'm slowly writing a backwards chaining n3 interpreter

00:08:23 <Arnia> chaalsMEL: I could apply a conditional predicate on you... but that may be a little too kinky

00:09:02 * chaalsMEL goes for a nice cup of tea and some breakfast

00:09:15 <Arnia> Just as I thought, one of the Establishment ;)

00:09:26 <IsoosI> since forward chaining seems (to me at least) to have scalability issues if you want to scale up to the size of the Internet.

00:09:53 <Arnia> IsoosI: They each have their place... a point I made in my Introduction to Adenosine :)

00:10:05 <danbri_dna> Sorry timbl, the five minutes is up.

00:10:17 <IsoosI> Arnia: True.

00:10:29 * sh1mmer waves to danbri_dna

00:10:36 <danbri_dna> 'ello

00:10:48 * crschmidt wonders what the _dna is for

00:10:54 <danbri_dna> laptop's name.

00:11:03 <danbri_dna> .google douglas n adams

00:11:05 <datum> douglas n adams: http://www.douglasadams.com/

00:11:07 <Arnia> Douglas Nigel Adams?

00:11:12 <Arnia> Ah yes :)

00:11:26 <crschmidt> ah.

00:11:38 <Arnia> DNA is also the name that h2g2 gave to the software that runs the BBC communities after he died

00:11:47 <sh1mmer> danbri I am agahst libby kept me out drinking ;)

00:11:54 <Arnia>http://www.bbc.co.uk/dna/hub

00:11:54 <dc_rdfig> A: http://www.bbc.co.uk/dna/hub from Arnia

00:12:01 <Arnia> D'oh

00:12:20 <sh1mmer> you can remove it. i forget how

00:12:24 <crschmidt> forget A

00:12:30 <crschmidt> hm. not like that ;)

00:12:39 <danbri_dna> A:=""

00:12:39 <dc_rdfig> Replacement must be a valid URL.

00:12:44 <Arnia> A:|DNA, the massive community information software developed for the BBC by h2g2

00:12:44 <dc_rdfig> Titled item A.

00:12:51 <Arnia> May as well title it for now

00:12:52 <danbri_dna> that's better :)

00:13:10 <sh1mmer> Arnia its realted

00:13:12 <sh1mmer> kinda ;)

00:13:21 <Arnia> In a vague sense.

00:13:35 <Arnia> A wonderful example of why the semantic web needs a concept of trust ;)

00:13:48 <sh1mmer> or a concept of not to trust you :)

00:14:08 <IsoosI> heh

00:14:28 <Arnia> Hey... I can have my own property - foaf:dontTrustArnia - that only works for me ;)

00:14:42 * sh1mmer wonders if chaalsMEL has actually developed a semantic spaming tool yet

00:14:58 <Arnia> Ooh ooh ooh! I want to spam Bale!

00:15:12 <Arnia> Damn... have to write a reference impl first *sulks*

00:15:14 <sh1mmer> Arnia you might come under foaf:evilIncarnate

00:15:28 <Arnia> I'm not evil, I'm misunderstood

00:18:26 <Arnia> Ok... sh1mmer, what's your opinion on sugaring. Good or bad?

00:18:40 <sh1mmer> if you want sugar use a tool

00:18:53 <Arnia> :p

00:18:54 * danbri_dna is away: zzzzzz

00:18:59 <sh1mmer> if you want to code by hand learn to learn

00:19:09 <Arnia> I am... its called the adenosine compiler

00:20:14 <sh1mmer> there needs to be significant advantage to a language format to justify another language

00:20:48 <Arnia> Ability to write code without representing abstract syntax trees in triples format?

00:21:01 <Arnia> A concise way to express functional closures?

00:21:51 <sh1mmer> im going to bed

00:21:57 <sh1mmer> some of us have work tmw

00:22:00 <sh1mmer> ;)

00:22:02 <sh1mmer> gnight all

00:22:04 <Arnia> :p

00:22:06 <Arnia> Night

00:22:25 <chaalsMEL> nn sh1mmer

00:22:40 * Arnia resists the reference

00:24:02 <chaalsMEL> Arnia, noodling aloud...

00:24:19 <chaalsMEL> Wouldn't it be easy to describe a particular web service as a method or function?

00:24:30 <Arnia> chaalsMEL: Yes

00:25:03 <Arnia> chaalsMEL: And you conditionally introduce code into a particular context with a combination of forward and back chaining inference

00:25:19 <Arnia> chaalsMEL: I think I make a reference to that in the Jabberwocky section...

00:26:02 <chaalsMEL> Since it is perhaps available at a URI - and therefore can be thought of as an RDF resource (modulo the question of whether a resource in http space also means what you get when you try it, or whether that's up to somebody to decide).

00:26:28 <Arnia> chaalsMEL: Yep :) This is kinda what I was hoping people would see in the idea

00:26:53 <Arnia> chaalsMEL: This coupled with Jabberwocky is immensely exciting for me... I can see all kinds of uses

00:27:07 <chaalsMEL> So you could just write web services... Then there is the question of how you process them locally (i.e. when you don't have access to them). I wonder how useful WSDL is for figuring out what Web Services really really actually do.

00:27:43 <Arnia> Ok, how about this style syntax - @function ns:name(arg1,..) { CODE } ?

00:27:54 <Arnia> That way all the weird stuff is kept in an @ rule

00:28:20 <Arnia> chaalsMEL: I guess this would test that, wouldn't it :)

00:28:28 <chaalsMEL> Yep.

00:29:03 <chaalsMEL> The approach I ahve been playing with is about taking simple bits of Xforms or CSS or XSLT code and trying to use them to represent javascript

00:29:29 <Arnia> The really fun thing for me is that if you set up Bale to distribute storage on the basis of context URIs, you can transparently have 'computational sandboxes' over the semantic web

00:30:11 <Arnia> I've also begun playing with methods of representing I/O (especially GUIs) using contexts and ontologies

00:30:17 <chaalsMEL> And then identify those pieces so that you can refer to things that do the same thing even though they are different. But I haven't got very far down the path of figuring out how you might actually decsribe them.

00:31:02 <Arnia> chaalsMEL: Sounds like the bastard son of dependent types :)

00:31:25 <chaalsMEL> There's a research group in France called Rainbow who have been looking into that. And there is a long-running accessibility project at Stanford I thin (It's heaedd by a Kiwi and it has a classical greek name) that essentially does the same thing.

