Experimental IRC log swig-2007-08-08

Available formats: content-negotiated html turtle (see SIOC for the vocabulary)

Back to channel and daily index: content-negotiated html turtle

These logs are provided as an experiment in indexing discussions using IRCHub.py, Irc2RDF.hs, and SIOC.

14:03:45<drrho>ACTION is away: (shopping food)
14:17:37<bblfish>phew... I am starting to understand pgp and all that stuff. Very good book I am reading called "Web Security, Privacy & Commerce" 2nd edition, by O'Reilly
14:18:34<bblfish>now I understand why Jermey Carroll had to develop his graph signing algorithm
14:19:17<bblfish>If I want to sing my foaf file, it is going to be problematic because currently it returns a number of representations.
14:19:44<bblfish>I could sign just the rdf file, and just the n3 file for example
14:20:58<bblfish>so I instead of signing http://bblfish.net/people/henry/card I would have to sing http://bblfish.net/people/henry/card.rdf or http://bblfish.net/people/henry/card.n3
14:21:08<timbl>You should sign just one.
14:21:56<timbl>You connect the sig to that file and anyone who wants can check the sig and check the data from the sigend document.
14:22:02<timbl>IMHO.
14:22:19<timbl>Tryingto sign a graph, whch invlves canonicalizing a graph, is unnecessary.
14:22:23<timbl>You sign the doc.
14:22:25<bblfish>So I add something like <http://bblfish.net/people/henry/card.rdf> signed <http://bblfish.net/people/henry/card.sig>
14:22:56<timbl>Or even card.rdf and sig.rdf
14:22:59<bblfish>and then I also say <http://bblfish.net/people/henry/card.rdf> iana:alternate <http://bblfish.net/people/henry/card.n3>
14:23:26<timbl>You can sign things with cwm, and the results you put in RDF or N3 output
14:23:37<bblfish>oh. nice
14:23:46<bblfish>using pgp?
14:24:11<timbl>It uses the raw public key algorithms from pycrypto.
14:24:12<bblfish>ACTION Ah I was wondering how to add a signature to the file itself
14:24:34<timbl>There is a cert chain example in the regression suite even I think.
14:25:09<timbl>Ah ... adding a sig to the file itself is always tricky as yo have to exclude the sig from the signing, ... a sep file is much easier.
14:25:50<timbl><> ex:signedBy <sig.rdf> You can point to teh sig from the file
14:26:00<bblfish>yes, that's what I thought.
14:26:37<bblfish>(I meant that it would be tricky to add a sig to the file itself without there being some convention in the language that is)
14:27:49<bblfish>yep I see. I would have two sig files, one for the n3 and one for the rdf represenatations
14:30:22<bblfish>mhh the web of trust ontology does not have a relation from a document to its signature... http://xmlns.com/wot/0.1/
14:30:40<bblfish>wrong!
14:30:41<bblfish>it does
14:31:14<bblfish><card.rdf> <http://xmlns.com/wot/0.1/assurance> <sig.rdf.asc> .
14:31:29<bblfish><card.n3> <http://xmlns.com/wot/0.1/assurance> <sig.n3.asc> .
14:32:18<bblfish>and better just use <> and put the correct statement into each file .
14:33:35<bblfish>thanks timbl. I am going to try to add a signature to my file and blog about this :-) Interesting one can even use this to publish one's public key...
14:37:19<bblfish>ACTION this is so much fun
14:44:18<bblfish>"and better just use <> and put the correct statement into each file " except that if I do that then if someone merges the graphs from the two documents, there will be falsity
14:45:59<timbl>bblfish, you can publish my pgp key too in your foaf file.
14:46:00<bblfish>so I have to use the full URIs and then point out that <http://bblfish.net/people/henry/card> iana:alternate <card.rdf>, <card.n3>
14:46:12<timbl>Then we have a chain of trust
14:46:28<timbl>iana: ?
