IRC log of dig on 2013-01-28

Timestamps are in UTC.

00:35:29 [timbl]
timbl (~timbl@119.225.8.250) has joined #dig
05:58:12 [rszeno]
rszeno has quit (Quit: Leaving.)
06:08:28 [danbri_]
danbri_ has quit (Ping timeout: 276 seconds)
06:12:07 [danbri]
danbri (~danbri@173-228-68-60.dsl.dynamics.napanet.net) has joined #dig
08:10:36 [jmvanel]
jmvanel (~jmvanel@110.0.88.79.rev.sfr.net) has joined #dig
08:21:57 [deiu]
deiu (~andrei@157.159.103.72) has joined #dig
08:21:57 [deiu]
deiu has quit (Changing host)
08:21:57 [deiu]
deiu (~andrei@unaffiliated/deiu) has joined #dig
08:22:19 [melvster]
melvster (~melvin@ip-94-112-34-211.net.upcbroadband.cz) has joined #dig
08:30:32 [jmvanel]
jmvanel has quit (Ping timeout: 255 seconds)
08:42:35 [melvster]
melvster has quit (Ping timeout: 260 seconds)
08:50:55 [melvster]
melvster (~melvin@ip-94-112-34-93.net.upcbroadband.cz) has joined #dig
08:53:08 [danbri]
danbri has quit (Remote host closed the connection)
08:56:11 [trueg_away]
trueg_away is now known as trueg
10:04:10 [timbl]
timbl has quit (Quit: timbl)
10:04:49 [bblfish_]
bblfish_ has quit (Remote host closed the connection)
10:28:49 [jmvanel]
jmvanel (~jmvanel@APuteaux-551-1-29-23.w90-62.abo.wanadoo.fr) has joined #dig
10:50:42 [trueg]
trueg is now known as trueg_away
11:29:40 [timbl_]
timbl_ (~timbl@119.225.8.250) has joined #dig
11:41:58 [timbl_]
timbl_ has quit (Quit: timbl_)
12:07:15 [scor]
scor (~scor@c-98-216-39-127.hsd1.ma.comcast.net) has joined #dig
12:07:16 [scor]
scor has quit (Changing host)
12:07:16 [scor]
scor (~scor@drupal.org/user/52142/view) has joined #dig
12:18:15 [Ralph_]
Ralph_ (RSwick@w3cvpn1.w3.org) has joined #dig
12:18:23 [Ralph_]
Ralph_ is now known as RalphS
12:19:32 [DIGlogger]
DIGlogger (~dig-logge@groups.csail.mit.edu) has joined #dig
12:19:32 [wright.freenode.net]
topic is: Decentralized Information Group @ MIT http://dig.csail.mit.edu/
12:19:32 [wright.freenode.net]
Users on #dig: DIGlogger RalphS scor jmvanel melvster deiu tyteen4a03 cheater_ manu1 mattl bergi_ Yudai___ trueg_away manu-db presbrey betehess ericP sandro dsheets
12:35:14 [bblfish]
bblfish (~bblfish@did75-4-82-66-46-164.fbx.proxad.net) has joined #dig
12:38:20 [trueg_away]
trueg_away is now known as trueg
12:44:12 [danbri]
danbri (~danbri@173-228-68-60.dsl.dynamics.napanet.net) has joined #dig
13:39:50 [bblfish]
bblfish has quit (Remote host closed the connection)
13:41:34 [bblfish]
bblfish (~bblfish@did75-4-82-66-46-164.fbx.proxad.net) has joined #dig
14:23:22 [melvster]
mattl: im thinking of releasing some software that i want to be compatible with GPL but also want it to be pretty liberal in letting it spread ... do you think LGPL would be a good choice?
14:24:04 [mattl]
"LGPL v3 or later", yeah
14:24:10 [melvster]
thanks!
14:24:28 [melvster]
any other choices like BSD or MIT?
14:24:39 [mattl]
if you had a second choice, Apache 2.0 license.