00:32:17 <chaalsMEL> Their approach is that instead of caring about how you create content, you can build AI systems poewrful enough to present it for people with disabilities.

00:32:50 <Arnia> chaalsMEL: I'm researching natural language generation from RDF as well

00:33:41 <chaalsMEL> have you seen the WWAAC stuff on using RDF to work across symbol vocabularies (which have the nice property that as natural languages they are small, concrete, and generally have simple grammar)?

00:34:05 * chaalsMEL did a lighting talk at the technical plenary on that kind of work.

00:34:16 <Arnia> chaalsMEL: No... I've been working with functional unification grammars

00:35:31 <Arnia> chaalsMEL: FUF/SURGE produces some impressive results for example... taking a semantic representation and using a lexicon to generate a syntactic statement of the semantics

00:36:12 <chaalsMEL> Aah. The [http://archimedes.stanford.edu/|Archimedes project] is the one I was thinking of.

00:36:31 <Arnia> But then again, in linguistic phrasal syntax I'm a unification rather than a transformation grammar man

00:36:36 <Arnia> Cool :)

00:37:17 <chaalsMEL> Arnia, there's lots of interesting stuff in that area trying to find it's way into Voice-based interaction (like VoiceXML - the group published a handful of stuff pointing work in that direction). Again it benefits from working most often in reasonably constrained circumstances.

00:38:00 <chaalsMEL> unification vs transformation: I'm feeling uncomfortable, with a few pickets digging into my backside.

00:39:15 <Arnia> chaalsMEL: We'll see :) I've got a good lead on a method to represent lexical information (such as that found in HPSG etc) in a useful way in RDF

00:39:28 <chaalsMEL> My practical approach to translating when I do it is a bit of both, and a big dash of brute-force dictionary lookup of phrases.

00:39:28 <Arnia> chaalsMEL: I'm going to run with it and see if I can make all the ends join up

00:39:40 <chaalsMEL> HPSG?

00:40:30 <Arnia> Head-driven Phrase Structure Grammar - a unification model of syntax which eliminates need for movement and allows bottom up generation of phrases from lexical components

00:40:36 <Arnia> Pollard and Sag 1994

00:40:39 * chaalsMEL suffers from having learnedd translation as an historian...

00:41:25 <Arnia> My area of interest includes cognitive linguistics :)

00:43:14 * chaalsMEL thinks about the shades of meaning conveyed by the phrase 'area of interest'

00:44:00 <Arnia> Its a sign that all my interests are subtly linked by the fact that I keep running into the same papers and authors again and again in all the fields

00:44:25 <Arnia> But at the moment, I'm interested in Jabberwocky and Adenosine :)

00:54:33 <Arnia> chaalsMEL: Do you reckon its sensible to represent program transformations as inference rules in RDF?

00:57:16 <chaalsMEL> Hmmm. Off the top of my head I am not sure.

00:58:20 <Arnia> chaalsMEL: If we adopt a method like that in Adenosine then it shouldn't be impossible... its just whether its worthwhile :)

00:58:26 <chaalsMEL> I don't have an instant reason why not thought.

00:58:47 <Arnia> chaalsMEL: Of course, that would allow people to distribute 'optimisations' through the semantic web

00:58:56 <Arnia> Quite a cool concept

00:59:13 <chaalsMEL> To some extent it is what having Web Services does.

00:59:24 <chaalsMEL> I think

00:59:45 <chaalsMEL> being able to optimise that for offline use, or to replace one service with a more usful one strikes me as valuable.

01:00:14 <Arnia> chaalsMEL: In adenosine, everything is first class. The power of RDF :)

01:00:21 <chaalsMEL> It's also kind of important to know that if you want to try and do choreography of web services (real-time compositing)

01:00:45 <chaalsMEL> Hmm. I don't quite understand how the mixed syntax works in practice.

01:00:53 <chaalsMEL> ...in adenosine

01:01:08 <Arnia> heh, it just occured to me that adenosine fixes 'problems' in both n3 and ECMAScript. To n3 it gives behaviour, to ECMAScript it gives a transparent datastore and modules

01:01:59 <Arnia> Well, when I talk about adenosine I can really mean two things. I either mean the surface syntax (however that pans out) or I mean the underlying RDF ontology that code gets compiled into

01:02:04 <chaalsMEL> well, cwm built-ins like math give n3 behaviour. The question is whether it is worth adding in functionality that breaks the model where everything is declarative.

01:02:33 <chaalsMEL> (talking about the syntax). Once it is compiled into nice underlying RDF it's fine.

01:02:46 <chaalsMEL> How do you handle variables in compiling back to RDF?

01:03:36 <Arnia> chaalsMEL: Variables are going to be bnodes in the context they are declared in.

01:03:51 <chaalsMEL> I don't think cwm really has a good handle on that at the moment - it makes an anonymous thing in the document, but it isn't prepared to use it in RDF/XML. I am not sure if that shows a difference in semantics between n3 and RDF/XML or it is just an artifact of the code.

01:04:01 <Arnia> I'm using contexts in a very similar way to closures from ECMAScript and LISP

01:04:32 * chaalsMEL has to go get some stuff done while it is business hours...

01:04:48 <chaalsMEL> THings I am thinking about: How to make a for loop...

01:05:11 <Arnia> chaalsMEL: For loops are just sugar for while loops

01:05:27 <Arnia> A for loop would act as a closure as well

01:05:31 <chaalsMEL> some nice way of declaring parameters

01:05:40 <chaalsMEL> how to make a while loop :-)

01:06:21 <Arnia> Fabl already has the beginnings of an ontology for abstract syntax trees like this

01:06:27 * chaalsMEL wonders if he has time for a nap. Maybe after lunch :-(

01:06:59 <Arnia> The major change I make is to eliminate global and local variables and use closures to control such information

01:11:26 <Arnia> chaalsMEL: Thanks for this discussion :)

01:29:53 <Arnia> Is 'distobj' here? :)

01:40:22 * mattb scutters the RDF starting from the blogroll of http://www.planetrdf.com and finds some sinners

02:51:16 <Arnia22> Arnia22 is now known as Arnia

02:51:51 <Arnia> Ok, just thinking outloud here but I'd appreciate corrections if I start saying silly things

02:52:35 <Arnia> A comment on the brief outline of Jabberwocky prompted me to think deeper about the choice of protocol for Bale. Ideally we want to reuse as much technology as possible.

02:54:10 <Arnia> If we want to use adenosine to code webservices, it would make sense to have some synergy between the protocol used for Bale and the adenosine function evaluation.

