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.
| 03:44:27 | <PovAddict> | o_O there's facebook apps for everything |
| 03:44:27 | <PovAddict> | look at this poll: "Toilet paper - Over or Under?" |
| 03:44:27 | <PovAddict> | http://pweb.st.tbxing.com/user_files/fb_1343154428/speedpoll/TCphpSYSFqI_1243624815.png |
| 03:44:27 | <PovAddict> | http://pweb.st.tbxing.com/user_files/fb_1343154428/speedpoll/TCphp7FdQPp_1243624820.png |
| 03:44:41 | <PovAddict> | ... dammit |
| 03:44:49 | <PovAddict> | ACTION slaps dc_swig |
| 03:45:14 | <dajobe> | ACTION slaps off topic PovAddict |
| 03:45:46 | <PovAddict> | joins and parts (= the past 8+ hours) are more on-topic |
| 03:46:09 | <dajobe> | it's sunday night in many places, never going to be busy here |
| 03:46:48 | <PovAddict> | A:=urn: |
| 03:46:56 | <dajobe> | it's a regex |
| 03:47:05 | <PovAddict> | ah no URI :P |
| 03:47:17 | <dajobe> | it's (http|....) I forget the exact |
| 03:47:18 | <PovAddict> | A:=http://example.com |
| 03:47:19 | <PovAddict> | B:=http://example.com |
| 07:46:09 | <drewp> | someone around to test my app? it is more polite for real people to make new accts on it as opposed to me making more sock puppets |
| 07:46:15 | <drewp> | it runs on rdf+sesame... :) |
| 08:50:57 | <melvster> | drewp: i can try and test, if you let me know what to do ... |
| 08:51:15 | <drewp> | i already blew another test account on it by accident |
| 08:51:35 | <drewp> | the site is for microdonations to charity, so i just donated another $.25 |
| 08:52:48 | <melvster> | oops |
| 08:53:44 | <melvster> | i hope it's a good cause! :) |
| 08:53:46 | <drewp> | yeah, how are these charity donations ever going to pay off! |
| 08:54:08 | <drewp> | this one was for teaching kids to read |
| 08:54:36 | <drewp> | apparently 'drewp4' supports that too |
| 08:54:53 | <melvster> | :) |
| 08:55:15 | <melvster> | i think charitable donations is the biggest app on facebook |
| 08:55:36 | <drewp> | havent' heard of that. They take real money from users? |
| 08:56:31 | <melvster> | not yet, i think it's called "causes", they get something like 20 million uniques a month iirc, though facebook have just announced their gift platform |
| 08:56:59 | <melvster> | which means apps can access payments from the facebook platform |
| 08:58:11 | <melvster> | On my account i have: Your Impact 0 recruits $0 donated $0 raised |
| 08:59:12 | <melvster> | maybe the do payments as i see: |
| 08:59:20 | <melvster> | Donation History Recurring Donations Payment Information You have made 0 donations. |
| 09:00:23 | <melvster> | s/the/they |
| 09:02:11 | <melvster> | is your system anything like: http://www.justgiving.com/ ? |
| 09:04:00 | <drewp> | mine is tipfight.com, and the difference is that it's got more game/score feel and it's micro (i suggest $0.25 a day) |
| 09:04:50 | <melvster> | wow looks excellent |
| 09:06:46 | <melvster> | i like your featured charities, most people dont realise that suicide kills more people anually than war does |
| 09:07:07 | <drewp> | heh- they're gone in 4 hours (until i rotate those back through) |
| 09:07:59 | <melvster> | shall i try and sign in? |
| 09:08:30 | <drewp> | sure. it should now (once again) work to click donate and then get automatically walked through a twitter oauth workflow |
| 09:08:37 | <drewp> | and a tipjoy one, which you won't even see |
| 09:09:14 | <drewp> | and then you'll have pledged $0.25 on tipjoy. I'll just pay you back on tipjoy, so you won't need a credit card to successfully donate |
| 09:09:28 | <drewp> | ah, did you go for the signin button first? |
| 09:09:47 | <melvster> | i just tried twitter, it signed me in... |
| 09:10:22 | <drewp> | no prob. The donate buttons would have included that step, but this is a fine test as well |
| 09:10:24 | <tobyink> | melvster: and mosquitoes kill more people than war and suicide combined. |
| 09:10:55 | <melvster> | tobyink: that's right tho not the ones that Bill Gates breeds :) |
| 09:11:37 | <melvster> | tipjoy: "We're celebrating the API with a Contest: make something cool with our API and win a MacBook Air!" |
| 09:11:53 | <melvster> | i wonder if you could triplify the tipjoy API and win TWO macbook airs :) |
| 09:12:15 | <drewp> | don't think that hadn't crossed my mind :) |
| 09:12:45 | <drewp> | i did essentially triplify twitter accounts and followers and tipjoy transactions on this project |