14:24:51 [melvster]
perfect, thanks for the info
14:25:19 [mattl]
melvster: http://www.gnu.org/licenses/license-recommendations.html
14:25:51 [melvster]
hah even better (but i should have known that page already) thanks! :)
14:27:47 [scor]
scor has quit (Quit: scor)
14:46:39 [bblfish]
bblfish has quit (Remote host closed the connection)
14:57:28 [rszeno]
rszeno (~rszeno@79.114.107.98) has joined #dig
14:58:23 [bblfish]
bblfish (~bblfish@did75-14-82-236-19-233.fbx.proxad.net) has joined #dig
15:07:26 [scor]
scor (~scor@132.183.243.108) has joined #dig
15:07:26 [scor]
scor has quit (Changing host)
15:07:26 [scor]
scor (~scor@drupal.org/user/52142/view) has joined #dig
15:57:35 [bblfish]
bblfish has quit (Remote host closed the connection)
16:08:39 [bblfish]
bblfish (~bblfish@212-198-252-43.rev.numericable.fr) has joined #dig
16:23:34 [danbri]
danbri has quit (Ping timeout: 256 seconds)
16:24:46 [danbri]
danbri (~danbri@173-228-68-60.dsl.dynamics.napanet.net) has joined #dig
16:46:23 [deiu]
deiu has quit (Quit: Leaving)
17:02:05 [danbri]
danbri has quit (Remote host closed the connection)
17:12:09 [GitHub153]
01[13data.fm01] 15deiu pushed 1 new commit to 06dev: 02http://git.io/fTFHVg
17:12:09 [GitHub153]
13data.fm/06dev 142486974 15deiu: Fixed a small bug related to trailing slashes.
17:12:29 [deiu]
deiu (~andrei@157.159.103.72) has joined #dig
17:12:29 [deiu]
deiu has quit (Changing host)
17:12:29 [deiu]
deiu (~andrei@unaffiliated/deiu) has joined #dig
17:13:16 [deiu]
presbrey, do you have anything against merging dev into master for data.fm?
17:15:22 [presbrey]
dont think so, taking a look..
17:15:36 [GitHub130]
01[13data.fm01] 15deiu pushed 1 new commit to 06dev: 02http://git.io/ewQEzA
17:15:36 [GitHub130]
13data.fm/06dev 149f0e915 15deiu: Small fix for length of empty list of WebIDs.
17:15:46 [deiu]
I've deployed it on my-profile.eu for some real tests
17:15:57 [deiu]
found some small bugs which I've fixed
17:16:34 [presbrey]
have you seen the updatesVia branch?
17:16:50 [presbrey]
I would like to merge that with dev+master too but it may require more work
17:16:51 [deiu]
nope
17:17:00 [deiu]
what's it about?
17:17:09 [presbrey]
it powers the Updates-Via header / WebSocket
17:17:23 [Pipian-Work]
Pipian-Work (~Pipian@2001:470:8b2d:7d8:c48d:37de:4f12:aa13) has joined #dig
17:17:23 [deiu]
ah right
17:17:27 [presbrey]
for real-time sparql updates
17:17:39 [deiu]
you might have told me some time ago
17:18:26 [presbrey]
we should start it as an option defaulted to off until stable
17:18:47 [deiu]
sure
17:19:06 [deiu]
hmm, I think it would be easier to merge dev into master for starters
17:19:42 [deiu]
at least for now since it's pretty much stable
17:19:49 [presbrey]
yes just looking at the branch list for low hanging
17:20:17 [deiu]
I'm working on ACL / dir, and I'd like to be able to push commits to dev :-)
17:21:00 [deiu]
I'm starting to see real issues with the php redland bindings :-(
17:21:28 [presbrey]
more segfaults?
17:21:29 [bblfish]
bblfish has quit (Remote host closed the connection)
17:21:46 [deiu]
not segfaults but huge memory leaks
17:22:08 [presbrey]
what py api would you want to use?
17:22:22 [deiu]
you mean framework?
17:22:40 [deiu]
I guess either flask or bottle (similar)
17:22:43 [presbrey]
graph/storage "adapters" ?
17:22:47 [deiu]
with rdflib
17:23:02 [presbrey]
ok
17:23:07 [deiu]
https://rdflib.readthedocs.org/en/latest/
17:23:27 [deiu]
it seems well maintained
17:23:58 [presbrey]
is that one pure python?