02:55:45 <Arnia> If we view adenosine's function execution mechanism as cloning the closure to form a new thunk, then the most natural form of parameter passing would be as the triples, directly

02:56:26 <Arnia> I therefore propose that Bale use SOAP over HTTP with the message format described in RDF/XML.

02:57:12 <Arnia> This does open up the rather cool prospect of using adenosine to write the graph modification, distribution and query features of Bale

02:57:49 <Arnia> Of course its 4am so that may have made no sense whatsoever :)

03:47:27 <Arnia>http://www.netalleynetworks.com/community/jgeldart/research/adenosine/semantics-syntax-functions

03:47:27 <dc_rdfig> B: http://www.netalleynetworks.com/community/jgeldart/research/adenosine/semantics-syntax-functions from Arnia

03:47:42 <Arnia> B:|Some notes on the compilation of functions in Adenosine

03:47:42 <dc_rdfig> Titled item B.

04:26:06 <MarkB> SOAP/HTTP sounds good to me so long as that includes respecting HTTP's application semantics, which Web services typically don't do

05:55:47 <matthewLee>http://rdfig.xmlhack.com/

05:55:48 <dc_rdfig> C: http://rdfig.xmlhack.com/ from matthewLee

06:46:06 <zImage> jsled- I've been thinking a little about what you said on recipeML about using RDF. I'd like to hear your ideas if you're around.

08:14:56 <arnarl> hi

08:31:10 <sh1m> moin

08:32:20 <dajobe> hi

08:32:22 <dajobe> whoa, I got a new google look

08:33:07 <dajobe> our comprehensive survey of 2 people says it's changed for everyone

08:36:30 <sh1m> it is a bit different maybe

08:36:47 <sh1m> i normally just use g search terms in fluffy opera

08:45:35 <sbp> yes

08:45:49 <sbp> the coolest part of the change was Froogle going live

08:46:08 <sbp> initially it displayed random searches on the homepage that people were making...

08:46:22 <sbp> but they've since replaced that with a boring set of categories that no one will use instead

08:46:44 <sh1m> live as in not beta?

08:47:02 <sbp> perhaps they were fed up with people refreshing the homepage to check out the often hilarious results. (fleece ferret hammock?)

08:47:12 <sbp> still beta, but advertised now

08:47:16 <sh1m> ah

08:47:35 <sh1m> well in google head quaters they have something like that above receptino

08:47:50 <sbp> yeah

08:48:02 <sh1m> i can just see myself waiting to be seen and using the google api to spam until one of my hilarious results came up

08:48:04 <sh1m> ;)

08:48:30 <sbp> heh, eh

08:59:01 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

09:23:18 <sh1m> hmmm maybe i should pull the logs off sh1mmer and see what happened with timbl, chaalsMEL and arnia last night

09:23:26 <sh1m> seemed like an interesting convo

09:46:21 <Arnia> Hello everybody :)

09:49:17 <ericP> hi Arnia

09:49:34 <ericP> you signed off with something about SOAP and RDF

09:49:49 <ericP> made me thing of the W3C search tool

09:49:51 <ericP> GET -H Accept:\ application/soap+xml 'http://www.w3.org/Search/Mail/Team/search?keywords=welcome&hdr-1-name=subject&hdr-1-query=&index-grp=Team__FULL+Member__FULL+Public__FULL&index-type=t&type-index=www-rdf-rules&resultsperpage=1'

09:50:02 <ericP> (if you have LWP)

09:51:54 <Arnia> Don't have access to that. If you give me a moment to gather my thoughts (I've only just got up :)

09:52:39 <ericP> odd about not having access

09:53:01 <ericP> i'm using a machine with no special credentials

09:53:21 <sh1m> Arnia: why are you up?

09:53:32 * sh1m waves in ericP's direction

09:53:42 <Arnia> Gives me an http auth box for 'W3C Team'

09:54:22 <Arnia> sh1m: I couldn't sleep properly so I may as well try and get back onto a normal bodyclock

09:54:25 <ericP> ya know, i was wrong, i guess i do have credentials

09:54:30 <ericP> try:

09:54:30 <ericP> GET -H Accept:\ application/soap+xml 'http://www.w3.org/Search/Mail/Public/search?keywords=welcome&hdr-1-name=subject&type-index=www-rdf-rules&resultsperpage=1'

09:54:35 <ericP> heya shim

09:55:05 * ericP tries to figure out how to make his machine untrusted

09:55:12 <Arnia> :)

09:55:20 <sh1m> ericP: isn't it some weird time or are you not in .jp?

09:55:40 <ericP> 18:55 -- not to wierd, as times og

09:55:40 * Arnia teaches ericP's computer how to lie

09:56:11 <ericP> 3am tends to be a bit wierder, but that may be the company i keep

09:56:14 <sh1m> fair enough

09:56:37 * sh1m keeps his company in a shoe box under the bed

09:57:00 <Arnia> And I like for him to let me out occasionally

09:57:15 <ericP> i keep mine on irc

09:57:24 <ericP> must have good connectivity in the shoe box

09:57:45 <Arnia> Its a very modern shoe box with all the convieniences

09:57:53 <Arnia> Except Odoureaters

09:58:02 <sh1m> Arnia: I gave you airholes didn't I

09:58:33 <Arnia> If nothing else, its a well-branded shoebox

09:59:31 <ericP> swoosh

09:59:46 <sbp> ooh, they've brought back the random results

09:59:53 <Arnia> Nah, that was so passe. I've given it my own light touch

10:00:43 * ericP notices a bug in the SOAP putput

10:00:49 <ericP> the rdf output works better

10:00:59 <ericP> ditto n3

10:03:29 <Arnia> Anyway, the idea is to use soap and RDF/XML with adenosine in such a way that parameters can be passed to adenosine functions just by merging the graph described in the soap request with the closure

10:04:49 <ericP> closure?

10:05:45 <Arnia> Adenosine's execution model uses contexts to model closures

10:06:26 <ericP> not following -- should probably read

10:07:10 <ericP> 1st guess interpretation: foward-chained conclusions are asserted int an n3 context.

10:07:17 <ericP> but not sure that's what you meant

10:08:03 <Arnia> Well, the way that LISP, ML, ECMAScript and similar manage information is by wrapping it up in bundles called closures. A closure isolates the local variables and processes of a function (functions can contain multiple threads for example) from the rest of the system

10:09:15 <ericP> oh, kind of like a lamda function ?