| 09:15:34 | <melvster> | very cool, i never worked out how to triplify a transaction, do you wrap it in a transaction triple? |
| 09:17:39 | <drewp> | uh, not that cool. I get a (financial) transaction record from tipjoy, and i stuff it in my graph linking to the two parties involved |
| 09:18:20 | <melvster> | im not sure i completely understand how tipjoy works, you tweet a payment, then what? |
| 09:18:36 | <drewp> | you've just "pledged" some money, essentially |
| 09:19:03 | <drewp> | tipjoy will give your account a negative balance, and the recipient can see that you mean to pay them money, but nothing else happens |
| 09:19:21 | <melvster> | very smart |
| 09:19:27 | <drewp> | if someday you choose to actively put money ($5 min) into your tipjoy accuont, then they can start paying off those and future transactions |
| 09:19:28 | <melvster> | but how do the balances get cleared? |
| 09:20:24 | <melvster> | im building something similar using sem web tech, danbris 'egobank' idea |
| 09:21:18 | <tobyink> | Have found out that "#ttl" is used on twitter as some sort of wine tasting tag. Anyone got an idea for a better tag to use for MicroTurtle? |
| 09:21:50 | <tobyink> | MicroTurtle = http://buzzword.org.uk/2009/microturtle/spec |
| 09:22:06 | <tobyink> | (formerly known as ttldent). |
| 09:22:13 | <drewp> | egobank.me curious! |
| 09:23:01 | <drewp> | i was telling someone who doesnt use computers much about web-of-trust kinds of things, and she said "what happens when someone several links away does something bad?" |
| 09:24:07 | <libby> | ACTION worries about http://www.theregister.co.uk/2009/05/29/hp_labs_closure/ |
| 09:24:38 | <tobyink> | You need a diminishing web of trust. |
| 09:25:46 | <tobyink> | So that anyone you trust directly gets a trust score of 1.0. Indirect trust falls off by 66.7% for each degree of separation. |
| 09:26:26 | <drewp> | tobyink: and then we don't do anything? once i know person4 is bad, i want him trusted at 0.0 |
| 09:26:26 | <melvster> | tobyink: where dd you get the vaue 66.7%? |
| 09:26:38 | <tobyink> | But if you trust Alice and Bob, and they both trust Carol, then you trust Carol 33% + 33%. |
| 09:26:57 | <BenO> | libby, ouch... |
| 09:27:14 | <drewp> | tobyink: -that- sounds odd. if i know carol 6 ways, do i trust her 200%? |
| 09:27:34 | <BenO> | libby, but a bit of a silver lining for me, erm... we are looking for sem web people :) |
| 09:27:57 | <tobyink> | melvster: it's roughly how PGP's WOT works. If you trust 3 people who all trust a fourth person, then you automatically trust the fourth person. IIRC. |
| 09:28:21 | <melvster> | interesting ill have to read up on that |
| 09:28:34 | <melvster> | there's also the grameen system |
| 09:29:09 | <melvster> | you have 5 people in your web, and if one of them doesnt pay (does something bad) you dont get any more money |
| 09:29:18 | <melvster> | i think he got a nobel prize for it |
| 09:29:56 | <melvster> | turns out this system of payments got one of the highest return rates theyve ever seen from the poor |
| 09:33:21 | <tobyink> | libby: yes, cutting R&D is precisely what a company needs to do to weather a recession... </sarcasm> |
| 09:33:53 | <melvster> | its more complex as you go out further nodes away, you probably need a good scoring algorithm, i wonder if the scoring/weights should be inside the semantic web of trust, or stored seperately in an algorithm |
| 09:37:34 | <melvster> | tipfight is pretty cool i have to say, also like the tipjoy idea, i wonder if there is a semantic version ... |
| 09:38:02 | <drewp> | got any suggestions for tomorrow's charities? i still haven't set them up |
| 09:38:17 | <drewp> | they should have twitter accts already (tipjoy not necessary) |
| 09:38:20 | <melvster> | AIDS i'd say |
| 09:38:34 | <melvster> | look at the copenhagen consensus |
| 09:39:36 | <melvster> | see: http://www.copenhagenconsensus.com/Default.aspx?ID=953 |
| 09:40:16 | <melvster> | hmm seems HIV has fallen down the lsit |
| 09:42:02 | <tobyink> | [ a :AutoCalculatedTrustLevel ; :truster <foo> ; :trusted <bar> ; :score 0.16 ; :calculatedFrom [ a :ManualTrustLevel ... ] , [ a :ManualTrustLevel ... ] ] . |
| 09:42:52 | <melvster> | tobyink: very nice |
| 09:43:47 | <tobyink> | It would probably even be possible for CWM to calculate trust levels over a graph of people from a handful of manual trust levels. |
| 09:44:34 | <melvster> | would it need the triples or could it say use sparql against sindice/swse? |