17:24:12 [rszeno]
static vs. dynamic linking?
17:24:43 [deiu]
yes it is
17:24:48 [presbrey]
if so it might be slower
17:25:11 [deiu]
worst case scenario, we can use the redland py api
17:25:26 [RalphS]
RalphS has quit (Ping timeout: 246 seconds)
17:25:30 [presbrey]
same leak problem there though?
17:25:36 [deiu]
no idea
17:25:38 [RalphS]
RalphS (RSwick@30-7-118.wireless.csail.mit.edu) has joined #dig
17:25:41 [presbrey]
perhaps it would be good to fix the leaks?
17:25:59 [presbrey]
or the leak may be in our wrapper?
17:26:19 [deiu]
may be
17:26:29 [deiu]
I'm not 100% sure what the cause is
17:26:56 [presbrey]
if we move to new apis I would like to see the serializer buffering improve
17:27:23 [presbrey]
when you have a big graph I think our current integration has to serialize the full output before it can start sending to the client
17:27:24 [deiu]
maybe we can benchmark them
17:27:46 [deiu]
I think most apis behave like that
17:27:50 [presbrey]
so if rdflib serializer uses generators => yield '', '', '', :-D
17:28:16 [deiu]
you could ask this in #rdflib btw
17:28:28 [presbrey]
most python ones do, but probably not the redland py api
17:29:41 [presbrey]
yep or we could test it out and see
17:29:53 [deiu]
got a big data graph?
17:29:58 [presbrey]
but I should go back to looking at the dev/master merge
17:30:02 [deiu]
ah yes
17:30:05 [deiu]
please :-)
17:30:09 [presbrey]
try serializing an empty graph
17:30:21 [presbrey]
and yes we have big graphs on data.fm
17:30:38 [deiu]
shit, my machine just became very unstable
17:30:47 [presbrey]
http://test.data.fm/triples/
17:30:49 [deiu]
lots of disk activity
17:30:52 [timbl_]
timbl_ (~timbl@119.225.8.250) has joined #dig
17:30:52 [timbl_]
timbl_ has quit (Client Quit)
17:30:52 [presbrey]
1K, 5K, 10K
17:31:07 [deiu]
nice!
17:32:10 [presbrey]
I would agree on targeting flask
17:32:47 [deiu]
I'll try to work on it
17:35:39 [deiu]
and this time we need to write proper docs ;-)
17:40:02 [trueg]
trueg is now known as trueg_away
17:48:59 [GitHub72]
01[13data.fm01] 15presbrey pushed 27 new commits to 06master: 02http://git.io/cxmn8Q
17:48:59 [GitHub72]
13data.fm/06master 149d54381 15deiu: Fixed the WAC rule names when generating new WebIDs.
17:48:59 [GitHub72]
13data.fm/06master 1422dff48 15deiu: Python script to generate WebID certs. Replaces the CA.
17:48:59 [GitHub72]
13data.fm/06master 1419c06e2 15deiu: Minimal interface for managing WAC rules.