10:09:58 <ericP> threads? really? that's got to be some crazy task manager.

10:10:12 <Arnia> Yes... ECMAScript and LISP actually have almost identical underlying semantics. Its a semantics very amenable to modelling in RDF

10:10:21 <Arnia> Well, threads is the wrong word really

10:10:43 <Arnia> A thunk represents an uncompleted computation... something that hasn't been resolved yet

10:11:37 <Arnia> A thunk spawns other thunks to fill in holes in its structure (this is how function calls work), but that thunk doesn't need to be physically run on the same location

10:11:44 <ericP> some code and variable bindings and an entry point into the code. passed to someone who will call that entry point?

10:12:11 <Arnia> This is how languages like Haskell claim that functional languages are more naturally paralisable

10:12:17 <Arnia> But yes, that's a thunk

10:12:28 <danbri> does Python have anything like closures yet?

10:12:47 <deltab> no

10:12:58 <ericP> 'cause thunks imply a marshalling of variables so you can pipe it over the wire as easily as execute it locally?

10:13:19 <ericP> (not run in same location)

10:14:05 <deltab> you can define functions that can access the names in their containing scopes, but they can't rebind them

10:15:57 <ericP> i interpreted this as: they can re-assign, but they are effectively functional, can't have any side effects on the passed variables.

10:18:43 * ericP mixes the conversation up a bit by throwing some declarations into the stream of interogatives

10:19:27 <kota_> kota_ is now known as kota

10:24:09 <Arnia22> Arnia22 is now known as Arnia

10:24:30 <Arnia> Sorry about that... I got cut off just after my last message

10:37:59 <sh1m> ericP: can you figure out a way to email me some sushi the canteen really isn't up to much and I'm hungry

10:41:51 * ericP deploys the sushipult

10:42:35 <sh1m> whole new meaning of flying fish

10:42:57 <sh1m> Arnia: I wonder if thought treasure could have come up with sushipult

10:43:14 <Arnia> sh1m: I'm tempted to install it and check :)

10:45:22 <Arnia> Anyway, ericP, did you respond to my comments about parameter passing and leveraging existing infrastructure?

10:46:24 <ericP> 2004-03-29T10:12:17Z <Arnia> But yes, that's a thunk

10:46:46 <Arnia> Damn... cut me off earlier than I thought

10:46:50 <Arnia> I then said -

10:46:51 <Arnia> If we use SOAP, we can do parameter passing into the thunk by a simple merge of a graph fragment describing the bindings of the parameters

10:46:56 <ericP> that's the last i saw from you before your unfortunate accident

10:46:59 <Arnia> Plus SOAP allows us to leverage the existing architecture of webservices and the web in general. We can bootstrap the semantic web quite nicely ;)

10:47:28 <ericP> interesting

10:47:49 <ericP> i wonder how much webservices arch will be usable for this

10:48:25 <ericP> s/arch/deployed tools/

10:48:41 <Arnia> Have you read my proposal for Bale? Defining a standard API for things that expose themselves in the form of RDF

10:48:59 <Arnia> Whether their underlying store is anything like RDF is irrelevant as long as they look like RDF across this API

10:49:27 <Arnia> And if everyone uses the same API then you can do transparent distribution of triples around the world

10:50:17 <ericP> haven't read it. sorry.

10:50:37 <Arnia> When queried for a triple (either directly or as part of a query operation) a Bale server can look at the URIs of the subject, predicate, object or context and apply a regexp which if true causes the Bale system to ask another server the same question

10:51:06 <Arnia> The triples become, effectively, placeless

10:52:11 <ericP> i'm tempted to dive into this, but i have several things on the stack and i'll never get out of here.

10:52:22 <ericP> hope to pick it up later

10:52:28 <Arnia> I'll be around ;)

10:53:56 <ericP> roger

10:56:01 <Arnia> Although if anyone else has comments, I'm around now too :)

11:03:27 <aharth> arnia: where can i read about this bale thing?

11:08:03 <Arnia> aharth: http://www.netalleynetworks.com/community/jgeldart/research/

11:08:17 <Arnia> aharth: Jabberwocky and Adenosine are the two bits

11:08:48 * aharth looks at the links

11:08:54 <Arnia> The docs are scarce because I'm still clothing the idea

11:09:20 <Arnia> I'm terrible for thinking out loud :)

12:03:37 <LeoBard> hi, has anybody heard of a WINFS (Microsoft, Longhorn) and SEMANTIC WEB connection ?

12:03:54 * LeoBard just read another article about winfs and there are some similarities

12:04:30 <sh1m> I was talking earlier about XAML and something they are doing which seems to be vaguely semantic web related

12:04:37 <sh1m> but I don't know if thats what you mean

12:04:53 <Arnia> A lot of their stuff is semantic web undercutting :p

12:05:12 <LeoBard> i just wondered if Microsoft digs SemanticWeb or if they shunn it at make their own stuff :-|

12:06:14 <sh1m> I am not the best person to ask about that, ericP might know how active microsoft are if he is still up

12:08:19 <LeoBard> thanx!

12:08:46 <sh1m> or danbri or DanC

12:11:29 <danbri> Here's a line I heard from someone at MS...

12:11:47 <danbri> (paraphrase) "we look at RDF somewhat like our attitude to functional programming..."

12:12:03 <danbri> ie. they're not _anti_ it, some of them use it, like it etc., but it isn't particularly mainstream for them

12:12:15 <danbri> (not much customer demand yet is something I hear occasionally)

12:12:45 <sh1m> is that ms corp. or ms research?

12:12:55 <danbri> I believe the winfs data structures are pretty RDFish

12:13:06 <danbri> but i've not studied, just heard from others

12:13:14 <danbri> ms generally, corp i guess

12:13:21 * Arnia has an evil plan to represent winfs as a bale

12:13:41 <sh1m> afaik ms research is one of the main driving people behind functional programming

12:13:51 <LeoBard> winfs is very near to rdfish. especially they treat files & emails & database more equally ( more like resources)

12:14:14 <Arnia> Yes, Simon Peyton-Jones works for MS Research

12:14:24 <danbri> they also have legit concerns about the danger of a fork in xml tech... that there is uptake in XML Schema, Query etc and that the RDF/SW stack isn't yet well integrated with investment in XML-level tools.