14:46:46<bblfish>http://bblfish.net/work/atom-owl/2006-06-06/AtomOwl.html
14:46:53<timbl>gen:contentTypeSpecific
14:47:18<bblfish>atom has link relation alternate that is specified by iana
14:47:27<bblfish>for linking to alternate representations
14:47:46<quoll_>timbl: wouldn't a chain of trust require a signature on the the foaf file? Sorry if I've missed it, but I haven't seen a description of how to do that
14:49:33<timbl>see http://www.w3.org/2000/10/swap/test/crypto/Makefile
14:49:59<timbl>for a public-key cert generation and signing etc using cwm
14:50:03<timbl>You were using pgp, though?
14:50:46<quoll_>well, I use gpg, but the principles are the same
14:50:59<bblfish>this is where the iana alternate relations comes from: http://www.atompub.org/rfc4287.html#rfc.section.4.2.7.2
14:51:21<quoll_>my concern was the signature not being incorporated into the file
14:52:06<quoll_>though I guess that's not really a problem (now that I think about it)
14:52:13<timbl>If you incorporate the signature into the file, you have to have a sig algo which excludes the signaturev itself, which is what a bunch of the complexity from XML DSig is.
14:52:56<timbl>ACTION adds card_rdf.asc : card.rdf
14:52:59<timbl>ACTION gpg --output card_rdf.asc --clearsign card.rdf to his Makefile
14:55:22<timbl>Hmmm . looks like me key has rotted
14:55:23<bblfish>What would be the advantage of putting your signature in my foaf file? I already point to your foaf file... Should you not have your foaf file point to your public pgp signature?
14:56:56<timbl>Sorry, you put my pup key in your foaf file
14:57:04<timbl>pub
14:57:22<bblfish>what for?
14:57:24<timbl>"I know a person whose name is tim and whose pgp key ois this"
14:57:44<timbl>Then when someone retreives my foaf file they can check it has not been tamprered with
14:58:05<timbl>Your FOAF file becomes a cert for me.
14:58:11<bblfish>ah. ok. Distributed public kehs
14:58:16<bblfish>keys. nice
14:58:19<timbl>yup
14:58:31<bblfish>still you should point to your public key :-)
14:58:42<bblfish>Are these things not super long though?
14:59:56<timbl>Well, quite long -----BEGIN PGP PUBLIC KEY BLOCK-----
14:59:57<timbl>Version: GnuPG v1.2.4 (Darwin)
14:59:57<timbl>mQGiBEESw2YRBAC3FASezIEFh1B/5dy6jjbd8ITDXXG2I+Jf+/WYzLJEwXNkik6h
14:59:57<timbl>C6T08UmHg1Cf2IDXHoyAMshM3bElJU0KRjRkNFXE3vCBKYjR0anNAtVLCQEdGPMS
14:59:57<timbl>1wpgG9vRqpq/uDp4fv94sr5CDLbXeDJv6Ope1zvboFENO8sT/YIvRUsGvwCgyD0F
15:00:01<timbl>vFX+AKGxxCnEEfUpHRgUhq8EAJTlu8cymTw4qIjtLTiKZ4FgAOXiy7hkm9C+TYC6
15:00:01<timbl>qIHRqZfPmGo6k3YJtSv8m17rnT8Wm7hBFCq9BnYSoqDAOqprvMJyCiMRBvf+pIh9
15:00:01<timbl>US9hAjqdNyfX1HnmXMc7gzRSuDrt4wZzIxkuRIZ0e01sYOTluqu0g7cqW5z+oIiJ
15:00:03<timbl>giobA/0cRK0qwPslCEPPeAcPZuWx+rpypPqFlseU18CoAlf25bgm4Gn9lz+L3EV1
15:00:05<timbl>ByQx6KlqrKu33wu8IGTr3oO/cSbwNxXfxbTZQwi6krRJSS+Fkfv/Q5K3oyKTja6r
15:00:07<timbl>nFq4iQfy/IS7i9EOPOkaqkPHPC0kmX8BiyMsuwI2lCcCuA2n4rQuVGltb3RoeSBC
15:00:09<timbl>ZXJuZXJzLUxlZSAoTUlUL0NTQUlMKSA8dGltYmxAdzMub3JnPohkBBMRAgAkBQJB
15:00:11<timbl>EsNmAhsDBQkFo5qABgsJCAcDAgMVAgMDFgIBAh4BAheAAAoJEDY6vzCfw9V+MvMA
15:00:13<timbl>nj/dwb+jphtqFOaHTAbwZ1P+MDsjAJ9Op7Z7GQYmT3YRwY6SjPMUEAbk2bkBDQRB
