IRC log of dig on 2006-06-22
Timestamps are in UTC.
- 17:37:19 [DIGlogger]
- DIGlogger (n=dig-logg@pink-panther.csail.mit.edu) has joined #dig
- 17:37:19 [brown.freenode.net]
- topic is: Decentralized Information Group (DIG) <http://dig.csail.mit.edu/>
- 17:37:19 [brown.freenode.net]
- Users on #dig: DIGlogger sheets djweitzner hmslydia timbl alerer lkagal evoli ryanlee yosi_s jamesdh eikeon claudiof RalphS tlr erim sandro DanC
- 17:37:22 [sheets]
- Adam: Have tested some cases. Some tabulations don't work
- 17:37:22 [timbl]
- agenda+ demo of constraints
- 17:37:36 [sheets]
- Adam: There may be some bugs in tabulating. I am now working on sparql
- 17:37:49 [sheets]
- Adam: Working on sparql->kb transforms
- 17:37:54 [sheets]
- Tim: sparql server?
- 17:38:19 [sheets]
- Adam: I want to use nearly the full sparql standard.
- 17:38:43 [sheets]
- Tim: maybe just a client for other people's sparql stores?
- 17:39:01 [sheets]
- Adam: We can attach the translation layer to anything and connect to independent people
- 17:39:15 [sheets]
- Adam: independent people = independent services
- 17:39:29 [sheets]
- Tim: A good milestone might be to point to a public sparql service
- 17:39:52 [DanC]
- http://esw.w3.org/topic/DawgShows
- 17:40:04 [DanC]
- http://sparql.org/sparql.html
- 17:40:08 [sheets]
- DanC: "DawgShows" is an example: http://esw.w3.org/topic/DawgShows
- 17:40:34 [sheets]
- DanC: govtrack has 10e7 triples
- 17:40:45 [sheets]
- DanC: XMLArmyKnife also interesting
- 17:41:26 [sheets]
- DanC: can type in zip code and do a sparql query for legislation that impacts you
- 17:41:40 [sheets]
- Jim: fixed horrible not loading triples in map bug
- 17:42:00 [sheets]
- Jim: implemented thumbnail view in map -- a bad hack, used filetype extensions
- 17:42:05 [DanC]
- the zip code thing was an idea/wish. It's doable in theory. In practice, XMLArmyKnife just has a bunch of data and a SPARQL service
- 17:42:22 [timbl]
- agenda+ rdfs for finding what an image is
- 17:42:39 [sheets]
- Jim: I should rewrite this to use RDF filetype data. We should talk about view interrelations.
- 17:42:45 [DanC]
- http://xmlarmyknife.org/
- 17:43:14 [sheets]
- Tim: one of our primary objectives is to use web architecture correctly
- 17:43:31 [sheets]
- Tim: I will be publically flogged if we use filename extensions
- 17:43:50 [DanC]
- http://xmlarmyknife.org/docs/rdf/sparql/
- 17:43:50 [sheets]
- Tim: we should definitely use the metadata
- 17:44:05 [sheets]
- Jim: there are some limitations to this approach -- need rdfs
- 17:44:24 [sheets]
- Jim: can open info bubble entries into new outline view
- 17:44:40 [sheets]
- Tim: same window?
- 17:45:15 [sheets]
- Jim: we can make it a different window. We need to rewrite some code but it's doable. Let's think about UI concerns.
- 17:45:36 [sheets]
- Adam: we would have to open a new tabulator instance
- 17:46:00 [sheets]
- Tim: you can use DOM to remote-control other windows
- 17:46:14 [sheets]
- Tim: we should clean up global variables
- 17:46:30 [sheets]
- Adam, Joyce: myquery shouldn't be a global variable!
- 17:46:34 [sheets]
- Tim: refactor!
- 17:46:51 [sheets]
- Jim: isn't there a memory constraint to doing this
- 17:46:55 [sheets]
- Tim: share the kb
- 17:47:16 [sheets]
- Tim: a map should be able to aggregate multiple queries. Maps should be singleton?
- 17:47:50 [sheets]
- Jim: rewriting map view to be standalone class
- 17:48:16 [sheets]
- Jim: making this modular so we can aggregate and reuse this -- leave it up to a later developer
- 17:48:38 [timbl]
- sheets:
- 17:48:49 [DanC]
- ScribeNick: timbl
- 17:48:53 [timbl]
- ... i hve nearly finished the new parser fopr RDF/XML.
- 17:49:10 [timbl]
- ... Have run through 20 odd test sets, each of whgich is 1..20 rdf and NT files.