12:16:59 <LeoBard> ahja: winfs: http://msdn.microsoft.com/longhorn/default.aspx?pull=/library/en-us/dnlong/html/winfsdevpersp.asp

12:17:50 <Arnia> I reckon that an ontology to define a mapping from the RDF triple model to and from an arbitrary XML dialect would be a Good Thing tm (and I've stated as much with Oslo)

12:29:51 * Arnia needs to stop writing docs late at night

12:30:02 <Arnia> Either that, or take a proofreading course

13:04:33 <Wack> foo!

13:04:50 <Wack> whoohoo! i can speak again :/

13:05:14 <kao> obviously :-)

13:05:29 <Wack> well, yes, my IRC bouncer ate everything I said in this channel

13:05:47 <kao> oooh, i see

13:06:31 <Wack> here I was, almost thinking people were ignoring me, chatting along with the WinFS stuff while none of it actually came through :P

13:07:07 <sh1m> yes... your irc bouncer... we weren't ignoring you...

13:07:15 <Wack> looking at the irc logs confirmed my suspicion ;)

13:07:42 <Wack>http://longhorn.msdn.microsoft.com/lhsdk/winfs/daconwhatiswinfsdatamodel.aspx

13:07:43 <dc_rdfig> D: http://longhorn.msdn.microsoft.com/lhsdk/winfs/daconwhatiswinfsdatamodel.aspx from Wack

13:07:53 <Wack> D|"WinFS" Data Model

13:07:56 <Wack> D:|"WinFS" Data Model

13:07:57 <dc_rdfig> Titled item D.

13:09:07 <Wack> D:Looks like RDF(S) with a sprinkle of OWL and some specific property types (holding/referential)

13:09:07 <dc_rdfig> Added comment D1.

13:09:29 <Wack> btw, on WinFS, http://anopinion.net/ has some interesting information

13:10:03 <sh1m> chump it as a comment then ;)

13:10:57 <Wack> D:Regarding WinFS, [a blog by Mike Deem|http://anopinion.net/] has some interesting information and comments

13:10:58 <dc_rdfig> Added comment D2.

13:11:57 <Arnia> Question in case anyone has a suggestion. Should code be related to the URI of the closure, or be inside the closure

13:13:04 <Arnia> As in is it like :someFunction std:code (); :- {} or :someFunction :- { _:ENTRY std:code () }

13:14:54 <Arnia> I reckon the former is more sensible (code shouldn't be able to change between instantiations of the function) but may lead to some complications with regards variable referencing within the code (variables are BNodes within the closure's context)

13:16:16 <Wack> Arnia: btw, regarding 'bale'; what does it's API offer that serving application/rdf+xml and having an RDF query submitted using a POST payload does not?

13:17:56 <Arnia> Wack: Well, what I'm proposing is nearly that. Using SOAP as the API's mechanism. I feel its important to standardise however.

13:19:14 <Arnia> Wack: The reason I'm not using a plain webserver however is simple... the semantic web shouldn't tie itself down to unneccessary physicalities. The information expressed shouldn't depend on where it is stored in the real world or in the network topology of the internet

13:20:59 <bijan> logger, pointer

13:20:59 <bijan> See http://www.ilrt.bris.ac.uk/discovery/chatlogs/rdfig/2004-03-29#T13-20-59

13:24:20 <Arnia> The point isn't the choice of protocol (although I explained earlier why I thought SOAP would be a good choice) but rather that a choice is made and is stuck to. I envisage a world filled with bales able to talk to each other, able to share information, able to distribute the storage (and querying and discovery) of the graph

13:25:19 <Arnia> And because bale doesn't make any commitment about the underlying storage, you can have your fridge act as a bale. It can produce a representation of its state in the form of a graph.

13:25:41 <Arnia> (of course, this abstraction isn't limited to fridges ;)

13:25:45 <kao> i think i missed something there... what is a bale? any links?

13:26:20 <Arnia> kao: http://www.netalleynetworks.com/community/jgeldart/research/ - adenosine and jabberwocky links.

13:26:41 <Arnia> Haven't updated them with the results of last night and today's discussion.

13:27:24 <Arnia> (I'm currently trying to decide between two different graph structures for adenosine functions and they both seem equally good, or bad, atm)

13:32:38 <Arnia> Ok... I'm going with the code outside of the thunk. Putting the code in the thunk raises all kinds of security implications I don't want to deal with

14:28:59 * DanC looks around for bijan...

14:29:52 <DanC> anybody know of any SWSI ftf meetings since Apr 2003?

14:35:21 <LeoBard> anybody knows how to CVS checkout this branch here: http://dev.w3.org/cvsweb/2000/10/swap/

14:35:53 <DanC> er... that directory? or do you really mean branch in the cvs sense?

14:36:28 <LeoBard> jup, but I just found http://dev.w3.org/cvsweb/ and there is the info :-0

14:36:43 <DanC> more specific instructions are available at http://www.w3.org/2000/10/swap/doc/cwm.html#dev

14:36:54 <LeoBard> thanks !

14:37:18 <DanC> please report your experience with it to public-cwm-talk@w3/prg (and/or public-cwm-bugs@w3.org)

14:37:20 <DanC> as;ldfjk

14:37:25 <DanC> public-cwm-talk@w3.org

14:41:40 <DanC>http://www.daml.org/services/owl-s/1.0/

14:41:40 <dc_rdfig> E: http://www.daml.org/services/owl-s/1.0/ from DanC

14:41:47 <DanC> E:|OWL-S 1.0 Release

14:41:47 <dc_rdfig> Titled item E.

14:41:59 <DanC> E:* OWL-S is a OWL-based Web service ontology...*

14:41:59 <dc_rdfig> Added comment E1.

14:43:50 <DanC> E:also: [http://www.daml-s.org/owl-s/1.0/owl-s.html|OWL-S: Semantic Markup for Web Services]

14:43:50 <dc_rdfig> Added comment E2.

14:49:11 <Arnia> DanC: Very cool :)

14:51:35 <DanC> E:I remember something odd about the encoding of preconditions, but I can't find it.

14:51:35 <dc_rdfig> Added comment E3.

14:52:08 <DanC> E:cwm is having trouble with [http://www.daml.org/services/owl-s/1.0/CongoProcess.owl|CongoProcess.owl]

14:52:08 <dc_rdfig> Added comment E4.

14:56:39 <DanC> llyn.URISyntaxError: Hash in document ID - can be from parsing XML as N3! -#http://www.w3.org/2001/XMLSchema

14:56:46 <DanC> grumble. cwm need to give me a line number!