17:49:07 [deiu]
deiu has quit (Remote host closed the connection)
17:49:42 [deiu]
deiu (~andrei@157.159.103.72) has joined #dig
17:49:42 [deiu]
deiu has quit (Changing host)
17:49:42 [deiu]
deiu (~andrei@unaffiliated/deiu) has joined #dig
17:49:53 [deiu]
wow, this time the whole computer crashed
17:50:29 [GitHub169]
01[13data.fm01] 15presbrey pushed 1 new commit to 06dev: 02http://git.io/GVHYBw
17:50:29 [GitHub169]
13data.fm/06dev 14b761821 15presbrey: Merge remote branch 'origin/master' into dev
17:53:19 [deiu]
brb
17:53:23 [deiu]
deiu has quit (Client Quit)
17:57:56 [GitHub171]
01[13data.fm01] 15presbrey 04force-pushed 06updatesVia from 140e47d6f to 146157d4c: 02http://git.io/NKbJlw
17:57:56 [GitHub171]
13data.fm/06updatesVia 146157d4c 15presbrey: updatesVia: add sparql query broadcast
17:58:18 [GitHub193]
01[13data.fm01] 15presbrey 04deleted 06dev at 14b761821: 02http://git.io/A7j7ag
18:29:28 [jmvanel]
jmvanel has quit (Ping timeout: 276 seconds)
18:38:09 [GitHub167]
01[13data.fm01] 15presbrey pushed 1 new commit to 06master: 02http://git.io/6pg0MA
18:38:09 [GitHub167]
13data.fm/06master 14252ea8d 15presbrey: [WAC] bugfix checks for trailing slash at root
18:38:43 [GitHub130]
01[13data.fm01] 15presbrey 04force-pushed 06updatesVia from 146157d4c to 1452845d6: 02http://git.io/NKbJlw
18:38:43 [GitHub130]
13data.fm/06updatesVia 1452845d6 15presbrey: updatesVia: add sparql query broadcast
18:44:10 [deiu]
deiu (~andrei@2a01:e35:8b67:4160:8008:a8db:a271:6381) has joined #dig
18:44:10 [deiu]
deiu has quit (Changing host)
18:44:10 [deiu]
deiu (~andrei@unaffiliated/deiu) has joined #dig
18:47:34 [presbrey]
welcome back
18:48:02 [presbrey]
slight bug with ACLs at the root domain
18:50:14 [presbrey]
check master for the fix
18:50:16 [deiu]
:S
18:50:40 [deiu]
hmm
18:51:09 [presbrey]
I thought I had stripped the trailing slash for root domains earlier in runtime
18:51:26 [presbrey]
so there may be a more complete fix hiding
18:51:47 [deiu]
I had some issues with that
18:51:54 [presbrey]
this is needed because eg.
18:51:58 [presbrey]
joe$ dirname http://test.data.fm/test
18:51:58 [presbrey]
http://test.data.fm
18:52:55 [presbrey]
hopefully we did not append a slash hard-coded somewhere?
18:53:33 [presbrey]
sry I did not check the dev commits too closely :\
19:02:47 [deiu]
no worries, I'll go over it one more time
19:07:13 [timbl]
timbl (~timbl@119.225.8.250) has joined #dig
19:08:15 [presbrey]
what we need more than docs is tests
19:08:28 [presbrey]
so next time WAC is changed we verify no breakage
19:09:36 [trueg_away]
trueg_away is now known as trueg
19:31:42 [deiu]
do you have a preferred way of implementing tests?
19:35:07 [deiu]
brb
19:35:09 [deiu]
deiu has quit (Quit: Leaving)
19:36:21 [deiu]
deiu (~andrei@2a01:e35:8b67:4160:f0b6:93b8:7127:80a5) has joined #dig
19:36:21 [deiu]
deiu has quit (Changing host)
19:36:21 [deiu]
deiu (~andrei@unaffiliated/deiu) has joined #dig
19:38:07 [presbrey]
been using nodeunit with rdflib.js
19:49:10 [jmvanel]
jmvanel (~jmvanel@110.0.88.79.rev.sfr.net) has joined #dig
20:02:40 [melvster]
wow did somebody say realtime updates via ...
20:04:18 [melvster]
so updates-via is the address of a websocket server?
20:06:42 [trueg]
trueg has quit (Remote host closed the connection)
20:09:27 [trueg]
trueg (~trueg@HSI-KBW-46-237-238-110.hsi.kabel-badenwuerttemberg.de) has joined #dig
20:11:08 [presbrey]
yes
20:11:18 [presbrey]
demo should be live on data.fm
20:11:47 [presbrey]
if the new ACL stuff didnt break it :)
20:13:06 [melvster]
amazing ...
20:14:02 [melvster]
ooh the locks are there!
20:15:12 [melvster]
Updates-Via: ws://test.data.fm:81
20:15:33 [melvster]
amazing
20:15:41 [melvster]
so how do you serialize updates?
20:17:09 [presbrey]
updates are serialized to: sparql + old etag + new etag
20:17:20 [melvster]
ah cool
20:18:30 [melvster]
so in the js you make a connection to it, then when the sparql updates come in, send them to rdflib, update the kg, and trigger a re render as necessary?