15:00:15<timbl>EsNoEAQAvB+EVh6pda14jrMgApACfHrUqq+8UvpO6gopKFADt3/IAINAeiHQdV/P
15:00:17<timbl>R8BbjbaaEi7VuC2otgPGy589YaHD8vLbpT4k/n5Tb7axNyiNWVhGhoaeGAb9BOVM
15:00:19<timbl>W+G/WLOBmDIJWEhL807C3Xoi9FOvZLLgF+8W03+YhhWAspHYh58AAwUEAIQeMgAl
15:00:21<timbl>ymUluuhesM8VCKnwghYDZz33H4hCUA0nGbGNQv61LUClqlnlbq3Aq6aZwEtRplx1
15:00:23<timbl>1XyRf+hcqq7JTcZFmdT/syUwRKlxUmkUxWx60GmAdKAMfWCPV3wilL6JyrGOfffy
15:00:25<timbl>wP/jytbW876cp9JKYJVEdCy8vZa8CFvJWo+OiE8EGBECAA8FAkESw2gCGwwFCQWj
15:00:27<timbl>moAACgkQNjq/MJ/D1X5B8QCdEFmFcSnvErlRNAL0wO57O3WA2XIAn1w4i13VIr9H
15:00:29<timbl>ZolpWlZNX1WNwrVU
15:00:31<timbl>=C+Je
15:00:33<timbl>-----END PGP PUBLIC KEY BLOCK-----
15:00:35<timbl>:)
15:00:37<timbl>ACTION apologizes for the long paste
15:00:53<bblfish>If I do this for everyone I know that's going to make for a long foaf file
15:01:07<bblfish>Perhaps not, if I don't know that many people.
15:01:38<timbl>That is better than knowing too many.
15:01:49<quoll_>what's the problem with a long foaf file? The computer shouldn't mind
15:02:19<timbl>We could certainly for example fix tabulator to recognize keys and suppress their display
15:02:25<quoll_>ACTION laugh when thinking of how many people timbl must associate with on a regular basis
15:03:40<bblfish> <http://xmlns.com/wot/0.1/assurance> points to a resource I think
15:03:42<hermit>There is a limited number of people I would be prepared to sign FOAF files of.
15:04:13<bblfish>I could just put the signature on my server, and point to it
15:04:35<hermit>hmm signin things I get gpg: no default secret key: unusable secret key
15:04:38<hermit>darn
15:05:54<bblfish>Are people happy with the wot ontology?
15:06:53<HarryH>Danja - please come to GRDDL WG meeting
15:07:24<bblfish>ACTION probably herding his cats
15:08:54<bblfish>timbl: are you going to use wos to point to your signature?
15:09:27<bblfish>I mean wot
15:10:01<timbl>IIRC wot is not properly online
15:10:58<bblfish>I know I sent it to danbri yesterday, content neg problem
15:11:27<bblfish>the file is at http://xmlns.com/wot/0.1/index.rdf
15:11:49<bblfish>there was a similar problem with foaf a few months ago
15:13:19<HarryH>ACTION [DONE]: jjc to organize library tests; link from http://www.w3.org/2001/sw/grddl-wg/library
15:13:37<HarryH>ACTION [CONTINUES]: jjc to dispose of @@s in http://www.w3.org/2001/sw/grddl-wg/library (and in docs of 3 txforms)
15:13:41<HarryH>Zakim, open item 13
15:13:47<dajobe>HarryH: wrong channel
15:13:50<dajobe>and server
15:13:52<IvanHerman>ACTION things Harry is on the wrong channel
15:17:41<dajobe>http://www.w3.org/2007/06/OWLCharter
15:17:52<dajobe>B:|OWL Working Group (proposed) Charter
15:18:07<dajobe>B:including possible (new?) XML syntax for it
15:20:00<bblfish>? I don't understand the need for a syntax for OWL
15:20:16<dajobe>B:I would suggest they start no earlier than both DAWG and GRDDL WG finish
15:20:17<bblfish>Use Turtle
15:20:59<dajobe>yeah I must get back to that, and hide from IvanHerman
15:21:07<dajobe>ACTION -> 7.5hrs of meetings
15:26:15<drrho>ACTION is back
15:26:37<bblfish>wot seems to make a distinction between the wot:PubKey and the wot:pubKeyAddress . Looks like they should be the same no?
