Experimental IRC log swig-2009-04-22

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.

00:00:00<snail>PovAddict: i can. it has been. but you're right, it shouldn't have been
00:00:17<snail>PovAddict: is there a particular problem you're tring to solve?
00:00:29<PovAddict>nope, just wondering... every single IDE seems to have its own
00:00:35<snail>s/tring/trying/
00:04:00<snail>PovAddict: you may be interested in DOAP, and specifically the apache version http://projects.apache.org/doap.html
00:04:21<snail>but that's not really an IDE
00:07:24<PovAddict>it would be great if there was a single file format that could let you:
00:07:28<PovAddict>know information about the project
00:07:34<PovAddict>download the project source code
00:07:48<PovAddict>load the project source code into an IDE
00:07:53<mattl>FILE_ID.DIZ
00:07:54<mattl>;)
00:08:00<PovAddict>download dependencies
00:08:05<PovAddict>and maybe even build it
00:08:53<snail>PovAddict: all but the IDE one are provided by the debian packaging format
00:10:03<mattl>and people don't really use IDEs to develop free software, with the exception of Java, maybe.
00:10:38<PovAddict>http://blogs.sun.com/bblfish/entry/replacing_ant_with_rdf
00:12:02<PovAddict>snail: can the debian packaging format build Windows software with Visual Studio? :)
00:12:07<snail>PovAddict: DOAP (mentioned in that article) is surprisingly widely used
00:14:05<snail>PovAddict: sure, check out wine
00:20:00<Anchakor>snail: Im sure he meant under windows... (cygwin? :))
00:23:52<MFen>what should data on the web look like
00:24:34<MFen>what i mean is, what are most people expecting in terms of an API to access this data? random piles of REST? rdf/xml dumps of your database? some application-specific api?
00:25:02<MFen>well take as given that rdf/xml or n3 or something is a standard that everyone wants to use.. is there any concept of a standard for getting to it?
00:38:03<Anchakor>MFen: sparql endpoints
00:40:07<drobilla>sparql?
00:40:32<snail>Anchakor: debian ships with windows (X11 etc)
00:41:06<snail>drobilla: http://www.w3.org/TR/rdf-sparql-query/
00:42:05<snail>MFen: Anchakor was right when they said sparq, it's pretty much the best way imaginable to allow access to data
00:43:31<snail>hmmm, confusingly, the sparql standard doesn't contain the expression 'endpoint'
00:44:22<snail>but a sparql endpoint is the web address of a sparql interface to an RDF database
00:44:38<Anchakor>MFen: for discovery of data (sparql endpoints) voiD is (should be) used: http://rdfs.org/ns/void-guide
00:48:49<Anchakor>http://www.lirmm.fr/~coletta/CaisePresentations/TutorialYAHOO.pdf
00:49:18<Anchakor>B:nice link I just found on crawling the web
00:49:59<Anchakor>B:comparison between distributed/centralized etc.
00:50:46<drobilla>snail: I know ;)
00:55:38<Anchakor>snail: debian ships with ms windows? news to me ;)
01:00:19<MFen>Anchakor: thanks, looking into this
01:00:56<MFen>i have actually encountered the term sparql endpoints before, but didn't know if i should care
01:02:21<snail>MFen: it's a bit like an JDBC connection in the relational world, except that it's connecting to a semantic database with a semantic query language
01:03:15<MFen>snail: does it address the wire protocol at all?
01:04:15<snail>MFen: I'm not sure what you mean by that, but use the accepts header to say what format you want the result set in
01:04:40<snail>and the normal http headers to say things like character set, etc, etc, etc
01:05:00<MFen>snail: i meant, does it specify http, and it sounds like it does :)
01:08:10<snail>http://rdfs.org/ns/void-guide is really interesting (and new to me)
01:08:19<snail>oops
01:08:49<snail>it seems like it much more likely to fly if someone wrote a stats-generator to generate all those stats though
01:09:55<MFen>ok, void already looks pretty awesome
01:10:01<MFen>i think this is exactly what i needed
01:10:42<MFen>*love* void:exampleResource. way more stuff should have that
01:16:03<PovAddict><snail> Anchakor: debian ships with windows (X11 etc) <- he meant Microsoft Windows...
01:16:05<PovAddict>and I did too...
01:16:28<PovAddict>can you use a .deb to compile software with Visual Studio? wouldn't it depend on makefiles?
01:16:46<mattl>please please no.
01:16:59<PovAddict>also, debs identify dependencies with names, not with URIs to other debs like DOAP would do :D
01:25:40<snail>PovAddict: your statement implies that URIs are not names, which is false
01:27:45<PovAddict>'libgtk-2.0' isn't globally-unique enough
01:28:25<MFen>it's not necessarily a good thing for that name to be globally unique
01:29:10<PovAddict>I can't grab a random .deb off the internet and install it with its dependencies
01:29:25<mattl>actually, you probably could.
01:29:34<mattl>assuming you're running Debian.
01:30:28<PovAddict>I grab a .deb from distroX and install it, but the dependency package names match the package names on distroY, and rely on peculiarities of how distroY built the programs?
01:30:56<snail> PovAddict: yeah, you can with debian, ubuntu, knoppix, etc, etc
01:31:10<snail>there might be some that are borken, but in general you can
01:31:11<mattl>no, ubuntu changes the names of things for some reason
01:31:39<PovAddict>are their packages compatible? I don't mean the dpkg format, I mean the *contents*... they have to be built with compatible C++ ABIs for example
01:31:55<snail>on my ubuntu i install random debs regularly, or is it the other way round it doesn't work
01:31:58<MFen>usually. if they're not, the dependency tree will usually exclude.
01:32:21<Anchakor>MFen: everything of inportance deserves an URI ( http://www.w3.org/TR/webarch/#identification )
01:32:22<snail>PovAddict: yes, the contents of the packages are compatible
01:32:53<PovAddict>http://apt.saurik.com/debs/ <- there are probably package names there that match Debian package names and yet they'd be totally incompatible (if Debian supported darwin-arm at all :P)
01:33:01<snail>mattl: i suspect that maybe ubuntu uses a slightly different linking fomat
01:33:40<MFen>Anchakor: well, you could make the argument that those packages deserve URIs, all i'm saying is i can't say for sure whether the debian package architecture would still work if package names as used by apt became URIs
01:34:31<PovAddict>how do I "recursively" build a program and its dependencies from source?
01:34:35<MFen>PovAddict: the architecture is part of the metadata. lots of debs have "all" as the arch, explicitly so that you can use them on architectures for which they were not specifically built
01:35:01<PovAddict>I thought the automatic way installed binary dependencies and then built the program alone from source
01:35:04<snail>PovAddict: apt handles dependencies
01:35:14<MFen>PovAddict: sudo apt-get build-dep package is how you get its dependencies, i don't know of any tool that lets you build those all the way down to the bottom of the turtles
01:35:43<PovAddict>I heard "bootstrapping" Ubuntu repository is nearly impossible
01:35:59<snail>PovAddict: there are tools that let you do that (mainly for testing purposes) ask in a debian channel
01:36:29<Anchakor>MFen: my guess it wouldn't, uris contain some strange characters
01:37:18<MFen>Anchakor: nothing to do with syntax.. there are features of the debian system that are enabled by names being non-globally unique
01:38:33<MFen>systems like fink are based on apt and do rebuild the whole system
01:38:50<MFen>and that's practically a one-click installer on osx
01:55:44<MFen>bbl to discuss void and sparql endpoints.. gotta play D&D now
06:42:40<mhausenblas>good morning Web of Data
07:06:49<mhausenblas>WWW09 official opening. on the stage (among others): TimBL and Vinton Cerf
07:09:23<mhausenblas>hei kwijibo
07:09:41<kwijibo>hey mhausenblas , how was lwod ?
07:09:41<phenny>kwijibo: 21 Apr 09:05Z <danja__> tell kwijibo http://semwebdev.keithalexander.co.uk/blog/feed.atom = 406 Not Acceptable ..?
07:10:11<mhausenblas>cool, kwijibo
07:10:12<kwijibo>ta phenny
07:10:14<mhausenblas>all well
07:10:32<kwijibo>wish i could have seen you presentation
07:10:38<kwijibo>did anyone video ?
07:10:58<mhausenblas>no, I don't think so, but Jun told me that she'd like it ;)
07:13:04<mhausenblas>the WWW09 panel seems actually nice: Mike Shaver from Mozilla and Dale Dougherty from O'Reilly
07:28:12<melvster>hi all, question about SPARUL, it works via HTTP POST, i thought POST is normally name value pairs, but in this case it's just free text, e.g.: INSERT { <http://foaf.me/test> <http://xmlns.com/foaf/0.1/interest> "remote testing" . } is that a normal use of POST?
07:28:30<melvster>this is the datawiki impl.
07:30:21<danja>http://twitterwall.me/www2009
07:30:46<danja>E:| Twitterwall for #www2009
07:30:55<danja>E: (good scribing guys!)