- 17:49:28 [timbl]
- .. It works for most. The difficulties are in the comparnf of te output files...
- 17:49:44 [timbl]
- I have been used hacked up code in [missed]
- 17:49:55 [timbl]
- DanC: Ca you get the results onto a disk?
- 17:50:28 [timbl]
- sheets: Currently I loop dearching for a match for each triple, but i can't do bnodes.
- 17:50:40 [DanC]
- tim, does the tabulator have a graph comparison function yet?
- 17:50:43 [timbl]
- ... It passes all of the major tests, but collections will need changes to thhe store.
- 17:53:59 [timbl]
- ... I tried to get unicode unescapingg o be done by jscript, but couldn't
- 17:54:30 [DanC]
- (you could just leave the strings in their literal form, for the purposes of comparison)
- 17:54:41 [timbl]
- .. The two failing test cases rae XML serialization and blank node ids not matching up
- 17:54:50 [timbl]
- danc: COuld test that with query
- 17:55:14 [timbl]
- sheets: I fixed the unicode escapes with regexp
- 17:55:23 [timbl]
- danc: Just leave them unescaped.
- 17:55:38 [timbl]
- sheets ... don't need to , unscaping done.
- 17:56:36 [sheets]
- Tim: Norm Walsh's foaf fails
- 17:56:46 [DanC]
- timbl: norm walsh's foaf file fails with the old RDF/XML .js parser
- 17:57:42 [timbl]
- Reseacrh Science Institute
- 17:58:03 [timbl]
- ... calendaring and paper generation and mailing lists ... some data will end up in DF
- 17:58:13 [timbl]
- s/D/RD
- 17:58:28 [DanC]
- re agendas in N3, see http://esw.w3.org/topic/MeetingRecords
- 18:03:08 [sheets]
- Tim: SQL data for dbview->tabulator?
- 18:03:14 [DanC]
- http://www.w3.org/2006/05/dbview/
- 18:03:21 [sheets]
- sheets: I have a music library in postgres
- 18:03:30 [sheets]
- Tim: sounds good, put it in DIG space
- 18:03:34 [timbl]
- ----------------------
- 18:03:37 [DanC]
- timbl: don't think I committed the dbview overview thing at all
- 18:03:49 [DanC]
- agenda = Calendar walkthru, DIG channel privacy, Parser walkthru, blanknodes, good/needed datasets
- 18:04:03 [sheets]
- DanC: I have been working towards turtle or n3 parser in js
- 18:04:03 [timbl]
- grep agenda+ <>
- 18:04:19 [timbl]
- check.py
- 18:04:26 [sheets]
- DanC: check.py hardening?
- 18:05:11 [sheets]
- DanC: i have n3 grammar in rdf. parsed grammar and got json data structure. Can do lexical analysis now.
- 18:05:22 [sheets]
- Tim: no branching table?
- 18:05:37 [sheets]
- DanC: not yet. I keep reproducing your (Tim) work
- 18:05:39 [DanC]
- http://www.w3.org/2000/10/swap/grammar/notation3.bnf http://www.w3.org/2000/10/swap/grammar/notation3.json etc.
- 18:05:53 [DanC]
- http://www.w3.org/2000/10/swap/grammar/Makefile
- 18:06:06 [sheets]
- Tim: first on the agenda: calendar demo
- 18:06:18 [timbl]
- http://dig.csail.mit.edu/2005/ajar/ajaw/tab
- 18:06:19 [sheets]
- Joyce: try on your computer
- 18:06:34 [DanC]
- I also have some stuff in dig-svn/2005/ajar/ajaw/grammar, but I haven't checked it in :-/
- 18:06:35 [sheets]
- Joyce: try the TAGMobile roadtrip
- 18:07:09 [sheets]
- Tim: sorry about the GPS data
- 18:07:58 [sheets]
- DanC: we need a link to it from DIG
- 18:08:04 [sheets]
- Jim: it's pretty high on google
- 18:08:15 [sheets]
- Tim: we should link from DIG
- 18:08:23 [DanC]
- http://dig.csail.mit.edu/2005/ajar/ajaw/tab.html
- 18:08:30 [sheets]
- Joyce: try ColorPicture and use date or date-time
- 18:08:33 [timbl]
- open "TAGmobile..."
- 18:08:44 [timbl]
- 1. openm tabulator proect
- 18:08:51 [timbl]
- 2. open testdataset: tagmobile
- 18:09:14 [timbl]
- 3. focus on Colour picture
- 18:10:11 [timbl]
- open http://www.w3.org/2004/lambda/Pictures/iPhoto%20Library/2006/06/12/IMG_1758.JPG
- 18:10:15 [timbl]
- typeexpandloadedColorPicture
- 18:12:57 [sheets]
- DanC: multiple select on *nix machine/x windows?