15:28:25<bengee>ACTION wonders if they are going to introduce new ontology language flavours until the SemWeb's entire marketability is destroyed...
15:29:04<bblfish>A:= http://xmlns.com/wot/0.1/
15:29:11<timbl>ACTION guesses they mean wot:publicKeyFile
15:29:17<timbl>whcoh is diff from key
15:29:32<bblfish>A:| Web of trust
15:29:38<timbl>when something has "addresss" which takes a string that is a pain.
15:29:42<Arnia>bengee: depends upon what you consider to be a good ontology semantics, or even if you believe ontology is the best way to represent knowledge
15:29:43<timbl>but logical.
15:29:50<bblfish>A: discussion on the wot ontology and adding these to our foaf files
15:30:05<timbl>Should have "keyfile" which takes <> not ""
15:30:35<timbl>There is a new syntax called out in the charter?
15:30:51<bblfish>Ok, well that is probably just me now understanding enough about a PubKey, and its relation to a pubkey file
15:31:02<Arnia>bengee: I personally don't think OWL is the right approach, most do. At least we have the freedom to pick the logic which works for us and our applications :)
15:31:12<Arnia>ACTION would hate to be forced to use OWL ;)
15:32:21<bblfish>the idea of a new syntax just for owl is pretty weird. -1 from me
15:32:25<bengee>yeah, maybe yet another ont option just accelerates the trend to focus on the more practical parts of the semweb menu..
15:32:35<KjetilK>oooooh, the Arnia has returned! :-)
15:32:54<Arnia>Indeed. Currently hitting prolog with a pointed stick
15:33:04<KjetilK>:-)
15:33:21<KjetilK>Any semantic desktop work lately?
15:33:45<bblfish>ACTION off to create a pgp public key for bblfish
15:34:08<KjetilK>ACTION notes that he should meet bblfish one day to exchange signatures
15:34:12<Arnia>KjetilK: some... working on other bits of the puzzle atm though (and my PhD of course)
15:34:27<bblfish>ACTION :-)
15:34:48<KjetilK>ACTION pokes timbl too about that signature... :-)
15:35:02<Arnia>KjetilK: I'm currently writing an experiment in preference systems for the semantic web by using a non-axiomatic logic as a semantics for RDF
15:35:28<Arnia>(this is in parallel to writing a category theory account of RDF/RDFS/SWRL etc)
15:35:37<Arnia>ACTION should stop splitting himself into so many bits
15:35:47<KjetilK>hehe
15:35:52<KjetilK>Arnia: cool
15:36:24<KjetilK>it seems like redland is going into KDE 4
15:37:02<Arnia>aha
15:37:35<KjetilK>which should expose a lot more real users to a semantic desktop
15:37:46<Arnia>I think my main problem is that I hate coding... it is so much more messy than thought :)
15:37:52<KjetilK>of course, it will take a long time, but when the infrastructure gets there
15:37:57<KjetilK>hehe
15:38:00<Arnia>Unsurprisingly I'm a fan of haskell
15:38:08<bblfish>how long should my key be valid for?