07:35:58<kwijibo>melvster: i think so; you can also send regular SPARQL queries with POST according to the the sparql protocol
07:37:18<dajobe>yay, another TUrtle keynote reference
07:38:41<melvster>kwijibo: ah ok thanks, looks like an underused technique, i think it's called POST RAW in some texts
07:39:27<kwijibo>we use it in the talis platform api for uploading data to the store
07:39:55<melvster>talis is java?
07:44:29<melvster>this is quite a useful use of http_post i think im going to try and use it more often ...
07:49:04<drewp>ACTION reads http://n2.talis.com/wiki/Changesets and remembers the sesame/allegro/etc transaction xml format (http://www.franz.com/agraph/support/documentation/current/http-protocol.html#header3-67)
08:02:52<mhausenblas>panel discussion over, now questions from the audience
08:03:30<mhausenblas>JimH asked about how we can increase the outreach, enable more people to use the Web (currently like 20-30% of the world population)
08:03:45<mhausenblas>mobile access might be the preferred route
08:07:10<melvster>only 5% of africa use the web, intel have a project to get 50% online by 2015
08:08:59<melvster>sorry that's AMD: http://50x15.amd.com/
08:11:21<melvster>current penetration: 23.56% Pop: 6,748 Bn ; Internet Users: 1,589 Bn
08:12:45<melvster>stats: http://50x15.com/en-us/internet_usage.aspx
08:26:51<mhausenblas>right, melvster
08:27:10<mhausenblas>couldn't remember the precise figure anymore but that was is
08:27:31<mhausenblas>now her royal highness, the princess is comming
08:27:40<mhausenblas>a short ceremony I hope :)
08:27:57<kwijibo>princess ?
08:28:11<mhausenblas>lemme check the precise title ;)
08:28:21<dajobe>the heir to the spanish throne
08:28:42<kwijibo>really? weird
08:29:03<mhausenblas>'Inauguration ceremony: The Princes of Asturias. Access forbidden between 10:15 and 10:30. '
08:29:10<mhausenblas>from http://dragoman.org/2009program.html
08:29:20<mhausenblas>they would lock us down here in the room
08:29:38<mhausenblas>feels like 20000 photographers around ;)
08:30:27<mhausenblas>ACTION notes that Access forbidden is not 403, in this sense ;)
08:30:37<kwijibo>lol - I did wonder ;)
08:36:04<mhausenblas>wow, should brush up my Spanish
08:36:29<mhausenblas>I wonder if they would be so nice to switch over to English
08:36:43<mhausenblas>now, this is cool
08:37:00<mhausenblas>the one guy who talks now (in Spanish)
08:37:19<mhausenblas>and we get a nice slide projected on the wall and are supposed to read this?
08:37:53<mhausenblas>I wonder if the prince and the princess at least speak English
08:40:30<mhausenblas>I can't believe it - he *is* able to speak English! ;)
08:42:51<mhausenblas>so, the next speaker did it the other way round. said two sentences in English and the rest of the talk in Spanish
08:43:01<mhausenblas>... and we are back to reading slides :)
09:03:53<mhausenblas>now we talk!
09:03:58<mhausenblas>TimBL's speak
09:04:28<mhausenblas>reflections at 20y on the state of the WWW
09:04:48<mhausenblas>he focuses on Web Applications and Open Social Networks
09:05:41<mhausenblas>and of course: ..... Open Linked Data! :)
09:08:09<mhausenblas>we learn that Web year to normal year ratio is 2.6 (by TimBL original definition)
09:08:18<mhausenblas>now going to 1:1?
09:09:40<mhausenblas>ouch ... battery going down .. cya later
09:10:47<dajobe>phising sites
09:10:57<dajobe>see twitter search for #www2009 for lots more
09:11:22<dajobe>look out for things that change slowly that head to a tipping point
09:11:30<dajobe>some things are the same...
09:11:35<dajobe>een though ther eis a lto of fancy stuff on the web
09:11:43<dajobe>there is alot of importnat static stuff
09:11:48<dajobe>the record, the sscientific record
09:11:51<dajobe>the human record
09:11:54<dajobe>is being archived
09:12:07<dajobe>e.g. the whitehouse changes and immediately all the web pages change
09:12:13<dajobe>is somebody looking after them? i hope so
09:12:28<dajobe>Need the web to be a more collaborative
09:12:30<dajobe>Twitter!
09:12:34<dajobe>need more intuituive ways
09:12:45<dajobe>Same tension for parts to be come non-standard
09:12:49<dajobe>become
09:12:51<dajobe>compete
09:13:00<dajobe>have to keep one web - use standards
09:13:10<dajobe>#1. Web Applications
09:13:33<dajobe>when we go to some pages now, we're really running a program in a web page
09:13:37<dajobe>a new computing environment
09:13:48<dajobe>the devel of this environment is struggling with dealing with trust
09:14:04<dajobe>pulling data into a mashup to produce some result
09:14:11<dajobe>using lots of different bits of program
09:14:22<dajobe>how do you trust that code to read my private data?
09:14:26<dajobe>(Javascript code)
09:14:41<dajobe>need more research in capability-based systems like Cja
09:14:44<dajobe>Caja
09:14:51<dajobe>Need decentralized modular installation
09:14:58<dajobe>"Webized debian?" says the slide
09:15:08<dajobe>apt-get install
09:15:54<dajobe>Privacy - huge amount of discussion
09:16:23<dajobe>how to manage geo location on behalf of the user
09:16:27<dajobe>just some info
09:16:41<dajobe>machines will know so much about people
09:16:49<dajobe>it's just the start of a project to deal with this
09:16:55<dajobe>--- Modularity In General
09:17:09<dajobe>simplicity
09:17:16<dajobe>extensibility points to replace parts
09:17:25<dajobe>bolting on parts that haven't been invented yet
09:17:47<dajobe>rather, designing parts for a system that hasn't been invented
09:18:16<dajobe>html must still be modular
09:18:21<dajobe>and connect to the next big thing
09:18:24<dajobe>SVG
09:18:41<dajobe>IE really should impl SVG
09:19:19<dajobe>this is slide 7 of 19, he's not going to finish by 11:30am
09:19:46<dajobe>extensibility by java implementation
09:19:58<dajobe>migrate from in page, installed, pre/compiled
09:20:06<dajobe>this doesn't work for everything of course
09:20:31<dajobe>#2 Social Networking
09:20:48<dajobe>state of them now is walled gardens all over again
09:21:05<dajobe>then: AOL, Prodigy now: Facebook, ...
09:21:27<dajobe>a walled garden is tempting... can take money
09:21:39<dajobe>the first time we had phones that said they had web browsers
09:21:54<dajobe>you could only access web sites where the owner had an agreement/paid the phone co
09:22:03<dajobe>only took off when they were truly open
09:23:39<dajobe>walled garden picture
09:23:53<dajobe>and diagram of walled gardens of twitter / identi.ca and tweet.ie
09:24:53<dajobe>just as we can email anyone, this stuff needs to be opened up
09:25:06<dajobe>the bus above the social networks is rdf + owl _ HTTP _ SPARQL
09:25:12<dajobe>and above that an ajax 3.0 app
09:25:35<dajobe>arguing for Open Social Networking
09:25:48<dajobe>"using flash won't help"
09:26:08<dajobe>--- Identity - Personas
09:26:37<dajobe>I can give you an HTTP URI for me
09:26:46<dajobe>don't need to see "I am X on twitter, I am Y on facebook"
09:26:50<dajobe>that's so pre-web
09:26:55<dajobe>openID qutie deployed
09:27:00<dajobe>foaf+ssl has asdvantages
09:27:03<dajobe>maybe needs a new name!
09:27:15<dajobe>simpler, working towards a secure ID
09:27:29<dajobe>so you can be sure I am the person auth
09:27:41<dajobe>like openID but doesn't require such a complex system with PIPs
09:28:30<dajobe>computers can figure out identities of anonymous people
09:28:35<dajobe>pseudonyms not so real
09:28:52<dajobe>[and if computers can, so can the law/police/govt. !]
09:28:59<dajobe>#3 Linked Open Data Movement
09:29:23<dajobe>the first time I showed this pic, the bubbles were big enough you could read the labels
09:30:38<dajobe>wikipedia and links
09:30:40<dajobe>checking
09:30:44<mhausenblas>now TimBL is speaking about LOD cloud
09:30:54<dajobe>mhausenblas: read the logs
09:31:06<dajobe>Linked Data everywhere...
09:31:52<mhausenblas>ACTION is sorry - just came up online again, didn't check first that you are scribing. thanks!
09:31:56<melvster_>to follow on twitter: http://search.twitter.com/search?q=%23www2009
09:32:02<dajobe>science data
09:32:13<dajobe>sitting on disks, not available by other scientists
09:32:18<dajobe>so linked open data science is very imp.