20:18:32 [presbrey]
so if you worried you have become out of sync, get the latest If-Matches: (new-etag), and you can skip forward up to the sparql stream
20:18:58 [melvster]
kind of epic
20:19:40 [melvster]
any kind of demo working at all, or exercise for the reader? :)
20:20:50 [presbrey]
tabulator has basic integration, proof of concept only
20:21:11 [presbrey]
if you open data.fm graphs in tabulator in separate browsers, updates on one should trigger the other to refresh
20:21:31 [presbrey]
the granularity of update processing on the tabulator UI is very low for various reasons
20:21:59 [melvster]
Fantastic!
20:21:59 [presbrey]
mostly weak graph/kb => dom api
20:22:04 [presbrey]
so far :)
20:22:21 [melvster]
presbrey: you've just made my month ... actually maybe my whole year!
20:22:38 [melvster]
didnt dream this would be possible so soon
20:25:44 [presbrey]
ha thx are you still running data.fm?
20:25:56 [melvster]
sure ... in lots of places
20:26:31 [presbrey]
cool try running the updatesVia branch
20:26:42 [presbrey]
it should have all the pieces you need :)
20:26:47 [melvster]
sweet
20:29:42 [presbrey]
you do have to install gunicorn if you havent already
20:29:53 [deiu]
presbrey, are you sure about that last fix for WAC?
20:30:24 [presbrey]
yes deiu, http://test.data.fm/ gives 40x without it
20:30:26 [deiu]
did you actually test it, or was it just something that you thought should be there
20:30:32 [deiu]
ah
20:30:48 [deiu]
that's because you're lacking some trailing slashes for dirs
20:31:06 [deiu]
we need to take slashes into account
20:31:41 [deiu]
the WAC editor should place trailing slashes at the end of dirs
20:32:19 [presbrey]
trailing slashes on dirs?
20:32:26 [presbrey]
I dont put those in my ACLs, do you?
20:32:39 [deiu]
yes, of course
20:32:56 [deiu]
how do make the difference between a dir and a resource?
20:33:45 [deiu]
it's even easier when reading the ACLs
20:34:33 [tyteen4a03]
tyteen4a03 has quit (Read error: Connection reset by peer)
20:34:57 [tyteen4a03]
tyteen4a03 (T4@n218250228096.netvigator.com) has joined #dig
20:37:00 [presbrey]
interesting
20:37:07 [presbrey]
I dont see trailing slashes in anyones ACLs on data.fm
20:37:39 [deiu]
because the editor didn't put them before
20:37:56 [deiu]
and it didn't work without removing them :-)
20:38:00 [presbrey]
ok I guess we can rewrite them
20:38:04 [deiu]
(the lines you added)
20:38:14 [deiu]
how many users are on data.fm?
20:38:47 [presbrey]
good question
20:39:34 [presbrey]
certainly more app users than domain admins
20:39:51 [presbrey]
I think melvster got a bunch of webids to a vote app
20:40:16 [melvster]
oh yeah for SOPA
20:40:49 [presbrey]
what does unix openafs nfs ... do? there must be some good precedence on this?
20:40:57 [melvster]
actually the vast majority didnt use WebID
20:41:06 [melvster]
most were google oauth
20:41:14 [melvster]
and lots of dns:
20:41:27 [deiu]
are there any real users of data.fm?
20:41:49 [melvster]
i just use it for testing currently
20:42:04 [presbrey]
not really sure, no logs :)
20:42:10 [presbrey]
we could take it down and see if anyone complains?
20:43:10 [melvster]
im guessing it would be OK
20:43:23 [deiu]
you don't need to take it down
20:43:26 [melvster]
it's been down a few times here and there ... i think im the only person that notices ...
20:43:35 [deiu]
just to revert the latest fix to WAC
20:43:54 [deiu]
and start using trailing slashes for dirs
20:44:13 [deiu]
if you're using the built-in WAC editor, you're safe already
20:44:17 [scor]
scor has quit (Quit: scor)
20:44:33 [melvster]
the wac editor is pretty cool i must say
20:44:54 [deiu]
it's lacking some features
20:45:05 [presbrey]
how do you see rules for root / ?