15:38:14<Arnia>KjetilK: still needs a frege-like layer I think
15:38:25<KjetilK>bblfish: mine doesn't expire
15:38:44<KjetilK>you can always revoke it if you have the secret key and remember the passphrase
15:38:57<Arnia>Actually, I'm beginning to think the semantic desktop is impossible within current dominant UI and SE paradigms
15:39:07<Arnia>ACTION will write a paper on that rather than babble though
15:39:16<KjetilK>Arnia: mmmm, ok
15:39:38<KjetilK>still, it would be enormously cool to have the data of digikam on the semweb
15:39:57<KjetilK>and a policy-aware-web thingie to declare who is allowed to see them
15:41:09<Arnia>KjetilK: oh yes, but I don't see it being particularly sensible until 'trust' (not as in Web-of-Trust, but as in the confidence one has in a given truth-valued statement) is part of the meaning of the statements
15:41:24<Arnia>But that's what my experiment is all about :)
15:41:31<KjetilK>ah :-)
15:41:32<bblfish>the trick is always to start small
15:41:52<bblfish>if you need to make huge changes to get something going, then you have started too big
15:41:55<KjetilK>Arnia: let us hope for a positive outcome
15:42:17<bblfish>I suggest helping out on the Baetnik address book, as a good starting point
15:42:33<bblfish>its at https://sommer.dev.java.net in the repository there
15:43:02<Arnia>bblfish: that's why I'm looking to give RDF a non-axiomatic semantics. No need to change the syntax... but you gain the additional semantics needed
15:43:20<Arnia>ACTION should prod someone to port Frege to Cocoa
15:43:53<bblfish>ah. look forward for that paper. Just learning about gpg right now...
15:44:06<Arnia>Don't have time myself, but I can provide a lot of notes making it mostly a matter of Ikea-coding
15:58:18<KjetilK>I've been on vacation, moving apartments and getting into a new job, so I have fallen a bit behind
15:58:41<KjetilK>I was just wondering what the status of the vcard stuff is right now?
16:00:45<bengee>they drifted off *slightly* and discuss a new RDF now..
16:01:02<KjetilK>hehe
16:01:32<KjetilK>yeah, I think I caught a brief glimpse of problems of literals in lists in RDF/XML
16:04:49<bblfish>Garret Wilson is a bit tiring I find
16:05:17<xi_>yeah he should make his point and get over it
16:05:46<xi_>i did find the 'literal resources' discussion useful. at least 20% of it - ended up doing something similar in the backend of a store - what comes out appears as normal RDF though..
16:05:58<bblfish>I think I have pointed out to him that 3 = "3"^^xxx:int; rel:int "3".
16:06:11<bblfish>I mean they are all equal
16:06:22<bblfish>yes. the discussion was helpful
16:07:18<xi_>speficailly, i added a 'read/write (non-meta)-data associated with this URI) and just throw the literals in a binary blob, and identify the blobs by the hash of the contents, and use that URI in place ofa literall internally
16:07:51<xi_>ACTION hasnt read 'contextual properties' brouhaha yet
16:45:18<bblfish>18:33:54 - ~/Documents/card
16:45:18<bblfish>hjs@bblfish:0$ gpg --list-keys henry.story@bblfish.net
16:45:19<bblfish>pub 1024D/C7DEAB05 2007-08-08
16:45:19<bblfish>uid Henry J. Story <henry.story@bblfish.net>
16:45:19<bblfish>sub 2048g/115243F3 2007-08-08
16:45:36<bblfish>I can't remember what my key length is, and I don't know how to interpret that
16:45:40<bblfish>is it 1024?
16:48:05<sbp>the public key, yeah. the subkey is 2048
16:48:09<sbp>I dunno what the D and g stand for
16:48:25<bblfish>ok thanks..
16:48:53<sbp>oh, type
16:48:57<sbp>D is DSA, g is ElGamal
16:49:05<bblfish>yep
17:19:23<bblfish>A: my new foaf file http://bblfish.net/people/henry/card has a pointer to my public key
17:19:50<bblfish>A: mail me if you find any problems with it
17:20:05<bblfish>ACTION got to go
17:49:39<bblfish>A: if possible mail me something encrypted with my signature :-)
19:00:30<bblfish_>mhh I can't set the base of a local file with cwm
19:00:44<bblfish_>hjs@bblfish:2$ cat test.n3
19:00:44<bblfish_><test> a <http://eg.com/Something> .
19:01:50<bblfish_>hjs@bblfish:2$ cwm test.n3 --ntriples --base=http://bblfish.net/people/henry/
19:01:51<bblfish_>
19:01:51<bblfish_> <file:///Users/hjs/Documents/card/test> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://eg.com/Something> .
19:02:38<timbl_>bblfish, cwm undocumented support for @base
19:02:49<bblfish_>ah. Thanks
19:04:15<bblfish_>how does it work?
19:04:28<bblfish_>@base <http://eg.com> .
19:05:57<bblfish_>?