09:32:55<dajobe>social networks, community data - open street map
09:33:14<dajobe>linkedgeodata.org
09:33:24<dajobe>--- Linked Data Challenges
09:33:41<dajobe>still need a really easy program for somebody with a legacy relational DB
09:34:20<dajobe>something conversational to get out the hiddent db table/columns semantics
09:34:24<dajobe>get the data out
09:34:26<mhausenblas>yes, I agree very much reRDB2RDF - hope we will be able to deliver the base for this in the upcoming W3C Working GRoup
09:34:42<melvster_>i wonder if the phpmyadmin interface could export to N3 it already does a variety of other formats
09:34:53<mhausenblas>nice idea melvster
09:34:55<dajobe>when there are very large datasets... federated or delegated query needed
09:35:08<dajobe>co-reference - which are the same people?
09:35:39<dajobe>--- Social concerns
09:36:11<dajobe>we have some controls preventing access except for friends
09:36:22<dajobe>but you sometimes want access in different roles
09:36:29<dajobe>friend, employer, polices...
09:37:05<dajobe>[at this rate it's a race for tim to finish, run out of time, or me to run out of power]
09:37:49<dajobe>socially aware web
09:37:53<mhausenblas>ACTION ;)
09:37:57<dajobe>fed into access control
09:38:03<dajobe>--- Web I would like
09:38:07<dajobe>clean
09:38:11<dajobe>not grow in complexity
09:38:17<dajobe>like it to grow in cleanliness
09:38:22<dajobe>bye bye rdf reification!!
09:38:34<dajobe>move from legacy data
09:38:42<dajobe>to make future simpler data that is easier to parse
09:38:49<dajobe>we are making tomorrow's legacy data
09:39:28<dajobe>want to ask a questuon and for somethign to go ff to the data sources, delegate it and return the answer
09:40:03<dajobe>want data integrated to be well integrated
09:40:07<dajobe>maybe a new programming language
09:40:15<dajobe>so it is very easy to operate over hte web of linked ata
09:40:42<dajobe>and when that happens, want a really powerful, extenible security model for web apps
09:41:39<dajobe>capability-based systems
09:41:44<dajobe>would like more research in that area
09:41:48<dajobe>--- Our ROle
09:41:55<dajobe>Build a platform for others to follow
09:42:06<dajobe>do as Vint and others have done
09:42:15<dajobe>design to be useful general facility
09:42:20<dajobe>smooth, solid platform for building
09:42:28<dajobe>so people all over the web can build
09:42:42<dajobe>things we cannot necessarily imagine
09:43:30<dajobe>out of power...
09:44:15<mhausenblas>final words of TimBL: do your bit and everyone else does as well
09:44:18<mhausenblas>coffee!!!
09:50:15<libby>do your bit and do everyone else's as well? ;-)
09:50:40<bblfish>I just saw that timbl mentioned foaf+ssl?
09:51:11<libby>I guess they re all off to get caffinated
09:52:51<mischat_>hehe libby
09:55:24<melvster_>bblfish: yes, was some mention on twitter too http://search.twitter.com/search?q=%23www2009
09:56:25<bblfish>yes, nice.
09:56:31<bblfish>are the slides up somwhere?
10:10:43<mischat_>does any want to go to the open hack day 2009 in london, i have a place but I can't make it, have a ball to go to instead, any takers? if not i will email and ask them to give it to someone random ...
10:15:50<melvster_>mhausenblas: it seems possible to write exporter plugins for phpmyadmin: http://phpmyadmin.svn.sourceforge.net/viewvc/phpmyadmin/trunk/phpMyAdmin/libraries/export/
10:16:50<CaptSolo>hey mischat_ :)
10:17:03<melvster_>i would say this is the most used RDBMS tool on the web, export to RDF/XML or N3 may be able to be based on the existing xml.php file perhaps combined with triplify
10:17:29<mischat_>hey CaptSolo ... if it is regarding open hack day, i just pressed the "send" button on my mail client
10:18:03<CaptSolo>mischat_: it is but not re. tickets to it.
10:18:12<mischat_>fine :)
10:18:19<mischat_>go on
10:18:26<CaptSolo>just that you are going (or planned to) to all the cool events :)
10:18:51<CaptSolo>ACTION will be at hackday
10:19:02<CaptSolo>anyone else here?
10:19:41<mischat_>eh, i have recently moved to london, and am up for meeting people, and am no longer at a uni which was willing to send to conferences ;) I think some of the beeb folks are going to the event
10:19:49<mischat_>s/beeb/bbc/
10:20:26<danbri_>i have a ticket too, not 100% sure if can make it either
10:20:57<CaptSolo>home to see some of you there
10:21:05<CaptSolo>s/home/hope/
10:21:39<CaptSolo>danbri_: hey, you've got to keep traditions live :D
10:21:54<CaptSolo>"semweb gang at hackday"
10:38:15<dajobe>http://www.w3.org/2009/Talks/0422-www2009-tbl/
10:38:25<dajobe>F:|Tim B-L www2009 keynote slides - web at 20
10:38:36<dajobe>F:W3C ACL protected right now - some w3c staff can fix?
10:38:48<mischat_>password protected
10:38:52<mischat_>its password protected
10:39:08<dajobe>that's what I said
10:39:45<mischat_>my w3c login doesnt work either :(
10:39:50<mischat_>ah well ...
10:41:24<EtnaRosso>I'm looking for a tool to convert events (specified using Events Ontology) in iCal
10:56:46<Anchakor>"bye bye rdf reification!!"? I think reification is necessary in some cases... or dump make rdf use quads... :)
10:56:57<Anchakor>s/dump//
10:58:54<shellac>what do you need rdf reification for? and why would quads help?
10:59:28<Anchakor>when you want to express somthing about a triple
10:59:42<shellac>example?
10:59:57<shellac>I really mean: use case
11:00:24<Anchakor>same as named graphs
11:01:04<shellac>well we have named graphs for that, so you don't need reification or quads
11:02:00<Anchakor>yes, but in some cases I think reification is already used and rdf/xml support is really bad AFAIK
11:02:02<shellac>most of the use cases I come across can be covered by named graphs or (non-rdf) reification
11:03:26<shellac>yeah, we've never written this down. WAI, for example, were surprised that their use of rdf rei was questioned. It's in the specs, after all
11:03:44<shellac>ACTION will do this holy task at one point
11:04:30<melvster_>shellac: I've often wondered, what about graphs where you want to add a numerical weight to each edge? In mathematics this is called network theory (as opposed to graph theory)
11:05:36<Anchakor>s/support/support for named graphs/
11:08:09<Anchakor>can rdf reification instance be converted into named graph? (looselessly) (reverse is surely possible)
11:09:14<shellac>if you adopt certain conventions, maybe, but the semantics for named graphs are much weaker
11:10:18<Anchakor>this should be researched... else you wont be able to say goodbye to rdf reification ever because of legacy data
11:12:54<melvster_>not exactly a use case, but here is an example of finding a shortest path in a graph by assigning weights to the edges: http://en.wikipedia.org/wiki/Dijkstra's_algorithm
11:13:49<shellac>Anchakor, the thing to do is probably to write down the alternatives and make it obvious
11:14:12<shellac>the reasons to avoid it are buried in W3C mail archives, for example
11:15:11<dajobe>it's my opinion that you shouldn't be organizing or annotating rdf at the triple level
11:15:20<dajobe>it should be triple-molecules or graphs
11:15:21<shellac>melvster_: things like qualifying distances, or degrees of beliefs are classic examples
11:15:36<dajobe>so that's named graphs with annotations to replace reification of a single tripel
11:16:32<dajobe>the 'legacy use' of reification was the only reason it wasn't killed in 2004 when rdf was revised. it was weak then, but we were cautious
11:17:26<shellac>dajobe: I concur, although that misses the other use cases which n-ary relations cover
11:17:39<dajobe>true
11:17:53<dajobe>maybe rdf should have had n-ary (& lists & ...) from the start
11:18:08<dajobe>but then it would be lisp or prolog
11:18:45<EtnaRosso>dajobe simpler is better
11:19:11<melvster_>seems that the techniques for describing subjects objects and verbs are almost infinitely deep, but adverbs are not so well covered, and perhaps require rephrasing the sentence?
11:19:12<EtnaRosso>binary relations are enough
11:19:19<dajobe>generally, but you need to have enough simplicity. sweet spots like JSON and imho Turtle
11:20:02<EtnaRosso>if you need to put an adverb, really you are telling something about an assertion
11:20:13<EtnaRosso>so you need a new individual to represent this assertion
11:20:31<EtnaRosso>(i.e. reification)
11:20:35<dajobe>melvster_: sometimes you just have >2 resources involved in a sentence, n-ary where n>2 and adding a blank node is all you can do, a little awkward
11:20:36<Anchakor>I agree
11:20:46<Anchakor>(with EtnaRosso)
11:21:19<melvster_>dajobe: that seems fine, i havent yet thought of a case where that's not enough
11:21:30<EtnaRosso>thank you Anchakor
11:22:43<melvster_>for example the trust vocab has a <TrustStatement> with two parties and at trust level
11:23:04<EtnaRosso>trust vocab?