- 18:13:09 [sheets]
- Tim, Joyce, sheets, Jim: we don't know but it's fixable
- 18:13:09 [timbl]
- select datetime OR date
- 18:13:13 [timbl]
- hit 'calendar
- 18:13:15 [timbl]
- ''
- 18:13:23 [timbl]
- ... wait.
- 18:13:47 [sheets]
- Tim: we're looking up images
- 18:15:06 [sheets]
- Joyce: not getting data when you don't refocus seems wrong
- 18:15:19 [sheets]
- Tim: it's not a bug. "mentions" is not a real link
- 18:15:27 [sheets]
- Joyce: it looks like a real link
- 18:15:46 [sheets]
- Tim: you want to narrow the search so refocus is a good idea
- 18:17:14 [sheets]
- Joyce: i don't like the stretch... calendar can get huge
- 18:17:31 [sheets]
- Tim: stretch is good (calendar view stretch)
- 18:17:51 [sheets]
- Joyce: maybe i will add minimize and maximize
- 18:18:11 [sheets]
- DanC: labels on pictures are not cheap?
- 18:18:14 [sheets]
- Tim: not cheap
- 18:18:46 [sheets]
- Tim: how to handle thumbnails is something to think about -- rdf? exif?
- 18:18:51 [DanC]
- labels on *these* pictures, I mean.
- 18:19:03 [sheets]
- Joyce: Jim loads the full picture
- 18:19:13 [sheets]
- Jim: I might be able to make it only load the image on event bubble opening
- 18:19:22 [sheets]
- DanC: let's use coordination group data
- 18:19:59 [sheets]
- Tim: there is no reverse of "mentions"
- 18:20:09 [sheets]
- Tim: that's why this may be a problem
- 18:20:44 [sheets]
- Tim: there's a whole section in the About document about inverting the outline view
- 18:20:59 [sheets]
- Tim: take the tree and pick it up by a different root
- 18:21:28 [sheets]
- DanC: I refocussed on the coord. group calendar
- 18:22:30 [sheets]
- DanC: I get nothing when i try to calendarize the thing of type vcalendar
- 18:22:51 [sheets]
- Joyce: when I only know the end of an event, I give up
- 18:23:18 [sheets]
- Tim: we should use RDFS to figure out how to handle this data too (as well as the images)
- 18:24:23 [sheets]
- DanC: i don't think that the calendar and the map use the same query paradigm
- 18:24:33 [sheets]
- Jim: you can use basedNear or lat/long
- 18:24:40 [sheets]
- Tim: we should think about consistent UI
- 18:25:07 [sheets]
- Tim: we should visually tag the properties that are viewable with our views
- 18:26:02 [sheets]
- DanC: we need to justify this with the internet
- 18:26:13 [DanC]
- http://www.w3.org/2002/12/cal/rfc2445.html
- 18:26:29 [sheets]
- DanC: (the way to handle the calendar data and map data selection UI)
- 18:26:44 [DanC]
- dtend http://www.w3.org/2002/12/cal/rfc2445.html#sec4.8.2.2
- 18:27:36 [sheets]
- Tim: iconify views next to viewable data
- 18:27:58 [sheets]
- Tim: the multiple selection interface is hard for Mac people
- 18:28:05 [sheets]
- Tim: piecemeal queries may be better
- 18:28:13 [sheets]
- Adam: Mac users are a minority
- 18:28:33 [sheets]
- Tim: Adam, you think about the power user. We should think of a range.
- 18:28:47 [sheets]
- Adam: I don't think that this is really useable to a wide range of people
- 18:29:10 [sheets]
- Tim: this is proof-of-concept. UI is a greater concern than technical boundaries
- 18:29:24 [sheets]
- Adam: too many tricks like "mentions"
- 18:29:49 [sheets]
- Tim: we should smooth things like that out. Agendize making mentions work logically
- 18:30:14 [sheets]
- Tim: "mentions" is totally in the view
- 18:30:30 [sheets]
- Tim: "mentions" is due to type
- 18:30:51 [sheets]
- Tim: this is probably not a good way to do this... too much data in some cases
- 18:30:59 [sheets]
- Adam: or maybe we should ditch mentions
- 18:31:15 [sheets]
- Tim: you have no way of bootstrapping a file then -- it's a kludge
- 18:31:40 [sheets]
- Tim: you could do some kind of serialization to get similar functionality?