| 09:45:05 | <melvster> | maybe im saying the same thing |
| 09:45:41 | <melvster> | sparql out the triples then run through the engine |
| 09:45:48 | <melvster> | ill have to try that |
| 09:45:52 | <tobyink> | Simple algorithm like: if A trusts B with score X, and B trusts C with score Y, then A autotrusts Cwith score min(max((X*Y - 0.1), 0), 1) |
| 09:46:45 | <tobyink> | Source of input data to cwm would probably be a local turtle/XML file, but that in turn could be exported from elsewhere. |
| 09:50:11 | <tobyink> | Trust levels added to http://vocamp.org/wiki/VoCampBristol2009#Vocabularies.2C_etc._I_want_to_to_work_on |
| 09:52:49 | <melvster> | bing is now live: http://www.bing.com/ |
| 09:54:29 | <melvster> | lol had to use autopager, to get it to have that live.com feel ... |
| 09:57:01 | <drewp> | i saw auto paging on the image search but not the main page |
| 09:59:37 | <melvster> | i use the firefox extension, is great, and also for finding xpaths, would love to see an rdfa version |
| 10:00:04 | <tobyink> | Just had a very quick play, but bing.com does actually seem to provide pretty good results. |
| 10:07:07 | <mischat_> | and it has a picture of santorini ! |
| 12:03:34 | <melvster> | tipjoy api is pretty interesting: |
| 12:03:35 | <mhausenblas> | hehe. /me connected, finally! thanks freenode crew :) |
| 12:03:35 | <phenny> | mhausenblas: 31 May 12:08Z <bengee> tell mhausenblas that the eswc confx system is online now (including signup, code pattern is in your mail) will work on widgets now |
| 12:03:36 | <melvster> | p $2 @charitywater to help build wells for those without clean drinking water |
| 12:04:48 | <melvster> | if @charitywater was a URI it would be a bit more global |
| 12:05:02 | <tobyink> | mhausenblas: "ttldent" is now "MicroTurtle". Also I need to choose a hashtag other than "#ttl" to act as an identifier because #ttl is used by some twitter wine tasting club. Do you have a suggestion of a hashtag to use? |
| 13:23:52 | <mhausenblas_> | maybe #mtu ? |
| 13:24:30 | <mhausenblas_> | tobyink: mtu == *m*icro *tu*rtle |
| 13:24:34 | <mhausenblas_> | btw, important announcement for those who are at ESWC2009: |
| 13:24:54 | <mhausenblas_> | the LOD gathering has been postponed to tomorrow, 2 June 4:30pm |
| 13:25:08 | <mhausenblas_> | we'll meet at the registration desk -please spread the word |
| 13:25:12 | <mhausenblas_> | ACTION back to workshop |
| 13:41:06 | <tobyink> | mhausenblas_: mtu seems sensible. Short; little existing usage. |
| 13:59:39 | <tobyink> | Does anyone know of an existing predicate with domain rdfg:Graph and range rdf:Statement which means "this graph contains this statement"? (The inverse would be fine too.) |
| 14:07:35 | <indifferen7> | hi! do you have any suggestions on how to approach authorization/authentication in linked data applications? |
| 14:09:10 | <mhausenblas_> | http://ouseful.wordpress.com/2009/06/01/plugnplay-public-data/ |
| 14:09:10 | <mhausenblas_> | indifferen7: yes, see: |
| 14:09:14 | <mhausenblas_> | http://esw.w3.org/topic/WriteWebOfData |
| 14:09:49 | <mhausenblas_> | indifferen7: basically it boils down to FOAF+SSL and WAC |
| 14:10:13 | <gromgull> | mhausenblas_: freenode unblocked our IP |
| 14:10:19 | <gromgull> | do you already know? |
| 14:10:39 | <indifferen7> | mhausenblas_: interesting, I'll check it out! what is your opinion on openid/oauth in this context? |
| 14:11:22 | <mhausenblas_> | work on both 'recently' started |
| 14:11:22 | <mhausenblas_> | some implementations are available (FOAF+SSL more, WAC less - check out Joe's Apache module) |
| 14:11:22 | <mhausenblas_> | C:| Plug’n'Play Public Data |
| 14:11:32 | <mhausenblas_> | C: hm. I certainly don't agree with the last paragraph ;) |
| 14:12:03 | <mhausenblas_> | C: 'PS just in case the Linked Data folks feel left out, I still think RDF and semweb geekery is way too confusing for mortals ... ' |
| 14:12:59 | <mhausenblas_> | D:| Concept of a true read-write Web of Data |
| 14:13:40 | <mhausenblas_> | gromgull: I noticed - otherwise we wouldn't be talking to each other ;) |
| 14:14:06 | <gromgull> | mhausenblas_: i saw you use a web irc client earlier - i thought maybe you still were! |
| 14:14:27 | <mhausenblas_> | indifferen7: I guess OpenID is widely supported and OAuth also increasingly |
| 14:14:43 | <mhausenblas_> | so, whatever we do it should also somehow work in a mix setup |
| 14:16:14 | <indifferen7> | mhausenblas_: okay, thanks for the feedback! :) |