14:58:05 <DanC> (Pdb) p value

14:58:05 <DanC> u'#http://www.w3.org/2001/XMLSchema#string'

14:58:37 <DanC> aha!

14:58:37 <DanC> <process:parameterType rdf:resource="#http://www.w3.org/2001/XMLSchema#string">

14:59:15 <DanC> <process:parameterType rdf:resource="#&xsd;#string"/>

14:59:45 <DanC> E:indeed, there's a bug in CongoProcess.owl. Hmm... I wonder if there are any tools that grok this format. They don't seem to have checked their examples by machine.

14:59:46 <dc_rdfig> Added comment E5.

15:25:39 <DanC> phtpht. chump search down? I lose at http://search.rdfig.xmlhack.com/cgi-bin/htsearch?config=rdfig&restrict=&exclude=&method=or&format=long&sort=score&words=sumo

15:25:52 <LeoBard> @DanC: just tried out lookout.py - hey, I like this. its real fast.

15:27:08 <dajobe> yeah it's broke, try the chatlogs search; it'll probably list the chumped items

15:27:21 <Arnia> Hey libby

15:28:51 <DanC> I think timbl wrote lookout.py, based on an earlier hack I did. Glad you like it, LeoBard.

15:30:14 <LeoBard> yup. Especially the schema related stuff. I am heavily thinking about schemas.

15:31:48 <LeoBard> f.e. timbl noted something about the iCal schema in rdfig recently, noting that it is not ideal.

15:32:17 <LeoBard> this is very interesting: scraping the ontologies out of existing metadata or doing new "better" ones....

15:32:18 <grault> grault is now known as Earle

15:37:23 <LeoBard> LeoBard is now known as LeoBard|out

15:42:32 <M7U90nt> M7U90nt is now known as md-sick|bed

15:42:42 <md-sick|bed> md-sick|bed is now known as md-sick|bed|afk

16:11:47 <howardck> Hi :-)

16:13:02 <Arnia> Hello

16:17:23 <howardck> Hi :-) Anyone around?

16:17:44 <Arnia> I think I'm here... don't know about anyone else though :)

16:18:09 <howardck> Congrats. You're the first person I've ever spoken to on irc. :-)

16:18:19 <Arnia> I feel honoured

16:18:39 <howardck> I'm horribly confused, but then that's not unusual.

16:19:36 <howardck> How do I see who else is online, if that's even the proper terminology?

16:20:05 <Arnia> What client are you using?

16:20:19 <howardck> mIRC

16:21:02 <Arnia> There should be a list of channel participants on the right hand side in a big long list

16:22:07 <howardck> I can see the list. These people are all connected and following this inciteful conversation?!

16:23:12 <Arnia> Yes

16:23:24 <Arnia> If they're watching the channel that is

16:23:41 <Arnia> I'm currently in 10 channels but only chatting in two. This one and #plone

16:24:06 <howardck> Right. Thanks. OK, I think I'll sit back quietly and watch for a while ...

16:24:16 <Arnia> :)

16:45:33 <larsbot> larsbot is now known as lars|away

17:19:15 <ndw> Hi all

17:20:07 <ndw> So I'm supposed to give a talk on RDF etc. at the Center for Document Engineering at Berkeley. I submitted an abstract and got back a polite "could you sex it up a bit" reply.

17:20:25 <ndw> Bleh. I'm not terribly good at that sort of thing, but I took as tab.

17:20:28 <ndw> s/tab/stab/

17:20:41 <ndw> Anyone like to proof it and tell me if it's unmitigated crap?

17:20:43 <mattb> sexed-up rdf?

17:20:44 * mattb boggles

17:20:59 <ndw> It's at http://norman.walsh.name/scratch/cde

17:21:01 <sh1mmer> get an RSS feed from playboy.com

17:21:18 <ndw> The guy who asked me to give the talk suggested pasting my head on a super model's body

17:21:36 <sh1mmer> oh god

17:21:41 <sh1mmer> Norm I feel for you

17:21:47 <mattb> gosh

17:22:10 <mattb> how about livening it up by introducing it as "why machines will take over the world, but you'll like it" or somesuch?

17:22:16 <mattb> or "why machines won't take over the world"

17:22:29 <ndw> that's an angle

17:23:05 <darobin> ndw: you could refer to zool's porn0graph

17:23:10 <ndw> The Day the Machines Took Over the Web: Or How I Learned to Stop Worrying and Love the Semantic Web

17:23:16 <darobin> that would sex it up

17:23:29 <ndw> I'm not quite sure that's the direction "sexing it up" was really meant to go

17:23:30 <sh1mmer> hehe nice

17:23:56 <darobin> well hey, it's RDF and sex :)

17:25:46 * Arnia adds a new mouse to his shopping list

17:25:51 <ndw> soccos, three or four of you have had a look? Total suckage or send it in, what do you say?

17:26:08 <ndw> Dagnabit. Everytime I try to start a sentence with "so," I bug poor soccos. Sorry.

17:26:13 <sh1mmer> Arnia Joe any thoughts on sexing up presentations on rdf?

17:26:38 <sh1mmer> heh

17:27:03 <sh1mmer> Norm it is a wee bit dry, I would go but I already know what RDF is.

17:27:23 <sh1mmer> They probably want you to make some boldly inaccurate statement to attract bored undergrads

17:28:15 <ndw> Still dry, huh. Sigh.

17:28:32 <mattb> it does sound a little dry

17:28:32 <sh1mmer> where did update

17:28:41 <ndw> Bugger

17:28:43 <mattb> no paragraph with the word "syntax" in it is going to excite undergrads

17:29:16 <sh1mmer> ai might be a good angle

17:29:43 <sh1mmer> "How to create a superbrain" ;)

17:29:50 <ndw> Silly me, trying to describe what I'm actually going to talk about. Snort.

17:29:51 * darobin would be deterred by a sin tax

17:30:19 * Arnia gets out the graphics tablet as a substitute

17:30:26 <sh1mmer> darobin you would be impoverished by one

17:30:38 <darobin> quite

17:30:50 <Arnia> sh1mmer: Its a wonderful tablet, its just a little cumbersome to use instead of a mouse :)

17:31:06 <Arnia> Oh, and re: ideas... do an ontology of the sex industry

17:31:28 * darobin shuts up his stupid puns and goes sex up his own article, or in fact, simply put some content in it

17:31:33 <sh1mmer> or maybe use that cia thing from the TP

17:31:39 <sh1mmer> that was a bit secret agent

17:33:13 * sh1mmer waves at dajobe

17:33:21 <dajobe> hi

17:33:26 <sh1mmer> ugh the number of nicks starting with 'da'

17:34:10 <Arnia> Lazy typist :)

17:34:14 <sh1mmer> yep

17:34:16 <sh1mmer> very

18:16:10 <ChanServ> [#rdfig] This channel is logged and blogged: http://logicerror.com/rdfIRCWelcome

18:21:17 <danbri> danc, is http://www.w3.org/2000/10/swap/pim/itin2ical.n3 tracking recent changes to the rdf-ical vocab?