11:23:51<melvster_>see: http://www.w3.org/2001/sw/Europe/reports/trust/11.2/d11.2_trust_vocabularies.html#2
11:23:51<EtnaRosso>go to lunch
11:24:09<EtnaRosso>melvster_, see trust ontology by Golbeck
11:24:17<EtnaRosso>i don't like it too much
11:25:06<EtnaRosso>or the mine
11:25:11<EtnaRosso>i don't like it too much
11:26:31<melvster_>EtnaRosso: nice, looks similar
11:27:27<melvster_>but uses hasMainProperties and hasAuxiliaryProperties as a wrapper for for key value pairs
11:48:43<tobyink>melvster_: I have a foaf.me-related proposition for you.
11:49:25<melvster_>tobyink: im all ears :)
11:50:52<tobyink>Basically, I want you to have a script (probably needs an Apache rewrite too) so that for any URI of the form <http://mbox.foaf.me/bf903ff8b42453a0bae78b87b4e939cace7f4b8e> an RDF document is returned with the following triples:
11:52:02<tobyink><http://mbox.foaf.me/bf903ff8b42453a0bae78b87b4e939cace7f4b8e> foaf:primaryTopic <#me> . <#me> a foaf:Agent ; foaf:mbox_sha1sum "bf903ff8b42453a0bae78b87b4e939cace7f4b8e" .
11:52:24<melvster_>sure i can do that
11:53:00<ldodds>cool, another infinite uri space
11:53:06<tobyink>Later on we can add some extra coolness allowing people to add more information (e.g. foaf:name, foaf:homepage, owl:sameAs, etc) which can be authenticated via e-mail confirmation.
11:53:31<melvster_>well I already have sparul set up
11:53:43<melvster_>so you can add triples with tabulator
11:54:30<tobyink>ldodds: the set of all valid sha1sums is not infinite. It's just very big.
11:54:45<melvster_>so you can already do something like: http://foaf.me/mbox/test
11:54:51<melvster_>that autocreates an account
11:54:58<ldodds>tobyink: you know what I mean :)
11:55:51<melvster_>e.g. http://foaf.me/mbox/bf903ff8b42453a0bae78b87b4e939cace7f4b8e
11:56:48<tobyink>melvster_: yes, that's close already.
11:57:16<melvster_>that was autocreated just by hitting the URI
11:57:16<tobyink>Changes in logic needed would be to remove the default firstName, givenName.
11:58:30<tobyink>And to put in place a rule that says when the URI matches "mbox/foo" to not set a foaf:nick, but instead set a foaf:mbox_sha1sum (and strip the leading "mbox/" from that).
11:59:18<tobyink>Bonus points for checking that the sha1sum is exactly 40 hexadecimal digits.
12:08:46<melvster_>tobyink: http://foaf.me/mbox/newmbox
12:10:20<tobyink>yay!
12:12:35<melvster_>you should also be able to insert new stuff with tabulator
12:13:02<melvster_>click on the + icon, or change things with Enter
12:13:34<melvster_>i can change the controller to check that you're foaf+ssl authenticated too
12:13:42<melvster_>but it's a liberal policy for now
12:16:15<melvster_>ldodds: re infinite URI space a friend of mine started http://reverse.me.uk/ to see how deep google would reverse spider md5, they went quite far in, it's limited i guess by query string length, but is still a pretty big number
12:17:16<tobyink>http://buzzword.org.uk/2009/webid-finder/WebIDFinder/form.php
12:17:30<tobyink>G:|WebID Finder (pretty form)
12:18:07<melvster_>tobyink: very cool!
12:19:21<tobyink>The idea is to make this part of libre.fm's profile editing interface, allowing a person to find an existing WebID. (By default libre.fm gives everyone a WebID of "http://alpha.libre.fm/user/USERNAME#me".)
12:19:46<tobyink>That way, libre.fm can avoid minting a new WebID for people - it can simply use their old one.
12:20:06<melvster_>very nice
12:20:11<melvster_>so you can tie a webid to an account?
12:21:25<melvster_>we're starting to get quite federated! :)
12:21:32<MacTed>G: <tobyink> The idea is to make this part of libre.fm's profile editing interface, allowing a person to find an existing WebID. (By default libre.fm gives everyone a WebID of "http://alpha.libre.fm/user/USERNAME#me".)
12:21:37<tobyink>Yes - see http://alpha.libre.fm/user/tobyink - the RDFa includes about="http://tobyinkster.co.uk/#i"
12:21:48<MacTed>G: <tobyink> That way, libre.fm can avoid minting a new WebID for people - it can simply use their old one.
12:21:50<MacTed>:-)
12:22:55<melvster_>awesome
12:24:17<tobyink>G: actively seeking suggestions about what other sites provide nice webids. (right now, just My Opera and Identica/Laconica are handled)
12:24:40<melvster_>friendfeed?
12:25:12<melvster_>e.g. http://friendfeed.com/johnrod/subscriptions?output=foaf
12:26:23<tobyink>G: [http://goddamn.co.uk/viewvc/misc/WebIDFinder/|source code]
12:27:23<tobyink>Problem with friendfeed is that the foaf:Person is a bnode.
12:27:41<melvster_>ah yes
12:28:03<tobyink>Ditto livejournal.
12:28:15<melvster_>hi5 would be the biggest: http://api.hi5.com/rest/profile/foaf/39290024
12:29:24<tobyink>hi5 - also bnode. (rdf:nodeID is not the same as rdf:ID)
12:30:28<shellac>that feels like a typo in hi5
12:30:55<shellac>why would you bother giving a bNode id there, especially one called 'me'
12:34:08<tobyink>Any FOAF sites that link to the FOAF using rel="meta" will work using the "Your Website" option anyway. The special support for identi.ca/laconi.ca/My Opera just skips a few HTTP requests because they have predictable URIs, thus speeding things up a bit.
12:34:26<oshani>If you at WWW2009, and interested in Linked Data, there's a BOF happening right now at L1
12:34:29<tobyink>e.g. http://danbri.org/ works.
12:36:15<melvster_>tobyink: really good, it's such an important tool, because it translates something a normal user can understand ( homepage / account / email ) into something useful with linked metadata
12:39:33<libby>ACTION waves at metade
12:41:13<mhausenblas>BOF meeting at WWW09, now in room L1 Paris
12:41:34<mhausenblas>Topic: A research agenda for linked data
12:44:50<oshani>mhausenblas, can u paste the URI in here pls?
12:44:56<yvesr>mhausenblas: can you scribe what's going on in the BOF?
12:46:49<scor>ACTION waves at oshani and mhausenblas
12:46:53<mhausenblas>http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData/MadridBOF
12:47:04<mhausenblas>ACTION waves back
12:53:45<kidehen>tobyink: have you considered using: http://lod.openlinksw.com (which has REST API for what you see on front page) or http://lod.openlinksw.com/sparql to driver your WebID locator/finder?
12:53:46<phenny>kidehen: 02 Apr 17:16Z <mhausenblas> tell kidehen see http://chatlogs.planetrdf.com/swig/2009-04-02.html#T17-09-11 - ETA?
12:53:47<phenny>kidehen: 06 Apr 23:32Z <CaptSolo> tell kidehen an interesting thought re. Actions. could you write more about it to the SIOC list?
12:56:09<kidehen>phenny, tell tobyink: have you considered using: http://lod.openlinksw.com (which has REST API for what you see on front page) or http://lod.openlinksw.com/sparql to driver your WebID locator/finder?
12:56:10<phenny>kidehen: I'll pass that on when tobyink is around.
12:56:31<tobyink>i'll take a look
12:56:31<phenny>tobyink: 12:56Z <kidehen> tell tobyink have you considered using: http://lod.openlinksw.com (which has REST API for what you see on front page) or http://lod.openlinksw.com/sparql to driver your WebID locator/finder?
13:30:38<juansequeda>Hi everybody. We are at WWW and just organized a BOF: Research Agenda for Linked Data
13:30:47<juansequeda>We will update the wiki http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData/MadridBOF
13:30:53<juansequeda>but if you have any ideas, please contribute
13:33:16<yvesr>juansequeda: just thinking, but integrating linked data in desktop apps (e.g. music player, web browser, calendar, etc.) would be great.