| 14:32:23 | <melvster> | indifferent7: openid works with foaf autodiscovery or rdfa |
| 14:32:35 | <melvster> | as long as the id is a URI |
| 14:33:36 | <melvster> | though once you've got an openid there's normally not much linked data associated with it |
| 14:33:49 | <melvster> | however with a foaf there almost always is |
| 14:34:45 | <melvster> | what are you trying to build? |
| 14:41:31 | <indifferen7> | melvster: I am currently creating a linked data content repository, and the content + metadata (rdf) need access control. the problem i see with openid/oauth is that it requires web browser interaction from the user |
| 14:43:37 | <melvster> | hmmm i think web browser is most common |
| 14:43:47 | <melvster> | you have to redirect the user agent from what i recal |
| 14:44:53 | <melvster> | you could simply add a username and password with a cookie/session |
| 14:45:00 | <PovAddict> | ACTION is all for non-html desktop tools interacting with the Web |
| 14:46:23 | <indifferen7> | melvster: yeah, but i'm thinking about when a machine crawls the linked data..if the user is not a person that is |
| 14:47:04 | <indifferen7> | the cookie solution could work as long as it is done RESTfully |
| 14:47:10 | <melvster> | machines can use cookies, but yes you can do the first authentication with SSL |
| 14:47:36 | <indifferen7> | yeah ssl seems more natural |
| 14:47:57 | <melvster> | you can tag a certificate onto an openssl command |
| 14:48:04 | <melvster> | i think also curl |
| 14:49:00 | <indifferen7> | perfect, apps like cURL need to be able to fetch linked data..openid and its relatives are more fitted for social networking sites and similar imo |
| 14:50:59 | <melvster> | what language are you using? |
| 14:51:12 | <indifferen7> | java :) |
| 14:52:44 | <melvster> | henry story has some implentations in java i think |
| 14:54:04 | <tobyink> | FOAF+SSL Java 0.3.1 - http://lists.foaf-project.org/pipermail/foaf-protocols/2009-May/000612.html |
| 14:54:30 | <melvster> | password to sommer is "guest" |
| 14:55:35 | <indifferen7> | alright, thank you for the feedback, now I know where to start looking :) |
| 14:57:39 | <melvster> | i guess it would be nice if the auth part of foaf+ssl could fall back to openid, for those openids that link to a FOAF |
| 15:16:39 | <timbl> | http://www.pwc.com/extweb/home.nsf/docid/1308AF8EA7929CCA852575BA00720F26 |
| 15:17:51 | <mattl> | timbl: does w3c have the ability to host hg/bzr/git for projects? librefm is looking for a new host. |
| 15:18:38 | <timbl> | E:|PriceWaterhouseCoopers: Spring 09 Technology Forecast: Semantic Web |
| 15:19:02 | <timbl> | E: Spinning a data Web Semantic Web technologies could revolutionize enterprise decision making and information sharing. Here’s why |
| 15:19:24 | <timbl> | E:Making Semantic Web connections: Linked Data technology can change the business of enterprise data management |
| 15:20:09 | <timbl> | E: Traversing the Giant Global Graph: Tom Scott of BBC Earth describes how everyone benefits from interoperable data. |
| 15:22:22 | <melvster> | mattl: why new host, i thought you guys used svn, assume you've thought about github, no bug tracker tho, then there's launchpad ... |
| 15:22:57 | <mattl> | melvster: we're moving away from svn... github and launchpad aren't free software. |
| 15:23:14 | <tobyink> | phenny, tell cygri would be nice if http://prefix.cc/foaf.redir did a 302 redirect to http://xmlns.com/foaf/0.1/ and so on for other prefixes. |
| 15:23:14 | <phenny> | tobyink: I'll pass that on when cygri is around. |
| 15:23:37 | <melvster> | true, what free software solutions are possible? |
| 15:23:52 | <mattl> | gitorious, for git... i'm not sure about git. |
| 15:24:05 | <mattl> | for bzr, sftp... |
| 15:25:08 | <melvster> | i started using git recently, i like it a lot |
| 15:26:09 | <melvster> | hmm gitorious looks good, i might try and run a node at some stage |
| 15:27:11 | <tobyink> | timbl: feature request - <http://www.w3.org/2000/10/swap/log#statement> a rdfs:Class ; rdfs:domain log:Formula ; rdfs:range rdf:Statement . |
| 15:27:22 | <tobyink> | s/rdfs:Class/rdf:Property/ |
| 15:27:40 | <mattl> | melvster: git seems a little too much for my liking |
| 15:28:05 | <kennyluck> | timbl, feature request - make cwm and tabulator into community projects. |
| 15:28:50 | <melvster> | mattl: a little too much what? :) |
| 15:29:07 | <mattl> | too much... complication. |