18:21:36 * danbri dohs, looks at date 2003/03/24, assumes it isn't

18:21:47 <DanC> dunno, danbri

18:22:05 <crschmidt> Mon Mar 29 12:21:36 CST 2004

18:22:31 <crschmidt> oh wait, that wasn't what you were asking, probably.

18:27:58 <sh1mmer> ndw got an update Norm? I would be interested to see what you sent

18:31:55 <howardck> Hi all. Wanna welcome a dazed and confused nubie to the wonderful world of irc? :-)

18:33:25 <sh1mmer> Hello

18:34:22 <ndw> sh1mmer, you can get the same URI again. I didn't end up changing much

18:34:41 <ndw> hello howardck

18:34:52 <sh1mmer> ndw I like the title :)

18:38:29 <howardck> hello shimmer and ndw. ndw, all that lovely rdf data attached to your blogged Nikon pix boggles my mind. I didn't know you had an RDF-capable camera! :-)

18:39:12 <ndw> It's just the exif data

18:39:43 <howardck> How'd you move it into rdf? I suspect xslt at work ...

18:39:55 <sh1mmer> actually it was the geosync on your calander itinary that i really liked

18:39:56 <ndw> uhm, jpegrdf.sf.net

18:40:15 <ndw> I'm still waiting to get a gps so I can give real locations for the photos

18:40:20 <howardck> merci. I'll go have a look ...

18:45:16 * ndw is also interested in geocaching

18:46:01 <LeoBard|out> LeoBard|out is now known as LeoBard

19:00:35 <peepo> I expect you're all well aware of the following issue:

19:00:47 <peepo> a search on google for: cc peepo svg

19:01:39 <peepo> produces a host of results related to drivers license rather than creative commons

19:02:18 <peepo> I suspect this is some sort of data mining of the web, but it is so bogus and defective, that it is a worry

19:02:35 <LeoBard> try link:cc.org ?

19:02:49 <peepo> a large percentage of the results are at fault, and I can only see this getting worse

19:03:55 <LeoBard> try: link:creativecommons.org

19:04:41 <LeoBard> and btw: http://www.google.com/search?hl=en&lr=&ie=ISO-8859-1&q=link%3Acreativecommons.org+peepo&btnG=Search

19:04:58 <LeoBard> this returned nothing, so peepo doesn't link to creativecommons.org

19:05:07 <LeoBard> hth ?

19:05:31 <LeoBard> in Semantic Web terms, you should search for uris rathre then terms :-)

19:07:07 <peepo> we use rdf in svg and link to http://web.resource.org/cc/PublicDomain

19:07:08 <LeoBard> interesting is that only 91 hits linkt to the Semantic Web home, this can't be true (google "link:www.w3.org/2001/sw" )

19:07:40 <kasei> i've always found the link: results slightly suspicious

19:07:42 <peepo> you can check this at http://www.peepo.co.uk

19:08:26 <LeoBard> +1 to the suspicous link: results. kasei is right

19:08:53 <kasei> they seem more like fuzzy results. where "fuzzy" means "completely wrong" :)

19:10:26 <LeoBard> but at least they show somehow what we mean with "semantic search" in the context of URIs. I dream of a Semantic Web where we have a URI aware google.

19:10:59 <peepo> It's not so much that I am concerned with google in this instance,

19:11:13 <LeoBard> (better: RDF aware google, that uses URIs to identify stuff)

19:11:52 <peepo> rather that some system similar perhaps to RSS is multiplying a simple error across a large number of sites, and that will squew any search engine

19:12:08 <peepo> It looks as though it might be true

19:20:17 <danbri> has anyone here managed to get an iCalendar file out of Amadeus' https://www.checkmytrip.net/ ?

19:20:52 <danbri> ...there is an 'add to calendar' feature, but it seems to require a plugin that ether isnt' working or i don't have... very confusing for what should be a simple file download

19:21:01 * danbri trying on linux and macosx

19:22:30 <danbri> they use www.inforetriever.com, which looks pretty cheesy

19:23:54 <danbri_dna> hmm src of http://my.infotriever.com/download/nspluginscript.html confirms that

19:24:21 <danbri_dna> if (detectNetscape6() == false) {

19:24:21 <danbri_dna> installNS4Plugin();

19:24:23 <danbri_dna> ...

19:24:33 <danbri_dna> also alert("This plug-in only runs on Windows 9x/NT/2000/Me/XP/2003.");

19:31:41 <soccos> ndw: :) wakes me up to RDF every now and again

19:32:18 <Arnia> danbri_dna: Can't get to that site

19:32:59 <danbri_dna> ... https://www.checkmytrip.net/ works for me

19:33:07 <ndw> sorry soccos

19:33:22 <Arnia> inforetriever I meant

19:33:23 <soccos> np

19:34:31 <LeoBard> inforetriever wants to install itself to windows. Its more than a plugin, it works only with Outlook and Act. https://my.infotriever.com/download/noactivex.html?

19:39:35 <danbri_dna> It seems to try to be a middleware between all calendar tools that you might have on desktop. Palm etc too.

19:39:54 <danbri_dna> Better imho to leave that to operating system dispatcher, rather than mess around inside browser like this.

19:41:19 <LeoBard> so leave it to Microsoft ? :-)

19:41:43 <LeoBard> My Semantic Desktop approach goes to the direction to provide RDF interfaces to all this stuff.

19:42:21 <LeoBard> would be nice to just say "I want this iCal/RDF in my calendar now"

19:42:58 <LeoBard> well, actually i am NOW hacking on this :-)))) java/jena app prototype that integrates outlook and ical and so on...

19:44:09 <LeoBard> you have the same problem with contacts. btw it all reminds me of good ol: http://esw.w3.org/topic/PersonalInformationDisaster

19:44:43 <LeoBard> so danbri, you are right now in http://esw.w3.org/topic/PersonalInformationDisaster

19:46:02 <danbri_dna> I see my current problem as the browser/desktop interface (plugins and helper-apps) is somewhat crude, ie. course-grained...