13:40:14<Wikier>juansequeda: we'll read the minutes on the wiki, thx
13:42:55<Shepard>http://www.infoq.com/presentations/mnot-http-status-1108
13:43:00<Shepard>I:|HTTP Status Report
13:43:22<Shepard>I:mnot talk on HTTP history, HTTPbis work, state of implementations, future of HTTP
14:03:27<tobyink>kidehen: http://esw.pastebin.com/mb15ebd3 seems to work nicely at http://lod.openlinksw.com/sparql
14:06:49<mhausenblas>BOF RALD
14:07:10<mhausenblas>we have split into two groups
14:07:13<mhausenblas>http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData/MadridBOF/TOP10
14:07:29<mhausenblas>J:| TOP 10 list of linked data research issues
14:07:37<mhausenblas>http://esw.w3.org/topic/SweoIG/TaskForces/CommunityProjects/LinkingOpenData/MadridBOF/Communities
14:08:01<mhausenblas>K:| Community liaisons for linked data
14:10:58<tobyink>What is "BOF"?
14:24:05<Anchakor>juansequeda: you should put on the wiki also what "BOF" is... ;)
14:28:25<mhausenblas>BOF ... bird of feather
14:29:30<mhausenblas>http://www.webshine.org/2009bof.html
14:29:53<mhausenblas>L:| BOF Track
14:29:53<mhausenblas>18th International World Wide Web Conference,
14:30:05<mhausenblas>L:| BOF Track 18th International World Wide Web Conference,
14:32:20<mischat_>tobyink: yeah i have emailed hi5 about the bnode
14:37:40<BenO>RDFa q: how should i escape quote characters in literals ? e.g.
14:37:57<BenO><span property="dc:description" content="This file [..] he "tick" data, [..]ed."></span> <-- is unescaped and therefore broken
14:39:56<Pipian>BenO: Should be the standard &quot; entity, although I could be wrong.
14:39:59<PovAddict>BenO: that's a general SGML/XML question :)
14:40:01<PovAddict>&quot;
14:41:32<BenO>Pipian, PovAddict, Just checking :) now to fix the rdfa serialiser!
14:41:54<PovAddict>you have to quote & and < and > too
14:42:19<PovAddict>(as &amp; &lt; and &gt respectively)
14:42:46<BenO>PovAddict, sure, its just that the rdfa serialiser Im using isnt doing that atm
14:42:52<mattl>w/in 14
14:42:54<mattl>erk
14:49:48<cerealtom>metade: ping
14:50:35<metade>cerealtom: hi
14:50:43<tobyink>Have added my auto-find-webid-thingy to libre.fm's test site: http://alpha.dev.libre.fm/edit_profile.php
14:55:20<cerealtom>metade: you still want me to setup the mailing list?
14:55:59<kurtjx>tobyink: cool just added my foaf to libre.fm
14:56:41<mischat_>mischat_: kurtjx thanks for e-meeting my friend :)
14:56:44<metade>yes please
14:57:01<kurtjx>mischat_: np :-)
14:57:53<mattl>tobyink: not sure about support for "my opera" ;)
14:58:10<cerealtom>ok metade, gimme two minutes
14:58:28<tobyink>kurtjx: if you click the "check" button near location, it should also find the geonames URI for your location and include that in your profile.
14:58:46<tobyink>mattl: I can comment that out if you like.
14:59:34<mattl>is geonames free? i am just concerned that we might be relying on nonfree stuff
15:00:15<mischat_>geonames can take a long time to respond too
15:00:49<JRV>"The GeoNames geographical database is available for download free of charge under a creative commons attribution license." (http://www.geonames.org/about.html)
15:02:47<tobyink>We don't actually include any geonames data (yet) - just store a link to a place on geonames.org against each profile.
15:03:21<tobyink>so licensing shouldn't be much of an issue yet. I was planning to retrieve long, lat and country from geonames int he future though.
15:07:52<mattl>tobyink: CC-BY is fine :)
15:08:36<JRV>:)
15:08:36<PovAddict>does CC-BY mean you can make derivatives and change license?
15:08:48<tobyink>mattl: I've also recently brought ARC2 into the svn respository. It's tri-licensed under W3C license, GPL2 and GPL3. That OK?
15:09:03<PovAddict>(no derivs has its own code, derivs-and-keep-license has its own code, so that's what's left)
15:09:04<leobard>in Turtle/N3, did we ever agree how to identify THIS document? I tried " : a foaf:Document." but that doesn't parse in CWM. any idea?
15:09:28<metade_>cerealtom: rdf-hackers-london ?
15:09:33<mattl>tobyink: yeah, that's fine.
15:09:34<leobard>ACTION says hi, btw.
15:09:48<mischat_>melvster_: oooo
15:09:57<mattl>PovAddict: yes it does.
15:10:16<mischat_><> : foafDocument .
15:10:22<tobyink>leobard: <> a foaf:Document should work .
15:10:25<mischat_><> a foaf:Document
15:10:27<mischat_>:)
15:10:40<leobard>ACTION testing...
15:10:45<cerealtom>metade_: or london-rdf-hackers? (more aligned with london-pm, as a related/comparable community?)
15:10:59<cerealtom>metade_: sorry for slowness, got distracted by giovanni :)
15:11:01<mischat_>yay to london-rdf-hackers
15:11:06<metade_>that's fine :)
15:11:18<cerealtom>ok, on its way
15:11:24<metade_>great!
15:11:41<cerealtom>metade_, mischat_, who dya want as list owners? which mailboxes?
15:11:49<cerealtom>you can be cryptic is desired
15:11:53<swh>+1 to london-rdf-hackers
15:11:54<cerealtom>s/is/if
15:12:11<mischat_>am happy to do it
15:12:16<mischat_>garlik email address
15:12:55<cerealtom>ok mischat_; can add more afterwards
15:13:02<leobard>tobyink, mischat_: thx, that does something but not exactly what I thought... and its undocumented... I will try what Jena does with it....
15:16:24<cerealtom>ACTION drums fingers on the table, waiting for web interface to return after creation of list for metade_ and mischat_
15:17:05<mischat_>nice one tom
15:17:06<tobyink>leobard: Anything in angle brackets is a URI; relative URIs are allowed; the empty string is a relative URI referring to the current document URI (or rather to the current base URI, which defaults to the current document URI).
15:17:49<tobyink>mischat_ , cerealtom : may I join london-rdf-hackers even if I don't strictly live in London?
15:18:45<mischat_>of course :)
15:19:07<mischat_>cerealtom: i just got an email from DH
15:23:51<leobard>@tobyink, @dajobe, @timbl: could you someone add this to the N3/Turtle documents? it was not crystal clear to me that <> could be used to identify the "current document" to say "<> a foaf:Document".
15:24:11<cerealtom>mischat_: great :)
15:24:24<leobard>ACTION hopes that this would help...
15:24:46<kasei>heya ceraltom
15:24:56<cerealtom>mischat_: did it include the password?
15:25:37<mischat_>yeah it did
15:25:47<mischat_>shall i paste it in there ... hehe
15:25:51<mischat_>s/there/here/
15:26:01<cerealtom>kasei: hey :) good to "see" you
15:26:07<mattl>mischat_: we'll also need your PIN, your bank details and your mother's maiden name ;)
15:26:23<cerealtom>mischat_: ok, cool, so you can add other list admins, inc metade_?
15:26:27<kasei>how'd vocamp go? (last week of the term here, so haven't been keeping up with posts)
15:26:29<mischat_>leobard: you can also do <#me> a foaf:Person . which is nice short hand
15:26:32<cerealtom>via the mailman interface mischat_
15:26:42<mischat_>via I can do that
15:26:54<mischat_>s/via/yeah/
15:27:13<mischat_>metade_: which email address should I use ?
15:27:36<cerealtom>kasei: vocamp ibiza was good, thanks. i had connectivity issues, so not much in the way of blog posts or tweets, but was a good couple of days
15:27:38<metade_>my gmail please
15:27:45<cerealtom>kasei: just been talking to jim about vocamps
15:28:15<cerealtom>kasei: didnt manage to convince him of VoCampTroy2009, yet, but i'll keep working on him
15:28:17<kasei>ah, nice (too bad about connectivity issues.. you'd think we could sort that sort of thing out at this point, but... :)
15:28:21<cerealtom>;)
15:28:35<mischat_>cerealtom: vocamp bristol ?
15:28:36<kasei>haha. convincing anyone to do things in troy is a challenge.
15:28:57<cerealtom>mischat_: cool, so will leave you to get metade_, swh and others setup and just shout if you need anything from me
15:29:14<cerealtom>haha kasei, what, other than drinking and fighting? ;)
15:29:29<mischat_>yeah good stuff, will do the adding now, will show if i need anything
15:29:39<kasei>something like that
15:29:51<cerealtom>mischat_: yep, VoCampBristol2009 is coming, looking like mid-Sept. shellac is coordinating stuff re rooms
15:30:08<mischat_>excellent
15:31:40<cerealtom>kasei: will you make it to vocamp nyc
15:31:57<mischat_>metade_: i just made you admin too ... let me know if you dont receive an email from the maillist thing
15:32:17<metade_>mishcat_: thanks!
15:32:17<ldodds>cool re: VoCampBristol2009
15:32:28<mattl>Mm, Bristol.
15:32:33<kasei>ceraltom: is there a firm date yet?
15:32:45<kasei>i'd certainly like to.