20:45:13 [deiu]
namely using groups (classes), append, control
20:45:44 [deiu]
I don't think you can, with the editor
20:46:25 [deiu]
there's no root / you can click on
20:46:48 [deiu]
the rules for root / are added by default if you don't already have a .meta file
20:46:56 [presbrey]
you should be able to get to them somewhere
20:47:27 [presbrey]
we cant just revert, there is still more fix needed if we want ACLs to append slashes for dirs
20:48:02 [deiu]
what fixes?
20:48:13 [presbrey]
when we are checking for parent path rules
20:48:23 [presbrey]
we have to append slash if the request is for a dir
20:48:38 [deiu]
let me see, I think it works
20:48:58 [presbrey]
eg. GET http://test/1/2/3/4/
20:49:04 [presbrey]
current code will check ACLs for:
20:49:18 [presbrey]
http://test/1/2/3/4/,">http://test/1/2/3/4/, http://test/1/2/3, http://test/1/2, http://test/1, http://test
20:50:14 [presbrey]
so if you write a rule for http://test/1/2/, sadly it will not trigger for http://test/1/2/3/4/
20:50:21 [presbrey]
the loop is closed by $p=dirname($p)
20:50:36 [presbrey]
(unix) dirname returns without trailing slash
20:51:20 [presbrey]
we can reappend slash or we can write a new dirname
20:51:48 [presbrey]
I think the former is more explicit and probably better
20:52:04 [deiu]
actually, why would that be a problem?
20:52:31 [deiu]
the rule for http://test/1/2/3/4/">http://test/1/2/3/4/ would just overwrite the one for http://test/1/2/
20:52:47 [deiu]
same as you have /home/ owned by root, and /home/user/ owned by user
20:52:50 [presbrey]
what about with no rule /1/2/3/4/ ?
20:53:21 [deiu]
it works if you have a rule for /1/2/3/4/
20:53:27 [deiu]
just tested
20:53:50 [presbrey]
yes this bug only applies to checking parent paths for acls
20:54:52 [presbrey]
put a defaultForNew+read/write rule at /1/2/
20:55:16 [bblfish]
bblfish (~bblfish@AAubervilliers-652-1-93-24.w83-114.abo.wanadoo.fr) has joined #dig
20:55:22 [deiu]
sec
20:58:04 [deiu]
it works
20:58:21 [presbrey]
did you remove all other rules?
20:58:40 [deiu]
I own /1/2/ with a defaultfornew, and the test user can see /1/2/3/
20:59:45 [presbrey]
do you have defaultForNew / ?
21:05:48 [deiu]
hmm
21:06:08 [presbrey]
here is the example .meta:
21:06:11 [presbrey]
[ acl:accessTo </>, </1/>, </.meta> ; acl:defaultForNew </1/> ; acl:agentClass foaf:Agent ; acl:mode acl:Read ]
21:06:23 [presbrey]
@ http://test.dev.data.fm/
21:06:58 [presbrey]
the acl:dFN </1/> should also give you access to </1/2/>
21:07:14 [trueg]
trueg is now known as trueg_away
21:07:23 [presbrey]
but does not, eg. visit http://test.dev.data.fm/1/2/
21:07:54 [deiu]
yeah
21:08:37 [presbrey]
if you add `echo $q` to WAC->can, you can see its again due to the trailing slash discrepancy
21:10:26 [presbrey]
I think its better to say never have slashes, not in ACL, not in verifiers.
21:11:18 [presbrey]
if its "maybe sometimes", then the good people of "lenient in what I accept" will surely check both always :)
21:12:15 [deiu]
yeah
21:12:21 [deiu]
I'll look into it :-)
21:13:47 [deiu]
I think I might have found another bug
21:16:53 [bblfish]
bblfish has quit (Remote host closed the connection)
21:21:31 [deiu]
ok, I've fixed
21:21:34 [deiu]
it*
21:21:34 [GitHub83]
01[13data.fm01] 15deiu pushed 1 new commit to 06master: 02http://git.io/6UpsXw
21:21:35 [GitHub83]
13data.fm/06master 14b62260d 15deiu: Added an extra check for the length of the WebID when setting ACLs.