| 15:29:10 | <melvster> | ahh |
| 15:29:20 | <mattl> | i like SVN, it has revision numbers that aren't ugly. |
| 15:29:51 | <tobyink> | I don't see why we can't all just stick with SVN. |
| 15:30:10 | <melvster> | harder to branch and merge |
| 15:30:17 | <Shepard> | tobyink: http://www.w3.org/DesignIssues/Reify.html has rei:RDFGraph |
| 15:30:28 | <mattl> | tobyink: because when the one host with svn on it goes down, it's a bit rubbish |
| 15:31:48 | <Shepard> | although it also has rei:Formula as a super-class of that. maybe that should be owl:equivalentClass log:Formula? |
| 15:36:39 | <melvster> | mattl: the advantage of gitorious is that laconi.ca is already there and is a similar project |
| 15:37:12 | <tobyink> | Shepard: thanks for that, but I think that's a bit heavyweight for me. I don't need to literalify URLs and stuff - I just need it as a mechanism for flattening a quad store into a triple store. |
| 15:37:13 | <mattl> | yup. |
| 15:37:28 | <tobyink> | BTW, Shepard, happy birthday! ;-) |
| 15:38:02 | <Shepard> | thanks ;) |
| 15:39:29 | <Shepard> | I think nonetheless it'd be worthy to equal those two Formula classes. as well as rei:RDFGraph and rdfg:Graph |
| 15:40:31 | <Shepard> | the example in the section RDF Graphs in that design issue is a bit broken. and the I'm not sure why rei:statements has a range of rdfs:Class... |
| 15:42:07 | <Shepard> | tobyink: just define the property on openvocab? |
| 15:42:43 | <tobyink> | That may be what I end up doing, but it would be nicer if it lived alongside the rest of log, as it seems generally useful. |
| 15:45:19 | <melvster> | both github and gitorious ask you to upload a public key before you can use it, now a foaf+ssl version would be able to get your public key on a page hit and upload it auotmatically |
| 15:47:19 | <melvster> | maybe its time to build a foaf+ssl+git system |
| 15:58:46 | <tobyink> | OK, my ttldent/MicroTurtle implementation is now fully quadded up! |
| 15:59:47 | <tobyink> | http://buzzword.org.uk/2009/microturtle/implementation.cgi?acct=tobyink&format=application/rdf+xml is essentially just my identi.ca feed, but with MicroTurtle triples added in reified form. |
| 16:00:44 | <tobyink> | Whereas http://buzzword.org.uk/2009/microturtle/implementation.cgi?acct=tobyink&format=application/rdf+xml¬ice=4778008 shows the non-reified graph of one of my notices. |
| 16:24:12 | <tobyink> | http://purl.org/NET/uri# |
| 16:24:27 | <tobyink> | F:|URI Ontology |
| 16:24:49 | <tobyink> | F: Updated yesterday. Allows you to describe multiple URIs for the same thing. |
| 16:25:04 | <tobyink> | F: Example usage at the bottom. |
| 17:16:00 | <mhausenblas> | ACTION waves to other ESWC09 participants. finally back from LarKC reception and at the hotel (with a proper network ;) |
| 17:20:40 | <mhausenblas> | tobyink++ for the work on the URI onto - will propose to (re)use in the AWWSW Task Force work |
| 17:21:17 | <bengee> | ACTION waves |
| 17:27:16 | <mhausenblas_> | ACTION waves to bengee ;) |
| 17:32:16 | <mhausenblas_> | http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData/HeraklionGathering |
| 17:32:56 | <mhausenblas_> | finally fixed the date and location: we'll meet up on Wednesday 3 June at the Aldemar Knossos Royal Village from 4:30pm to 6pm at the Conference Hall 1 (the big room) |
| 17:33:03 | <mhausenblas_> | please spread the word |
| 18:34:50 | <mhausenblas> | oshani? |
| 18:35:05 | <oshani> | hello mhausenblas |
| 18:35:12 | <oshani> | how is ESWC? |
| 18:35:50 | <mhausenblas> | oh there are worse places I can imagine :) |
| 18:36:22 | <mhausenblas> | some 20 degrees, sitting outside, beer, nice music, a network |
| 18:36:34 | <oshani> | sounds like fun |
| 18:36:43 | <mhausenblas> | yeah ;) |
| 18:37:23 | <mhausenblas> | still, the RDForms <-> SPARQL Update stuff causes me headache |
| 18:37:56 | <mhausenblas> | the thing is, the way RDForms are currently modelled ... they convey the entire operation semantics already |
| 18:38:22 | <mhausenblas> | so SPARQL Update (using just INSERT ....) is sort of not clean |
| 18:38:43 | <oshani> | not clean... as in? |
| 18:38:44 | <mhausenblas> | same as RPC/SOAP sort of misusing HTTP, you know |
| 18:39:09 | <mhausenblas> | if we go for that it will work, for sure but its not precisely a good design |
| 18:39:15 | <mhausenblas> | two options I see: |
| 18:39:41 | <oshani> | as a side note: I noticed in the pushback method of the demo, that even if you post the jsonoutgraph I'd still get a result from the server |