15:32:46<mischat_>good work on vocamp stuff cerealtom
15:33:09<cerealtom>thanks mischat_ :)
15:35:22<PovAddict>"Given that the local loops alone in the USA used 58 light-minutes of copper wire[...]" holy crap
15:35:24<cerealtom>kasei: yes, there is a date set for vocamp nyc, just one day i think
15:35:53<cerealtom>kasei: just trying to check on vocamp.org but connection at www2009 is bad
15:36:13<kasei>the page I found on vocamp.org says Date TBA...
15:39:11<cerealtom>kasei: hmmm
15:40:11<cerealtom>in case anyone missed it, there is a now a new list for london rdf hackers: http://lists.linkeddata.org/listinfo.cgi/london-rdf-hackers-linkeddata.org
16:24:25<cerealtom>kasei, you there still? July 18th looks like the day for VoCamp NYC, but asking Christine Connors to confirm on the wiki if that date is fixed
17:18:05<melvster_>bengee: nice post, im also looking at automation rich internet applications
17:18:15<melvster_>i finally found a todo list ontology: http://tr.im/todoont
17:18:57<bengee>thx :) rdf/ical has todos, too, IIRC
17:19:11<melvster_>oh great, will look there
17:19:14<melvster_>which OS do you use?
17:19:22<bengee>and danja has a project vocab
17:19:43<bengee>win and mac
17:20:20<bengee>aiming at in-browser apps, though
17:20:30<melvster_>same with me, but i develop more for win, ill shoot you some of my apps when they are ready, they will be desktop based with web interaction, foaf+ssl is the glue
17:21:02<bengee>interesting
17:21:41<melvster_>this is the kind of thing i build off: http://www.autohotkey.com/forum/topic38309.html
17:23:12<melvster_>ive been modifying this to allow it a keystroke based todo list that interacts with other resources, and also can be a heads up display, ill allow it to log in to other services as your web id
17:23:25<bengee>aha, hardcore keyboarder ;)
17:23:40<melvster_>sort of, i need a todo list to be fast
17:23:45<melvster_>otherwise i ignore it
17:24:32<melvster_>but you can build very rich desktop apps this way
17:24:50<melvster_>as it can interface with any other app,. desktop or web quite easily
17:25:17<bengee>I saw some pretty cool mozilla labs experiments that worked similarly
17:25:19<melvster_>and you're set up in only a few lines, which is better than the overhead of flex / silverlight / xul
17:25:38<melvster_>yes the mozilla stuff is awesome, firefox 3.6 will change the way we do things
17:26:46<melvster_>but once you have an identity decoupled from your application you can be much more flexible with how and what you build, in theory everything should work with everything else
17:26:50<kasei>dajobe, around?
17:28:08<melvster_>so basically an in browser app should be the same functionally as a desktop app
17:30:21<melvster_>and all of the different apps, whether they be browser or web, desktop should be able to interoperate, then you have a truly powerful federated platform
17:31:08<melvster_>much like UNIX IPC, but the (semantic) web is the operating system
17:33:01<PovAddict>the day our operating systems become fullscreen browsers I'll stop upgrading
17:34:51<melvster_>might take a while ... :)
17:36:52<PovAddict>but I'm only 18 so I'll probably be still alive by then
17:36:53<drobilla>rdfcat /proc/apps | sparqlgrep '?app a semnix:Browser'
17:37:23<PovAddict>after seeing Debian packaging format, I do agree that we need more RDF client-side :D
17:41:37<melvster_>ah that looks better: http://tr.im/icalont
17:43:18<drobilla>yes we do. rdf in plugins and packages and such works extremely well
17:44:21<MacTed>bengee - just saw your new blog post... have you looked at ODS? OpenLink Data Spaces satisfies virtually all your technical wishes, tho beauty is in the eye of the beholder (or the skinner -- UI improvements are always possible)...
17:48:10<bengee>heh, yep, different simplicity/beauty requirements here, I'm afraid (no offense ;)
17:49:40<MacTed>call it a stupid question ... why not work on (at least by feedback, if not by actual work) a better skin for the apps that already do (most) everything technical?
17:50:28<MacTed>I ask because, to put it bluntly, you've described ODS in a nutshell ... and effectively suggested it doesn't exist
17:52:22<bengee>I prefer to understand the code I'm building on. that seems to impossible with openlink stuff that is marketed as being able to do just everything someone can ever imagine
17:52:34<bengee>seems to *be* impossible
17:52:58<MacTed>hm.
17:54:33<MacTed>perhaps I'm a bit oversensitive, but "marketed as" has a whiff of "of course it can't *really*" to it ... while to me, it seems like the functionality really *is* there.
17:55:05<MacTed>(not for *everything* you can imagine ... but the feature set you outlined? yeah, it can do or does now...)
17:55:17<bengee>the messaging is just kinda offputting for developers like me who like to do *something* themselves.
17:55:56<MacTed>perhaps I misunderstood your post.
17:56:01<MacTed>you want to *create* this suite?
17:56:06<MacTed>or you want to *find* it?
17:56:15<bengee>the german term is "eierlegende wollmilchsau", which is sort of the opposite of "do one thing, and do it right"
17:56:35<bengee>I guess I'll create the apps, yep
17:57:29<MacTed>the whiff has become a stench... the opposite of "do one thing, and do it right" would seem to be "do everything, but none of it right"
17:57:52<MacTed>thanks for the "compliment"
17:58:04<PovAddict><bengee> I prefer to understand the code I'm building on.
17:58:07<PovAddict>NIH syndrome
17:58:20<bengee>I'm sure ODS does "everything, and right", but I wouldn't know where to start to find it out
17:59:21<MacTed>a crazy suggestion -- by registering yourself at http://myopenlink.net/ ?
17:59:30<bengee>PovAddict, not really in that case. I've not seen a simple, usable RDF issue tracker or calendar yet
18:00:17<MacTed>ODS-Calendar can be used directly, or you can publish iCalendar to ODS-Briefcase. both deliver RDF of the events...
18:00:45<PovAddict>well, if you think you have to write an app from scratch, go ahead; as long as you use the existing formats :)
18:01:05<MacTed>the Virtuoso Sponger currently includes a Cartridge for Bugzilla; others won't be that hard to produce for other trackers
18:01:10<bengee>PovAddict, sure, that's the idea
18:01:38<kidehen>bengee: best I reply you via a blog post. If you don't know ODS well enough then don't speculate, really
18:01:39<PovAddict>for a moment I misunderstood you disagreed with an *ontology* and wanted to make your own; then I saw ODS was a service, not a format
18:01:59<MacTed>ACTION I will suggest a rewording of your blog post -- for I'm sure others will have the same interpretation I did -- "I want these apps, and they don't exist!"
18:02:24<kidehen>bengee: you put out a boiler list and ODS has handled a good chunk of that for eons
18:03:12<kidehen>MacTed: bengee is putting out a feeler, but irrespective, ODS handles what he wants to build, and in an open world, competition if very very good
18:03:47<kidehen>MacTed: better to have something that is like you when trying to describe yourself. Imagine trying to describe "Adam"
18:03:59<kidehen>MacTed: or "Eve"
18:05:08<MacTed>kidehen - agreed
18:06:36<kidehen>MacTed: there are 3 stages to all innovations: Crazy (everyone thinks you're crazy), Obvious (ah! so obvious, even co-opt it in many cases), and then Religion. This pattern has existed for ever, pre computer industry, and absolutely pre Web.
18:07:29<kidehen>MacTed: I do believe Alan Kay actually described this phenomena in one of his many presentations re. OOP and his misgivings about the OOP term as he saw C++ and MS-DOS etc..
18:08:31<kidehen>MacTed: what's most important is that the cohesion of the Web (courtesy of Linked Data) confines stage 2 to: the Obvious stage, the co-opting doesn't work, especially when the Web becomes a full DBMS and supports Outer Joins
18:09:34<kidehen>MacTed: so I don't worry about how and when the Data Space epiphany, its going to happen, just as the Linked Data epiphany is now taking shape etc..
18:17:41<MacTed>*nods* sure. I just got caught on the difference between "love to see" and "love to make". :-)
18:19:35<PovAddict>I'd love to see so many things
18:20:42<bengee>I get intermitting errors when I try to connect to myopenlink.net
18:21:43<MacTed>there seems to be something network-ish going on. lemme see what I can do.
18:22:28<bengee>no worries, had a quick test drive of the address book
18:25:07<kidehen>MacTed: simply demo site is http://demo.openlinksw.com/ods ; myopenlink.net is virual domain based and 8890 is sometimes more predictable for experimental initial / exploratory use.
18:39:52<Anchakor>melvster_: why care about platforms and not develop crossplatform desktop apps?
18:42:50<PovAddict>I'd rather use a Mac app made in Cocoa than one made in wxWidgets
18:43:33<mattl>PovAddict: can you tell the difference?