19:07:26<bblfish_>ah I found ./test/n3/n3-full.tests:
19:07:38<bblfish_>I mean ./test/n3/n3-full.tests it uses @base
19:13:37<timbl_>ACTION is happy the syntax was obvious :)
19:13:48<timbl_>Not it can take a relative uri, rel to previous base
19:14:36<bblfish_>is this right:
19:14:37<bblfish_>hjs@bblfish:3$ cat test.n3
19:14:37<bblfish_>@base <http://bblfish.net/> .
19:14:37<bblfish_><test> a <http://eg.com/Something> .
19:15:19<bblfish_>I get this
19:15:20<bblfish_>hjs@bblfish:3$ cwm test.n3
19:15:20<bblfish_>#Processed by Id: cwm.py,v 1.194 2007-08-06 16:13:56 syosi Exp
19:15:20<bblfish_> # using base file:///Users/hjs/Documents/card/test.n3
19:15:21<bblfish_> # next char: u'@'
19:15:21<bblfish_> Failed to parse file:///Users/hjs/Documents/card/test.n3
19:15:21<bblfish_> file:///Users/hjs/Documents/card/test.n3
19:15:23<bblfish_>Traceback (most recent call last):
19:15:25<bblfish_> File "/Users/hjs/Programming/w3c/2000/10/swap/cwm.py", line 746, in <module>
19:15:27<bblfish_> doCommand()
19:15:29<bblfish_> File "/Users/hjs/Programming/w3c/2000/10/swap/cwm.py", line 454, in doCommand
19:15:31<bblfish_> why=myReason)
19:15:33<bblfish_> File "/Users/hjs/Programming/w3c/2000/10/swap/webAccess.py", line 225, in load
19:15:35<bblfish_> p.feed(buffer)
19:15:37<bblfish_> File "/Users/hjs/Programming/w3c/2000/10/swap/notation3.py", line 217, in feed
19:15:39<bblfish_> "expected directive or statement")
19:15:41<bblfish_>swap.notation3.BadSyntax: Line 1 of <file:///Users/hjs/Documents/card/test.n3>: Bad syntax (expected directive or statement) at ^ in:
19:15:44<bblfish_>"^@base <http://bblfish.net/> .
19:15:46<bblfish_><test> a <http://eg.com/Somet..."
19:16:28<timbl_>you need a pretty recent cwm
19:16:39<bblfish_>that's right from CVS
19:16:41<bblfish_>I think
19:16:45<timbl_>ACTION wonders wehther he checked in then
19:22:04<bblfish_>just in case
19:22:05<bblfish_>hjs@bblfish:2$ cat CVS/Root
19:22:05<bblfish_>:pserver:anonymous@dev.w3.org:/sources/public
19:36:51<bblfish_>timbl: would it not be more interesting instead of me having your public key in my foaf file, for me to point to a file that signs your public key?
19:37:05<bblfish_>with my public key
19:38:15<bblfish>ACTION got to go
21:25:48<lordi>is there any RDF database engine thats not written in java and is fast enough for a website? e.g. a query over thousands of triples stored
21:26:28<kasei>redland?
21:28:51<lordi>thx, I'll have a look
22:22:55<lordi>given that rdf, http://dpaste.com/16393/, is it possible to write an sparql query returns all the parent classes of "poodle"... here its "dog, animal"
22:39:54<tommorris>lordi: got it
22:41:09<tommorris>see http://dpaste.com/hold/16398/
22:50:22<lordi>tommorris, hm doesn't work for me.
22:50:38<tommorris>i tested it in sparql.org - it may be buggy
22:51:00<CaptSolo>nice lift-off of the space shuttle
22:51:20<tommorris>Oh, you need to put in the URI of your file
22:51:27<lordi>but thx anyways
22:51:30<tommorris>go to http://sparql.org/sparql.html
22:51:33<tommorris>paste the query in
22:51:42<tommorris>then pop http://dpaste.com/16393/plain/ in the target graph uri field
22:53:04<lordi>yea, that does it
22:53:47<lordi>but I'm not succesful with redland
22:53:52<lordi>maybe that is buggy, then

Back to channel and daily index: content-negotiated html turtle