| 18:39:50 | <mhausenblas> | either the client (RDForms processor) turns the diff graph into 'proper' SPARQL Update |
| 18:40:03 | <oshani> | so, I wonder whether there's a very high reliance between the RDForm and SPARUL |
| 18:40:04 | <mhausenblas> | yeah. buggy, isn't it |
| 18:40:17 | <mhausenblas> | well, that's the point |
| 18:40:43 | <mhausenblas> | so, either the client does something or we redesign the RDForm vocabulary |
| 18:41:07 | <mhausenblas> | it boils down to a triple-centric view vs. form-based/key value view of the data |
| 18:41:37 | <mhausenblas> | but in the current setup, a HTTP POST would actually be appropriate (no SPARQL Update at all) |
| 18:42:08 | <oshani> | ACTION looking at the diagram |
| 18:42:25 | <mhausenblas> | however, as timbl and I agreed at WWW, SPARQL should be the interface, I guess we'd need to change the RDForms bits |
| 18:43:04 | <timbl> | E: From folksonomies to ontologies |
| 18:43:04 | <timbl> | Uche Ogbuji of Zepheira discusses how early adopters are introducing Semantic Web to the enterprise. |
| 18:43:16 | <timbl> | E: Lynn Voge How the Semantic Web might improve cancer treatment |
| 18:43:17 | <timbl> | M. D. Anderson’s Lynn Vogel explores new techniques for combining clinical and research data. |
| 18:43:27 | <timbl> | E: Semantic technologies at the ecosystem level |
| 18:43:27 | <timbl> | Frank Chum of Chevron talks about the need for shared ontologies in the oil and gas industry. |
| 18:43:53 | <oshani> | mhausenblas, you said "either the client (RDForms processor) turns the diff graph into 'proper' SPARQL Update" - I think that's a good option |
| 18:44:00 | <timbl> | ooops |
| 18:44:20 | <timbl> | E5: Semantic technologies at the ecosystem level : Frank Chum of Chevron talks about the need for shared ontologies in the oil and gas industry. |
| 18:45:52 | <timbl> | I think there is a lot to be said for making application run on a generic ACL-controlled writable storage, with SPARL/Update and HTTP |
| 18:47:49 | <MacTed> | ACTION wonders if E4 and E6 might also want fixing from timbl? |
| 18:47:56 | <MacTed> | D: |
| 18:48:00 | <MacTed> | er, |
| 18:48:01 | <MacTed> | E: |
| 18:48:45 | <mhausenblas_> | oshani: sorry, network is shaky here as well ;) |
| 18:48:47 | <mhausenblas_> | yes, oshani - sounds like an option |
| 18:49:27 | <mhausenblas_> | timbl: ahm. which URI? |
| 18:49:50 | <mhausenblas_> | ah, sorry, saw it |
| 18:49:59 | <oshani> | so, these RDForms diffs would be the triples we are are inserting thru SPARUL? is this the protocol now? |
| 18:50:29 | <mhausenblas_> | well, that's were I'm not entirely sure you see |
| 18:50:59 | <mhausenblas_> | the idea behind RDForms is, as you know, to enable a single parser ... |
| 18:51:18 | <mhausenblas_> | however, that conflicts with the SPARQL stuff |
| 18:51:44 | <oshani> | not sure, I am seeing the conflict here |
| 18:51:50 | <mhausenblas_> | hm |
| 18:52:08 | <mhausenblas_> | well, we defined SPARQL INSERT now |
| 18:52:10 | <mhausenblas_> | right? |
| 18:52:16 | <oshani> | ok |
| 18:52:17 | <mhausenblas_> | for *all* operations |
| 18:52:33 | <mhausenblas_> | say, you want to delete a certain field |
| 18:53:02 | <timbl> | A one-off action form should basically be able to say what sort of bit of graph is inserted into which document. And editing form should have a (potential) statement in a certain document corresponding to each form field. The latter siutauin is what happens when editing in tthe outline mode in tabr. |
| 18:54:49 | <melvster1> | i wonder if the form is in "pure" rdfa can you detect and send the deltas as SPARUL |
| 18:54:59 | <mhausenblas_> | you'd do an INSERT ... and the operation semantics is encoded in the pb:operation |
| 18:55:36 | <mhausenblas_> | hm. not sure if I understand timbl |
| 18:55:42 | <mhausenblas_> | can you give a concrete example |
| 18:56:36 | <mhausenblas_> | ACTION very happy changing the RDForms voc as I very much believe in SPARQL Update being the standard interface for all updates |
| 18:57:00 | <mhausenblas_> | melvster1: yeah, that's not so hard |
| 18:57:41 | <mhausenblas_> | still I don't wanna get into this nasty situation misusing basic protocols such as HTTP and SPARQL |
| 18:59:27 | <timbl> | The form should have in it the knowledge "when press submit, then the fields will be filled in in { [ a Person; name ?n; email ?e] is foraf:knows of $me } in $me.foafProfile " or something |