21:22:00 [presbrey]
cool thx
21:22:22 [deiu]
hmm
21:22:27 [deiu]
I should revert the slashes then
21:22:58 [presbrey]
I like the 'Actions' :)
21:23:10 [presbrey]
revert slashes?
21:23:28 [deiu]
yeah, when generating the default .meta
21:23:40 [deiu]
I'm adding a slash after the domain name
21:24:15 [deiu]
<http://www.w3.org/ns/auth/acl#accessTo> <http://test.my-profile.eu/>
21:24:27 [presbrey]
yes
21:25:06 [presbrey]
thats not in webidgen, is that JS?
21:26:25 [presbrey]
yeah common.js nice :D
21:26:42 [deiu]
yes
21:27:36 [deiu]
done
21:27:47 [GitHub34]
01[13data.fm01] 15deiu pushed 1 new commit to 06master: 02http://git.io/IZK4kg
21:27:47 [GitHub34]
13data.fm/06master 14101c9de 15deiu: Reverted to no trailing slashes. Default #owner rule also applies to the .meta document.
21:28:14 [presbrey]
do you have any ideas on how you would like to register w/ $rdf.graph to get updates/callback for data changes?
21:28:29 [presbrey]
the updatesVia is probably already sync'ing the .meta graph
21:28:38 [deiu]
not sure
21:28:47 [deiu]
I've been thinking of a global notification system
21:29:07 [deiu]
similar to webbox, where you append triples to a resource
21:29:11 [presbrey]
ideally we could share event registration scheme with WAC editor + tabulator + vote.data.fm + etc..
21:29:20 [bblfish_]
bblfish_ (~bblfish@AAubervilliers-651-1-188-243.w83-200.abo.wanadoo.fr) has joined #dig
21:30:14 [deiu]
btw, append should only work for POST
21:30:32 [deiu]
we're not testing for it right now in other methods
21:30:56 [presbrey]
I think thats correct
21:31:12 [presbrey]
but only POST text/turtle
21:31:14 [deiu]
PUT implies a hidden DELETE
21:31:17 [deiu]
yes
21:31:29 [presbrey]
or other graph types
21:31:33 [presbrey]
no query types
21:32:40 [presbrey]
since query could also be a hidden DELETE
21:32:52 [deiu]
of course
21:35:44 [deiu]
btw, I'd like to rename .meta to .acl
21:36:14 [deiu]
to re-purpose .meta for LDP
21:36:50 [deiu]
it makes sense to also change to <Link rel=acl> at this point
21:38:43 [presbrey]
the URI is server-specific right?
21:39:02 [deiu]
user-specific even
21:39:11 [presbrey]
someone could use .htmeta or .htacl as long as they put Link header pointing to it
21:39:46 [presbrey]
I think consumers are all following their nose already
21:40:20 [presbrey]
but they are looking for rel=meta, so that would need to be changed if there we need to separate rel=meta from rel=acl into 2 Link headers?
21:40:22 [deiu]
Link:<http://user.example.com/.meta>; rel=acl
21:40:44 [deiu]
but are they looking for ACLs in the .meta?
21:40:49 [deiu]
or just any meta data
21:41:23 [presbrey]
you have to look wherever they say
21:41:26 [presbrey]
it might even be
21:41:40 [presbrey]
Link:<http://user.example.com/drupal/acl-generator.php>; rel=acl
21:42:03 [deiu]
yeah, I don't care about the actual link
21:42:13 [deiu]
I care about the rel= value
21:42:55 [deiu]
the ACL resource will be dereferenced anyway, however the link looks like
21:43:01 [presbrey]
ok
21:43:17 [presbrey]
I did not pick the rel=, maybe check with timbl or melvster?