18:44:31<PovAddict>also, GTK on Windows feels quite ugly/alien
18:44:59<mattl>what about Qt?
18:45:14<PovAddict>I don't remember using Qt apps on Windows
18:46:33<PovAddict>even GTK on KDE is easily noticeable from the horrible file picker
18:46:58<mattl>GTK's my favourite.
18:47:09<PovAddict>if you're on GNOME :)
18:47:46<PovAddict>I have never *made* a GUI, mind you...
18:48:01<PovAddict>well, I once wrote a little game using Windows API from C :/
18:48:12<Anchakor>I have no favourite but gtk dialogs and keyboard shortcuts are aweful, qt seems most ok now... I coded with wxwidgets in the past
18:48:13<PovAddict>*pain*
18:48:40<mattl>Anchakor: which keyboard shortcuts are dumb on GTK?
18:49:26<Anchakor>mattl: for example in save dialogs...
18:50:09<Anchakor>do you know vimperator? it is the heaven of usebility for people using mainly keyboard... people should learn from it...
18:50:42<mattl>i know of it, but it's vim, so I don't use it
18:50:47<PovAddict>vim <3
18:51:08<Anchakor>mattl: it isn't vim... :)
18:51:22<mattl>Anchakor: it looks like vim, feels like vim and acts like vim.
18:51:40<Anchakor>PovAddict: gtk is beautiful compared to what windows apps look like in wine :)
18:52:01<PovAddict>Anchakor: well... yeah :)
18:52:12<PovAddict>I think mono winforms apps looked worse though
18:52:19<Anchakor>mattl: it looks like firefox, acts like firefox & vim + it has awesome feature called hints - press f and type...
18:52:48<mattl>Anchakor: yeah, i don't want anything that acts like vim ;)
18:53:06<Anchakor>mattl: you fundamentalist ;)
18:53:15<PovAddict>ACTION :wq's mattl
18:53:28<mattl>i'm putting the fun back into fundamentalism.
18:53:43<mattl>i've also been using emacs for a long time, and have no desire to move.
18:53:50<PovAddict>editing html with vim > Microsoft Word
18:54:04<PovAddict>or latex, but I have little experience with it
18:55:01<Anchakor>mattl: sure, but vimperator != vim and Im sure emacs people enjoy it to... (after all, browser is becoming an operating system as well... :))
18:55:32<mattl>Anchakor: i really hope the browser never becomes an operating system :)
18:56:01<Anchakor>me too (maybe because I like vim and unix philosophy :))
18:57:17<PovAddict>Anchakor: http://news.povray.org/49ef4c67%40news.povray.org
18:57:30<PovAddict>I got him a collection :D
19:04:01<melvster_>Anchakor: eventually it will be portable people are working on a mac and linux version, and wine works too, but the main reason I use the OS is because I built it for myself to use, and wanted it to be utlra fast
19:06:19<Anchakor>melvster_: well any why not use crossplatform libraries? I coded in wxwidgets and it worked without any change on windows when I tried...
19:13:45<Anchakor>MacTed: for some people (me included) it can be a turn off if the desired application depends on ton of other things, which I suspect might be the case of ODE...
19:14:39<Anchakor>MacTed: and definately some things people prefer to make, not just see... love for example :)
19:24:07<kidehen>Anchakor: we have a wonderful way of being perceived as the opposite of what we are :-( Our products support many protocols so as not to have N^2 component dependencies. That's the base architectural philosopy
19:25:44<kidehen>Anchakor: adhering to industry standards is supposed to be a wonderful thing, right? Compare Virtuoso, ODS, or ODE to your typical LAMP stack, no comparison re. dependencies. Even more, we are not Linux only, we run across all the major platforms (again base philosophy, hence our company name etc..).
19:26:33<kidehen>Anchakor: we write in 'C' not due to some form of masochism, just because it is still the language that runs on most operating systems. In many cases, we use Virtuoso as our sort of VM :-)
19:27:56<MacTed>Anchakor - as Kingsley says -- ODE (OpenLink Data Explorer) and ODS (OpenLink Data Spaces) depend on Virtuoso. which depends on, um, the OS. and the network. and not much more.
19:29:25<MacTed>the browser-hosted version of ODE doesn't even depend on Virtuoso. it does need an RDFizing proxy, and defaults to Virtuoso -- but any suitably standards-compliant target can be used.
19:30:09<kidehen>Anchakor: problem with many things today is that History is worth zilch. We remake history from today onwards, and then we loose all the lessons that history truly teaches.
19:30:09<kidehen>Anchakor s/loose/lose
19:31:45<Anchakor>kidehen: you are right, people judge by impressions, I guess my impression was formed by the all-in one webserver concept, and might be wrong
19:32:08<kidehen>Anchakor: don't conflate standards compliance with component hell, really. We are about sanitizing the mess we have today via the concept of a Universal Server Platform. A platform you simply switch on (LAN or Internet WAN). It deals with Data is very sophisticated ways, doesn't seek to recreate paper, and absolutely doesn't see the document as ground zero. Data Space is ground zero in our world view.
19:33:48<PovAddict>I just imagined a world where you could explain something in machine-readable format, and the computer could turn it into a document... in any natural language
19:34:09<PovAddict>I think it would work better than trying to auto-translate from a natural language to another
19:34:10<kidehen>Anchakor: I did a podcast with Jon Udell a few years ago, worth a listen, I explain our vision. See: http://www.openlinksw.com/blog/~kidehen/index.vspx?page=&id=965
19:34:23<PovAddict>since the original would have no ambiguities
19:34:52<bengee>ACTION updates the blog post, starting to sound a little like an anti-rdf troll in the making ;)
19:35:08<Anchakor>I really like what openlink is doing and did try virtuoso and will use it in future on server witch just recently became available to me, but I understand some people don't have dedicated server, so they for example rely on having their apps run on common php+mysql webhosting
19:35:24<Anchakor>s/witch/which/
19:35:25<bengee>hope I don't become the next crschmidt..
19:40:22<mattl>bengee: what does that involve?
19:40:38<bengee>err, now this sounds like crschmidt is a troll, he clearly isn't.
19:40:56<mattl>heh. he's a nice guy.
19:41:22<PovAddict>ACTION has to go study boring stuff, will then come back here to explain a problem that RDF could solve
19:41:42<bengee>just a formerly active RDFer who discovered that he could be more productive without using RDF
19:42:05<Anchakor>ACTION does sturdy boring stuff in C all day
19:42:27<Anchakor>s/sturdy/study and code/ :)
19:42:54<Anchakor>(hopefully would be sturdy too :))
19:43:05<PovAddict>mine's more boring than you
19:43:27<PovAddict>yours*
19:43:59<Anchakor>ok, I admit it probably is :)
19:48:20<MacTed>huh. spam snip? really?
19:48:58<MacTed>ACTION because I said that the features you explicitly highlighted -- SPARQL and RDF -- were delivered by Virtuoso?
19:52:37<snail>PovAddict: re your "just imagined a world where you could explain something in machine-readable format, and the computer could turn it into a document... in any natural language" comment, you've seen GRDDL, right? http://www.w3.org/TR/2007/REC-grddl-20070911/
19:55:31<PovAddict>hmm
19:55:45<kidehen>Anchakor: you do know that Virtuoso and ODS exist in AMI form on EC2. Even more, AMIs exist for: DBpedia, MusicBrainz, NeuroCommons, Bio2RDF, and soon entire LOD-Cloud (so can instantiate your own: http://lod.openlinksw.com instance via an AMI)
19:57:13<kidehen>Anchakor: as with most things, the really hard issues have been addressed and are on display re. http://lod.oepnlinksw.com (scalable RDF management and faceted search + find, VoiD based stats, all the stuff that has challenged the RDF realm for a long time).
19:57:13<Anchakor>kidehen: AMI?
19:57:30<kidehen>Anchakor: Amazon EC2 AMI
19:58:02<Anchakor>aah
19:58:05<kidehen>Anchakor: links coming: http://virtuoso.openlinksw.com/wiki/main/Main/VirtInstallationEC2
19:59:30<kidehen>Anchakor: AMIs are about "Data as a Service" (the new electricity), just put on your AMI, get loaded and functional data spaces, do your thing via whatever you offer the world via your Linked Data oriented app. or service. Again, solving real problems.
20:00:10<kidehen>Anchakor: We will even go retro by making USB keys available with the entire LOD-Cloud data on board :-)
20:01:14<kidehen>Anchakor: or you get a key for you desired bubble from the LOD-Cloud e.g. if you want DBpedia, then you can order an DBpedia key etc.. Load the data into Virtuoso or any other RDF store capable of handling the task.
20:01:38<Anchakor>kidehen: I see, that is one way to do that, perhaps too new for most people (and maybe expensive)... to something a bit different: how far is development on federated query across the whole web of data?