| 18:59:52 | <mhausenblas_> | timbl: I see. thanks! |
| 19:00:14 | <mhausenblas_> | ACTION will have to leave soon, forgot charging unit. will sum up issue I see and send over a new proposal to you ... |
| 19:00:15 | <oshani> | timbl, mhausenblas_ 's RDForm allows to have any form structure (whereas Tab's outline is just one structure). You can then query the values on this form using RDForm semantics. So, as you say I think linking part of a document to a field is not a problem |
| 19:00:15 | <timbl> | So that when the form is filled the actual foaf file server just serves. |
| 19:00:32 | <melvster1> | the way pushback wraps the triples is pretty smart, what i imagined was an rdfa page, with "edit in place" js to turn it into an rdfa form, so lets say a span become maybe an <input>, then when that is changed some more js, works out what has changed and generates some SPARUL, this I imagine is an rdfa page similar in function to tabulator, but that might be quite tricky ... |
| 19:02:34 | <timbl> | When you are editing something like "The content of this field corresponds to ?x in { <..#jill> foaf:name ?x. } om <http://.... faf.rdf>. |
| 19:02:58 | <mhausenblas_> | ok |
| 19:03:07 | <timbl> | So every time the form is changed by the user the given doc is changed using SPARUL immediately. |
| 19:03:16 | <mhausenblas_> | right |
| 19:03:57 | <timbl> | The small granularity makes conflices with other users much less likely, and also reduced the chance of loisig a lot of data if the browser crashes/exits/is forgotten etc |
| 19:04:08 | <mhausenblas_> | indeed |
| 19:05:39 | <mhausenblas_> | ACTION will ping oana again - the google pushback should really be available soon ... |
| 19:05:58 | <oshani> | looking forward to it |
| 19:06:06 | <mhausenblas_> | +1 |
| 19:06:06 | <mhausenblas_> | :) |
| 19:07:10 | <mhausenblas_> | problem is: I sit in Greece, she in Ireland - tomorrow morning ;) |
| 19:08:24 | <mhausenblas_> | anyway. once that is fixed, we'd need a proper evaluation as well |
| 19:09:34 | <mhausenblas_> | met a great Scotsman here (Norman, http://nxg.me.uk/) |
| 19:10:32 | <mhausenblas_> | he works on something in this realm as well - had a good discussion with him today |
| 19:11:31 | <melvster1> | tell him to update his foaf:depiction! :) <depiction rdf:resource="jpg"/> |
| 19:11:45 | <mhausenblas_> | ok, seems I have to shut down. thanks oshani and thanks timbl |
| 19:11:52 | <mhausenblas_> | melvster1: will do ;) |
| 19:12:41 | <mhausenblas_> | ACTION wonders if Norman maybe intentionally did this ... sort of early social semantic obfuscation method :D |
| 19:13:26 | <melvster1> | his foaf should appear here if he fixes it: http://foaf.me/?webid=http://nxg.me.uk/norman/rdf |
| 19:14:52 | <melvster1> | ill try and see if i can get an edit in place with sparul working |
| 19:16:45 | <mhausenblas_> | cool |
| 19:21:07 | <mhausenblas_> | good nite Web of data |
| 20:43:35 | <MacTed> | E4: From folksonomies to ontologies: Uche Ogbuji of Zepheira discusses how early adopters are introducing Semantic Web to the enterprise |
| 20:43:35 | <MacTed> | E5: How the Semantic Web might improve cancer treatment: M. D. Anderson’s Lynn Vogel explores new techniques for combining clinical and research data |
| 20:43:35 | <MacTed> | E6: Semantic technologies at the ecosystem level: Frank Chum of Chevron talks about the need for shared ontologies in the oil and gas industry |
| 22:27:46 | <mischat> | http://www.youtube.com/watch?v=v_UyVmITiYQ&eurl=http%3A%2F%2Fwww.facebook.com%2Fhome.php%3Fref%3Dlogo&feature=player_embedded |
| 22:28:02 | <mischat> | H:| Google Wave Preview 2009 |
| 22:36:50 | <melvster1> | mischat: is a nice demo, i think it works over XMPP or are they using some kind of websocket ... i signed up for a sandbox account , didnt hear back tho ... |
| 22:37:31 | <PovAddict> | what *is* it? |
| 22:37:43 | <mischat> | yeah it looks interesting, and yes, I think it is based on xmpp, i will look into it after i finishing watching |
| 22:38:08 | <mischat> | ACTION wants a sandbox account, and it will be open source |
| 22:38:11 | <mischat> | :) |
| 22:39:42 | <melvster1> | they said the 'lions share' will be open source, whatever that means |
| 22:40:36 | <mischat> | well they said bits of it requires google gears, am not sure if that is open source |
| 22:41:11 | <mischat> | but i agreed with the basic idea, email is pretty rubbish |