- 18:32:20 [sheets]
- Tim: we could make a subject list to bootstrap a file
- 18:32:50 [sheets]
- djweitzner: I'm looking for my power adaptor
- 18:33:20 [eikeon]
- eikeon has quit ()
- 18:33:32 [sheets]
- Jim: how do we want view data selection to work?
- 18:33:57 [DanC]
- the stuff about dtend default is in http://www.w3.org/2002/12/cal/rfc2445.html#sec6 ; sort of a wooly part of the spec. In practice, the lack of a dtend is just like when somebody says "the meeting is at 2pm". It ends some time later.
- 18:34:06 [sheets]
- Jim: should we use selections in the iconic view creation?
- 18:34:49 [sheets]
- Tim: if selection then use those data. if not, then just graph point. Aggregate on one map. One day in another window.
- 18:34:55 [alerer]
- alerer has quit (Read error: 104 (Connection reset by peer))
- 18:35:26 [DanC]
- dtstart is required, evoli
- 18:35:38 [DanC]
- if you only have dtend, you cry foul ;-)
- 18:35:39 [sheets]
- Tim: there should be a big global symbol for the map
- 18:35:45 [evoli]
- DanC: alright then!
- 18:36:23 [sheets]
- Jim: if you break the map out into a separate window and then you wanted to send more data to that map... i was getting tripped up
- 18:36:42 [DanC]
- a wish: a trip view, which is a slider + a map. When you move the slider, you see where the trip passenger is
- 18:36:52 [sheets]
- Jim: I was trying to figure out how to code so the UI is the most intuitive for map viewing
- 18:37:03 [evoli]
- slider: time ?
- 18:38:13 [sheets]
- Jim: we might be able to use a scrollbar with event registrations... i think the map API can support it
- 18:38:24 [sheets]
- Tim: time slider is post release
- 18:39:15 [sheets]
- Tim: calendar walkthru: check!
- 18:39:25 [DanC]
- agenda = DIG channel privacy, Parser walkthru, blanknodes, good/neededdatasets
- 18:39:49 [sheets]
- Tim: channel privacy?
- 18:39:55 [sheets]
- Tim: logged but group only?
- 18:40:03 [sheets]
- DanC: that is my preference
- 18:40:18 [sheets]
- DanC: passwords are hard. I can't get at the mail archives.
- 18:40:43 [sheets]
- Tim: high priority phone call
- 18:40:55 [eikeon]
- eikeon (n=eikeon@dsl092-168-155.wdc2.dsl.speakeasy.net) has joined #dig
- 18:40:57 [jamesdh]
- sheets: snappy scribe commentary
- 18:41:26 [sheets]
- Tim: my car is [missed]
- 18:41:56 [sheets]
- Tim: #swig is a good place to discuss interesting web tech
- 18:42:23 [sheets]
- Tim: URIs get blogged
- 18:42:58 [sheets]
- Tim: annotations about URIs are helpful
- 18:43:26 [sheets]
- DanC: regexes are used to chomp URIs
- 18:43:38 [alerer]
- alerer (i=alerer@30-6-168.wireless.csail.mit.edu) has joined #dig
- 18:43:52 [DanC]
- http://swig.xmlhack.com/
- 18:44:02 [DanC]
- http://swig.xmlhack.com/2006/06/21/2006-06-21.html#1150909926.146386
- 18:44:15 [sheets]
- Tim: parser walkthru?
- 18:45:14 [sheets]
- http://dig.csail.mit.edu/2005/ajar/ajaw/rdf/rdfparser.js
- 18:46:59 [DanC]
- why is store both a param and an instance variable?
- 18:52:06 [DanC]
- dom.getAttributeNodeNS(RDFParser.ns.RDF,"about") <- oh, the DOM API is ugly!