21:43:52 [scor]
scor (~scor@drupal.org/user/52142/view) has joined #dig
21:43:55 [deiu]
I think we need to check if rel= values are standardized or not
21:44:08 [presbrey]
I don't have a strong preference, its only coded a few places
21:44:14 [deiu]
I think there's an rfc about them
21:44:55 [deiu]
http://tools.ietf.org/html/rfc5988
21:45:45 [deiu]
it seems that we'll have to register the new rel=acl value
21:46:45 [melvster]
http://www.iana.org/assignments/link-relations/link-relations.xml
21:47:36 [deiu]
weird, there's no rel=meta there
21:48:05 [deiu]
might as well register acl and ldp
21:48:54 [presbrey]
what does rel=ldp contain?
21:50:20 [presbrey]
just metadata for containers?
21:50:36 [presbrey]
contains metadata for containers, wow :)
21:50:48 [deiu]
yes
21:51:05 [deiu]
which also contain metadata for resources
21:51:37 [rszeno]
http://support.google.com/webmasters/bin/answer.py?hl=en&answer=139394
21:51:39 [presbrey]
why do acls not fall within: 'metadata for resources'?
21:53:52 [rszeno]
being interpreted by google this way doesn't end with a problem for web develpers?
21:54:17 [presbrey]
rszeno, I think that refers to a different rel=
21:54:18 [rszeno]
with rdfa is already a problem
21:54:40 [rszeno]
is usualy used with link in the head
21:55:55 [rszeno]
except canonical is also nofolow
21:56:13 [rszeno]
probably few others
22:24:18 [rszeno]
presbrey, you are right. What i'm worry about is combining markup for different readers and avoid confusions, i'm forced to follow google rules for seo, i also need to combine schema.org with rdf to provide linked data.
22:26:05 [rszeno]
is realy annoying when i look to how different user agents see same page
22:27:34 [deiu]
presbrey, how do you delete a triple with rdflib.js?
22:28:17 [deiu]
graph.remove() but what exactly?
22:32:30 [deiu]
or a whole rule (#rule)
22:35:22 [presbrey]
hm let me check the docs... ;)
22:39:06 [presbrey]
apparently it takes a:
22:39:20 [presbrey]
{ subject: '', predicate: '', object: '', why '' }
22:40:30 [melvster]
melvster has quit (Quit: Leaving.)
22:41:10 [deiu]
hmm, why doesn't it take g.sym()?
22:41:14 [deiu]
not really consistent
22:42:42 [deiu]
it doesn't work
22:42:49 [timbl]
timbl has quit (Quit: timbl)
22:43:07 [deiu]
I've tried g.remove({ subject: g.sym('a'), predicate: g.sym('b'), object: g.sym('c'), why: ''}) and g.remove({ subject: 'a', predicate: 'b', object: 'c', why: ''})
22:45:57 [deiu]
it's definitely with g.sym()
22:46:07 [deiu]
Array did not contain [object Object]
22:46:09 [deiu]
but it does
22:52:49 [scor]
scor has quit (Quit: scor)
22:53:49 [deiu]
oh, so I have to provide all 4 (including why:)
22:53:56 [scor]
scor (~scor@c-98-216-39-127.hsd1.ma.comcast.net) has joined #dig
22:53:57 [scor]
scor has quit (Changing host)
22:53:57 [scor]
scor (~scor@drupal.org/user/52142/view) has joined #dig
22:54:39 [cheater]
cheater (~cheater@p4FD0E4EB.dip.t-dialin.net) has joined #dig
22:56:03 [deiu]
I'll finish this tomorrow
22:56:10 [deiu]
o/
22:56:17 [deiu]
deiu has quit (Quit: Leaving)
22:57:45 [cheater_]
cheater_ has quit (Ping timeout: 260 seconds)
23:01:17 [scor]
scor has quit (Quit: scor)
23:01:17 [RalphS]
RalphS has quit ()
23:04:41 [scor]
scor (~scor@drupal.org/user/52142/view) has joined #dig
23:08:46 [melvster]
melvster (~melvin@ip-94-112-34-211.net.upcbroadband.cz) has joined #dig
23:13:26 [melvster]
melvster has quit (Ping timeout: 255 seconds)
23:33:13 [jmvanel]
jmvanel has quit (Ping timeout: 240 seconds)