| 22:41:13 | <melvster1> | did they open it, im not sure ... i think they are basing things on google gadgets a la open social |
| 22:41:51 | <melvster1> | hmm google gears doesnt work with opera |
| 22:42:50 | <melvster1> | id like to run a federated node and see how it interacts |
| 22:43:43 | <dajobe> | I captured a bunch of pointers & analysis about it on my delicious count |
| 22:44:08 | <dajobe> | at http://delicious.com/dajobe/wave |
| 22:45:09 | <melvster1> | dajobe: thanks |
| 22:45:10 | <mischat> | great stuff |
| 22:45:28 | <dajobe> | also try ##wave |
| 22:46:06 | <melvster1> | i did read that it was a bit on the slow side |
| 22:46:36 | <melvster1> | would love to make a semantic version |
| 22:48:41 | <melvster1> | from wiki: Gears, formerly Google Gears,[1] is software offered by Google that "enables more powerful web applications, by adding new features to your web browser."[2] Released under the BSD license,[3] Gears is free and open source software. |
| 22:51:07 | <melvster1> | i have a feeling google gadgets framework may play a heavy role |
| 22:53:11 | <melvster1> | Gadgets: Build gadget extensions to provide a new way for users to interact. Gadgets provide a means to run small programs in waves. For more information, see the Wave gadgets Tutorial. |
| 22:54:11 | <melvster1> | it seems like wave gadgets are like open social gadgets but with a few more features |
| 22:56:09 | <PovAddict> | (hypothetical) okay so how do I go about replacing the "private message" feature on my website with waves? :) |
| 22:57:04 | <PovAddict> | http://wave.google.com/help/wave/images/ss1.gif looks "heavy" and definitely not small-screen-friendly, heh |
| 22:57:30 | <dajobe> | the video has a 'curses' terminal client |
| 22:57:43 | <dajobe> | don't confuse the client from the protocol/server |
| 22:58:05 | <PovAddict> | yeah I know |
| 22:58:09 | <PovAddict> | well, good to know the system isn't tied to HTML |
| 22:58:37 | <PovAddict> | what about 3rd party "gadgets"? those will remain tied to the browser I guess? |
| 22:59:30 | <melvster1> | why tied to the browser? |
| 22:59:56 | <melvster1> | surely google desktop |
| 23:00:02 | <PovAddict> | well, don't you write them in AJAX? |
| 23:00:06 | <dajobe> | oh, it's tied to html + js + open social + ... |
| 23:01:17 | <dajobe> | have a look at this http://intertwingly.net/blog/2009/06/01/Wave-Robot-Ruby-Client |
| 23:01:17 | <PovAddict> | I mean, can 3rd party functionality be used from a curses client? :P |
| 23:01:35 | <dajobe> | if you can run the js engine & html engine to render into text |
| 23:01:56 | <PovAddict> | nope, if it's tied to html it's not what i want |
| 23:01:58 | <dajobe> | ACTION mtgs & |
| 23:03:23 | <melvster1> | i guess you could build a wrapper around the gadget framework |
| 23:03:35 | <melvster1> | but quite a lot of effort |
| 23:04:22 | <melvster1> | mind you lots of people are writing open social containers |
| 23:04:48 | <melvster1> | so wave containers will probably arrive at some stage |
| 23:07:04 | <melvster1> | http://www.w3.org/2005/Incubator/socialweb/wiki/TopSocialNetworkingSites |
| 23:07:17 | <melvster1> | A: list of top social networking sites |
| 23:24:47 | <melvster1> | the part of wave i fine most interesting: |
| 23:24:51 | <melvster1> | Google Wave Federation Protocol is extension to the XMPP protocol. It's designed for near real-time communication between the computer supported cooperative work wave servers. |
| 23:24:51 | <melvster1> | Still currently in development, the Wave Federation Protocol as an open protocol is intended to parallel the openness and ease of adoption of the email protocol so waves may succeed email, as the dominant form of Internet communication. |
| 23:26:27 | <melvster1> | s/fine/find |
| 23:28:46 | <melvster1> | so the identifiers are all email addresses |
| 23:30:19 | <melvster1> | and they are going to use hg: hg clone http://wave-protocol.googlecode.com/hg/ wave-protocol |
| 23:31:37 | <PovAddict> | <3 |
| 23:34:06 | <melvster1> | ah nice they're using SSL X.509 |
| 23:35:57 | <melvster1> | hmm they want to have root CA's |
| 23:37:04 | <melvster1> | "we suggest that each server sign a key with an X.509 certificate signed by one of an agreed-upon set of parties, much as those used for SSL in browsers" |
| 23:59:12 | <mischat> | ACTION chuckles at office space reference in the wave preview ... |
Back to channel and daily index: content-negotiated html turtle