20:06:02<kidehen>Anchakor: SPARQL federation works starts after we are done with Virtuoso 6.0 initial release (this week or next) . Remember, we federated SQL in 1998 with location aware query optimization. We want to do the same for SPARQL. Discoverability of sparql endpoints is vital, ditto stats about the state of the data behind the endpoint. Anyway, it is coming, and it will be delivered in typical OpenLink fashion (i.e. leveraging years of experie
20:06:35<kidehen>Anchakor: but we want to close the lid on the Cluster Edition which is more or less now done.
20:08:39<Anchakor>kidehen: sure, I look forward to it... do you plan to provide some limited free service for it?
20:09:47<kidehen>Anchakor: yes, of course, we are all ears about risk free evaluation options. At very least there will be 15-30 day free evals, and AMI options re. EC2
20:10:10<melvster_>what exactly is federated sparql?
20:11:00<kidehen>melvster_: automatically routing SPARQL out to another SPARQL endpoint based on analysis of a graph pattern that is scoped to one of more Named Graphs.
20:11:12<LeeF>Or not so automatically, for that matter.
20:11:39<Anchakor>melvster_: basically a sparql query spread across multiple endpoints
20:11:39<LeeF>Federated SPARQL is about delegating portions of a query to other query endpoints, and then combining the results to answer the original query
20:11:39<melvster_>to exactly one sparql endpoint, or is there scope for more than one?
20:11:50<LeeF>melvster_: more than one, in the usual scenarios
20:11:50<kidehen>LeeF: SPARQL endpoint discoverability is the key here, do we maintain a list or discover over DNS etc..
20:12:12<LeeF>kidehen: Maybe *you* think it's key :-), but I don't think that's an absolute truth at all.
20:12:31<melvster_>that sounds pretty awesome
20:12:41<kidehen>LeeF: endpoint discoverability? of the use of DNS?
20:13:32<kidehen>LeeF: We need to find an endpoint and ideally be able to determine what it has, statistically. No stats, no smarts, basically.
20:14:04<LeeF>kidehen: I was just saying that automated federation is a great goal, but it's not the only useful use of federated query, so it's a bit misleading to define it as such
20:14:31<LeeF>kidehen: lots of people get a lot of benefit from explicitly directed federation (c.f. SERVICE in ARQ and (i understand) pragmas in virtuoso)
20:14:55<kidehen>LeeF: Yes, but we want to go beyond that, of course
20:15:35<melvster_>im not sure dns is the best option for discovery, as it has a setup overhead
20:16:16<kidehen>melvster_: DNS is an option, not panacea.
20:16:18<PovAddict>Link: header?
20:16:58<kidehen>melvster_: have you read: http://www.floop.org.uk/eagle/discovering-sparql ?
20:17:42<kidehen>melvster_: if you have an dns-sd, try: dns-sd -B _sparql._tcp floop.org.uk
20:18:15<kidehen>melvster_: or dns-sd -B _sparql._tcp openlinksw.com
20:19:34<kidehen>melvster_: these things take time to get going, but when we put heads together constructively, we can make the erstwhile challenging, possible :-) The building blocks are taking shape.
20:20:12<kidehen>melvster_: building a VoiD graph for the LOD cloud hosting instance of Virtuoso: http://lod.openlinksw.com/void/Dataset is also part of the ground work
20:20:48<melvster_>nice article, i hadnt read it, i know about piggybacking off dns, is that how bonjour works?
20:21:09<kidehen>melvster_: Bonjour is mDNS, yes
20:22:36<melvster_>ah multicast he always wanted to do that
20:22:56<kidehen>melvster_: melvster_: the goal is to make creation of entire LOD cloud instances, or specific bubbles easy for anyone that needs to use this data corpus as a starting point. Goal is to diminish the pivotal role of the LOD cloud, DBpedia over time, they are are important seeds, not panaceas re. Linked Data Web and eventual "Data as a Service" based economny
20:23:07<kidehen>melvster_: s/economny/economy
20:23:50<kidehen>melvster_: grand scheme is to enable easy realization and comprehension of: URI as new medium of value exchange reality
20:24:00<kidehen>melvster_: re. Web aspect of Internet
20:25:28<kidehen>melvster_: the real fun starts when old media companies realize that there is immense value to be culled from meshing their data with LOD-Cloud data with the details of those meshes grounded in URIs minted from their domains etc..
20:25:32<melvster_>yes is very interesting, i need to give it a bit more thought, the guy who wrote bonjour taught me to program when i was a kid, but i didnt read up on it till now, i think it's also important for devices or clients to be sparql endpoints, but as you say you need a system 'that just works' like bonjour ... it's a nice page i need to think a bit more about discovery ...
20:27:45<kidehen>melvster_: yes. we need to remember what's staple on the desktop and apply to burgeoning Linked Data Web re. higher level apps and interfaces
20:27:51<melvster_>id never thought of it as data as a service, but you're right
20:28:17<kidehen>melvster_: data is the new electricity, just takes a while for it to sink :-)
20:29:15<melvster_>it's funny for an economy you need transactions, but transactions is just data
20:29:16<kidehen>melvster_: put valuable stuff behind your URIs and it will be a source of vitality in the dense mesh we know as the Web
20:29:24<kidehen>melvster_: yep!!
20:29:47<kidehen>melvster_: the economic crisis is putting the finishing touches to old currency as I type :-)
20:31:17<melvster_>according to the IMF we havent had the crisis yet, we're less than 1/4 of the way through it, but that's another story ... :)
20:31:52<melvster_>i guess we need technology to be able to allow us to do more on a limited budget, to help us recover
20:32:45<melvster_>yes perhaps data is the currency of the new ecoomy, i handnt though of it that way before, but it makes some sense
20:32:47<kidehen>melvster_: Yes, and that will always be the case. "Do more with less", what technology enables you "Do more with less" and then virally propagate this new kind of value through the ecosystem.
20:33:00<kidehen>melvster_: the opposite happened en route to current crisis :-(
20:33:18<kidehen>melvster_: spend Billions producing nothing of value bar NOISE.
20:34:18<kidehen>melvster_: so in the past Marketing had sway of Quality. The scales are about to tip the other way, because: "you have to do more with less".
20:35:51<kidehen>melvster_: customers can't afford the: "buy my way out of trouble" approach anymore. Example, look how long data integration has been a mess globally, and its only gotten worse each year.
20:36:07<kidehen>melvster_: Linked Data is the ultimate Data Integration Middleware :-)
20:36:29<kidehen>melvster_: and the unit of value is simply the URI, what data meshes it exposes etc..
20:36:57<kidehen>melvster_: anyway, baby calls, gtg :-)
20:45:34<melvster_>:)
20:55:37<Anchakor>would be interesting if later, when data as service is fully deployed, some kind of currency would evolve, which you could trade with other people - "Would you trade me <commodity> for 30000 triples from my web of data cache?" :)
20:58:06<kidehen>Anchakor: yes, but its not about a bland # of triples, rather, access to your data space. Basically, its all about "You" :-) "You" are the new media magnate :-) This is why your Web ID is so vital :-)
20:58:24<kidehen>Anchakor: see my blog posts for more, truly gtg now :-)
20:59:09<Anchakor>kidehen: yes... well more than you rather your data :) (data=money - we are back to where we are now! :))
21:00:11<kidehen>Anchakor: but currency changes :-)
21:00:14<kidehen>bye
21:00:38<Anchakor>bye
22:14:40<melvster_>"If you?re running on OSX, then mDNSResponder is already installed; if you?re running on Linux, then you?ll have to download, compile and install the package from Apple (again, see later episodes); if you?re on Windows, you?re on your own."
22:24:39<snail>"if you're melvster_ you have character encoding issues"
22:42:13<PovAddict>melvster_: package from apple? doesn't avahi work?
22:46:27<melvster_>snail: yes sorry, cut an paste from http://www.floop.org.uk/eagle/discovering-sparql PovAddict: yes avahi works
22:49:10<snail>Does someone have a good example of how to encode "Agent A believes predicate X relation Y" ideally in RDF/XML ?
23:04:23<kidehen>http://knol.google.com/k/twain/the-global-brain-the-semantic-web/31fjy9fjsu1x2/19#
23:04:44<kidehen>M: Interesting page developing on Knol re. Global Brain
23:19:29<kasei>snail: I'd imagine you'd reify on X (giving you statement Z), and then assert A believes Z.
23:21:14<kasei>oh, i may have misread the "predicate X relation Y" bit... what does A believe? (X :rel Y)?
23:24:26<kasei>something like this: <rdf:Description rdf:about="X"><relation rdf:ID="Z"><rdf:Description rdf:about="Y"/></relation></rdf:Description>
23:24:31<kasei>... <rdf:Description rdf:about="A"><believes rdf:resource="Z" /></rdf:Description>
23:25:14<kasei>this produces both (X relation Y) and the reification of that statement. it's more wordy if you just want the reification (without the actual statement).
23:44:51<snail>kasei: i just want to reification. i'm going to have to read up on reification, i suspect
23:45:22<snail>kasei: thanks

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