19:48:07 <LeoBard> well, at the moment the solution many apps provide is to integrate to outlook....

23:25:55 <DanC> .time

23:25:55 <datum> Mon, 29 Mar 2004 23:25:56 GMT

23:26:04 <DanC> hmm... I seem to have lost contact with MIT

23:26:19 <Arnia> Is that like losing contact with Houston?

23:26:34 <DanC> "ground control to major Tom..."

23:27:32 * Arnia is sure there will be an ontology of losing things soon

23:27:42 <Arnia> The semantic web equivalent to BrainFuck

23:33:44 <danbri> hey, where'd the mothership go?

23:33:57 <DanC> yeah, really, danbri.

23:34:26 <danbri> closest a traceroute here sees is B24-RTR-2-BACKBONE.MIT.EDU

23:34:43 <DanC> 'sup, eikeon? I'm still curious about cwm/rdflib API convergence, on occasion. you?

23:35:04 <DanC> in particular, I'd like to try to port toIcal.py to rdflib

23:35:27 <eikeon> DanC, sure. I'd like to do more.

23:35:37 <eikeon> ... and am back making releases.

23:36:00 <Arnia> yay :)

23:37:11 * DanC wonders if rdflib is faster than cwm for the toIcal.py task...

23:37:22 * eikeon shifts into evening gear... have time to work on something tonight. Where does toIcal.py live?

23:37:52 <danbri> MIT :/

23:38:16 <DanC> . http://www.w3.org/2000/10/swap/pim/toIcal.py

23:38:39 <DanC> hmm... I can get it.... I guess I have it cached

23:38:39 <eikeon> DanC, Thank you.

23:38:45 <Arnia> What's happened to MIT?

23:38:58 <danbri> It's playing hard to GET

23:39:02 <eikeon> lol

23:39:14 <Arnia> What is it with academia and awful puns?

23:39:27 * DanC tries apt-cache search rdflib... loses

23:39:40 * eikeon is able to GET it... wonder if I had it cached too.

23:39:49 * DanC swats away a cookie from rdflib.net

23:40:36 <eikeon> And the cookie is not even being used for anything :(

23:40:38 * DanC grabs http://rdflib.net/2004/03/15/rdflib-2.0.1.tgz

23:41:07 <DanC> so you're using some web server foo that has cookies on by default? do you know what it is so I can bonk them?

23:41:32 <eikeon> Yeah... it's Redfoot... I'm at fault :|

23:41:39 * DanC bonks eikeon

23:41:54 * danbri nudges DanC in the direction of uk/us slang dictionaries

23:42:21 <danbri> . http://www.netfunny.com/rhf/jokes/95q4/uk.html etc

23:42:32 <DanC> sigh. yes, all words mean "copulate".

23:42:43 <danbri> I doubt Eikeon minds, he's American too

23:43:17 * Arnia likes British slang

23:43:28 <DanC> in the midwestern united states, folks use "bonk" to mean "hit on the head, hard enough to go 'ow!' but not hard enough to need, say, medical attention"

23:43:33 <eikeon> Wow... Ireland going smoke free!?!

23:43:45 <eikeon> ... in Pubs.

23:43:49 <danbri> It means that too here, just to complicate things :)

23:44:08 <danbri> s/too here/here to/

23:44:40 <Arnia> danbri: Newcastle's the most fun place to be dropped if you like getting beaten up for saying the wrong thing :p

23:44:42 <DanC> darn; why doesn't "python setup.py test" Do The Right Thing?

23:45:10 <DanC> ImportError: No module named distutils.sysconfig

23:45:10 <DanC> ew.

23:45:26 <Arnia> danbri: Actually, Sunderland is worse because foreigners (ie everywhere else but the northeast) tend to call Sunderlanders 'geordies' which is grossly offensive to them

23:45:30 * DanC installs python2.3-dev

23:46:03 <DanC> ok, now I'm up to:

23:46:05 <DanC> error: invalid command 'test' (no module named 'distutils.command.test')

23:46:22 <DanC> ah... run_tests.py

23:46:32 * eikeon looks for distutils documentation

23:47:33 <DanC> the example.py seems to work.

23:47:40 <DanC> hmm... FOAF["Person"] ... why not FOAF.Person ?

23:48:42 <DanC> and oh yeah... Literal("donna") is independent of the store.

23:48:49 <eikeon> Yeah... nobody seems to like the choice of [] over . -- I've even started using .

23:49:01 * DanC isn't sure which way Literal should work

23:49:31 <eikeon> I remember you commenting on that before... am not sure either.

23:50:02 <DanC> ok, wanna hold my hand while I port toIcal.py ?

23:50:18 <eikeon> Sure.

23:51:07 <DanC> 1st swap/cwm-ism: from myStore import Namespace, load, setStore # http://www.w3.org/2000/10/swap/

23:52:23 * DanC hunts for rdflib equivalent of load

23:52:48 * eikeon tries to get to http://www.w3.org/2000/10/swap/ -- no joy.

23:53:09 <DanC> try web4.w3.org/...

23:53:10 <eikeon> load is a method on TripleStoore

23:53:23 <eikeon> er, TripleStore

23:53:41 <DanC> so you don't have a convenience function that takes a filename, opens it, parses it, and hands back a store?

23:55:12 <DanC> but this should work, yes? sts = TripleStore().load(addr)

23:55:20 <DanC> er... does load() return self? prolly not...

23:55:47 * eikeon checks... but guesses not.

23:56:41 <eikeon> ... which seeing as I came from Objective-C years ago... I would be returning self.

23:57:54 <eikeon> I should make the 2.0.2 release later tonight as well. Meant to make it last Monday.

23:57:56 <DanC> I no longer like returning self as a rule. it creates a false data dependency in many cases.

23:58:05 <eikeon> Maybe I can add some of these changes.

23:58:23 <DanC> ok, after making the naive changes, I lose thusly:L

23:58:25 <eikeon> ... not sure it is very Pythonic either.

23:58:25 <DanC> AttributeError: 'TripleStore' object has no attribute 'each'

23:58:51 <eikeon> Hum.

23:59:01 <DanC> perhaps each() is like triples()

23:59:12 <DanC> for cal in sts.each(pred = RDF.type, obj = ICAL.Vcalendar)

23:59:29 <eikeon> each is a CWM method... that is like rdflib's triples... IIRC