- 18:54:02 [DanC]
- connolly@dirk:~/dig-svn/2005/ajar/ajaw$ firefox test/jsunit/testRunner.html
- 18:55:40 [DanC]
- ~/dig-svn/2005/ajar/ajaw$ ls test/rdf-tests/horst*
- 18:55:41 [DanC]
- test001.rdf test002.rdf test003.rdf test004.rdf
- 18:58:32 [DanC]
- walking thru jsunit stuff
- 18:59:03 [sheets]
- http://www.w3.org/2000/10/rdf-tests/rdfcore/
- 19:00:47 [DanC]
- and open $DIG/2005/ajar/ajaw/test/jsunit/testRunner.html
- 19:01:05 [DanC]
- and in there, browse to $DIG/2005/ajar/ajaw/test/rdf/rdfparser.test.html
- 19:02:16 [DanC]
- for me, that's /home/connolly/dig-svn/2005/ajar/ajaw/test/rdf/rdfparser.test.html
- 19:03:02 [DanC]
- I get Runs: 21 Errors: 0 Failures: 2
- 19:07:12 [DanC]
- FYI, what we just went over is in mail that I sent:
- 19:07:12 [DanC]
- From: Dan Connolly <connolly@w3.org>
- 19:07:12 [DanC]
- To: Tabulator team <tabulator@csail.mit.edu>
- 19:07:12 [DanC]
- Subject: RDF/XML parser, jsunit progress
- 19:07:13 [DanC]
- Date: Mon, 19 Jun 2006 22:18:05 -0500
- 19:07:33 [DanC]
- and
- 19:07:34 [DanC]
- From: David Sheets <dsheets@mit.edu>
- 19:07:34 [DanC]
- To: tabulator@csail.mit.edu
- 19:07:34 [DanC]
- Subject: [Tabulator] Parser Update
- 19:07:34 [DanC]
- Date: Wed, 21 Jun 2006 03:07:32 -0400 (02:07 CDT)
- 19:11:30 [DanC]
- agenda = blanknodes
- 19:12:03 [lkagal]
- lkagal has quit ()
- 19:16:19 [DanC]
- "Subject: [Tabulator] Proposal for joint SwapMe-Tabulator-TAMI meeting 17 July"
- 19:17:37 [DanC]
- PROPOSED: to have another tabulator meeting 6 July 1pET
- 19:18:00 [DanC]
- so RESOLVED
- 19:18:15 [timbl]
- ____________________________________ Adjouned
- 19:19:56 [DanC]
- no, I think the IRC log suffices...
- 19:20:05 [DanC]
- ... I'll be happy if sheets mails a copy to tabulator-team
- 19:20:11 [sheets]
- will do
- 19:21:43 [djweitzner]
- djweitzner has quit ()
- 19:23:33 [erim]
- erim has quit (Read error: 110 (Connection timed out))
- 19:25:59 [djweitzner]
- djweitzner (n=djweitzn@30-7-146.wireless.csail.mit.edu) has joined #dig
- 19:26:11 [hmslydia]
- hmslydia has quit ()
- 19:39:34 [lkagal]
- lkagal (n=lkagal1@30-5-242.wireless.csail.mit.edu) has joined #dig
- 19:46:06 [evoli]
- evoli has quit (Remote closed the connection)
- 19:50:19 [sheets]
- sheets has quit (Read error: 110 (Connection timed out))
- 19:55:52 [evoli]
- evoli (n=JoyceC@30-7-248.wireless.csail.mit.edu) has joined #dig
- 20:06:31 [Yuhsin]
- Yuhsin (n=JoyceC@30-7-248.wireless.csail.mit.edu) has joined #dig
- 20:06:53 [evoli]
- evoli has quit (Nick collision from services.)
- 20:07:17 [Yuhsin]
- Yuhsin is now known as evoli
- 20:10:42 [sheets]
- sheets (n=dsheets@SIMMONS-TWO-TWELVE.MIT.EDU) has joined #dig
- 20:17:33 [claudiof]
- claudiof has quit ()
- 20:30:54 [alerer]
- alerer has quit (Read error: 104 (Connection reset by peer))
- 20:44:09 [RalphS]
- RalphS has quit (Read error: 110 (Connection timed out))
- 21:07:20 [yosi_s]
- yosi_s has quit ("Chatzilla 0.9.72 [Firefox 1.5.0.4/2006060216]")
- 21:11:32 [timbl]
- timbl has quit ()
- 21:24:14 [lkagal]
- Danny ?
- 21:33:34 [djweitzner]
- djweitzner has quit ()
- 21:34:26 [djweitzner]
- djweitzner (n=djweitzn@30-7-146.wireless.csail.mit.edu) has joined #dig
- 21:44:38 [eikeon]
- eikeon has quit (Read error: 110 (Connection timed out))
- 22:09:00 [djweitzner]
- djweitzner has quit ()
- 22:14:31 [djweitzner]
- djweitzner (n=djweitzn@30-7-146.wireless.csail.mit.edu) has joined #dig
- 22:24:06 [eikeon]
- eikeon (n=eikeon@dsl092-168-155.wdc2.dsl.speakeasy.net) has joined #dig
- 22:31:32 [djweitzner]
- djweitzner has quit ()
- 22:35:16 [ryanlee]
- ryanlee has quit ()
- 23:02:05 [tlr]
- tlr has quit ("leaving")
- 23:22:58 [evoli]
- evoli has quit ("Leaving")
- 23:37:23 [lkagal]
- lkagal has quit ()