IRC log of dig on 2013-08-10
Timestamps are in UTC.
- 00:10:57 [bblfish]
- bblfish has quit (Ping timeout: 276 seconds)
- 00:53:10 [timbl]
- timbl has quit (Quit: timbl)
- 00:57:24 [scor]
- scor (~scor@drupal.org/user/52142/view) has joined #dig
- 01:03:36 [scor]
- scor has quit (Quit: scor)
- 01:23:50 [melvster]
- melvster has quit (Ping timeout: 240 seconds)
- 02:03:09 [scor]
- scor (~scor@drupal.org/user/52142/view) has joined #dig
- 02:03:10 [scor]
- scor has quit (Client Quit)
- 07:32:47 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-323-225.w83-200.abo.wanadoo.fr) has joined #dig
- 10:41:01 [cheater__]
- cheater__ (~cheater@p57AE847C.dip0.t-ipconnect.de) has joined #dig
- 11:34:12 [melvster]
- melvster (~melvster@89.176.108.70) has joined #dig
- 12:50:20 [timbl]
- timbl (~timbl@184.151.63.241) has joined #dig
- 13:15:17 [cheater__]
- cheater__ has quit (Quit: leaving)
- 13:35:49 [cheater__]
- cheater__ (~cheater@p57AE847C.dip0.t-ipconnect.de) has joined #dig
- 14:01:07 [timbl]
- timbl has quit (Quit: timbl)
- 14:05:06 [zuzak]
- zuzak has quit (Read error: Connection reset by peer)
- 16:10:30 [zuzak]
- zuzak (~zu@pdpc/supporter/professional/zuzak) has joined #dig
- 16:17:38 [danbri]
- danbri (~danbri@146.90.244.38) has joined #dig
- 16:23:59 [timbl]
- timbl (~timbl@184.151.63.189) has joined #dig
- 16:38:13 [melvster]
- bblfish: re your question about rel="meta" it was michael hausenblas talking to mark nottingham about registering it, see the link on my previous post
- 16:39:10 [melvster]
- timbl: betehess suggested rel="acl", rather than, rel="meta" ... do you have any preference?
- 16:56:25 [danbri]
- is there a proposed definition for rel="meta"?
- 16:56:51 [danbri]
- we have it in http://xmlns.com/foaf/spec/#sec-autodesc because it seemed the most plausible at the time
- 17:04:08 [melvster]
- danbri: best info I have is http://lists.w3.org/Archives/Public/ietf-http-wg/2010JanMar/0244.html
- 17:04:36 [melvster]
- mh said: Use case would be ACLs as described in [2] (see
- 17:04:36 [melvster]
- section 3.2)
- 17:04:51 [melvster]
- referring to : http://dig.csail.mit.edu/2009/Papers/ISWC/rdf-access-control/paper.pdf
- 17:05:40 [bblfish]
- melvster that mail by michael is from 2010. So clearly no action was taken
- 17:06:13 [bblfish]
- so no registration process was started. Can you perhaps clarify this? Your mail makes it sound like a process is going on right now.
- 17:06:43 [danbri]
- 'meta' is pretty weak, by design
- 17:06:49 [danbri]
- 'some info probably to do with this info'
- 17:07:12 [melvster]
- bblfish: sorry if i gave that impression but michael hausenblas began the process when he was at deri, there was some follow up, but im not sure where we are now
- 17:07:34 [bblfish]
- The process has to be completed in a few months as it says in the RFC
- 17:08:39 [melvster]
- maybe that's why it didnt become official, mh isnt at deri anymore
- 17:09:08 [melvster]
- in any case that was the previous history of this, the use case is written up in the ISWC paper
- 17:09:19 [bblfish]
- danbri: yes I think that is why betehess and others thought that a more precise acl relation would be better. An ACL is meta data about a resource true, but one could argue ACLs are very important for any resource.
- 17:10:12 [melvster]
- bblfish: which do you prefer acl or meta? (or both)?
- 17:10:48 [melvster]
- i know that currently papers and some implementations use rel="meta" ... but changing this is a possibility
- 17:11:00 [bblfish]
- if you are going to go around the web fetching resources and determining access control, you want the data to be very relevant to that
- 17:11:07 [bblfish]
- so I'd say more precise here is better
- 17:11:24 [bblfish]
- ie. acl is better
- 17:11:30 [bblfish]
- rel="acl"
- 17:12:27 [bblfish]
- but presbrey's opinion here I think is important and so is Oshani's
- 17:12:40 [bblfish]
- and timbl's of course
- 17:15:14 [melvster]
- bblfish: i can see the advantages, and if someone wants to propose this, that's fine though I suggest it's important to gather consensus and/or register the link relation with IANA
- 17:15:50 [bblfish]
- that's what we are doing here :-)
- 17:16:15 [bblfish]
- trying to gather consensus. The missing parties are presbrey, and a few others....
- 17:16:50 [bblfish]
- Perhaps I'll edit the wiki later with votes from different people.
- 17:17:02 [melvster]
- im opposed to voting
- 17:17:12 [bblfish]
- that's new
- 17:17:33 [bblfish]
- you used to push for voting a lot on the WebID mailing list
- 17:17:51 [bblfish]
- anyway, it's not anyone's vote that counts. Just implementers really
- 17:18:09 [bblfish]
- and votes are just there to get some feedback.
- 17:19:28 [melvster]
- bblfish: i got a lot reading pete resnick's consensus gathering process at the IETF ... they reject dictator and voting, but work on rough consensus and running code ... it's explained nicely here .. the argument is that votes are too easily stuffed http://tools.ietf.org/id/draft-resnick-on-consensus-02.txt
- 17:19:53 [bblfish]
- that's why I keep mentioning implementors
- 17:28:16 [melvster]
- bblfish: yes I agree that straw polls can be useful
- 17:31:37 [melvster]
- things have been working well with rel="meta" so far, if you or behetess wants to make the case to change it to "acl" that's also OK, but bear in mind that rel="meta" has been in the 'wild' for 4+ years, and we might not know everyone that has used the pattern
- 17:32:29 [melvster]
- i can see the advantages
- 17:51:59 [bblfish]
- ok, I added this http://www.w3.org/wiki/WebAccessControl#rel.3Dacl_or_rel.3Dmeta.3F
- 17:54:45 [bblfish]
- anyway, it's more a question of getting a discussion going.
- 18:19:23 [melvster]
- bblfish: great, iirc rel=meta is mentioned somewhere in rdflib.js, tabulator, and rww.io ... i suspect toby's impl used that too, but that's just from vague memory
- 19:27:47 [timbl]
- timbl has quit (Quit: timbl)
- 19:36:57 [timbl]
- timbl (~timbl@184.151.63.189) has joined #dig
- 19:54:30 [timbl]
- I think rel=meta is a better choice
- 19:54:41 [timbl]
- because we have running code which picks it up
- 19:55:00 [timbl]
- and fragmenting the market does less damage.
- 19:56:32 [timbl]
- I can see also an ACL file having other stuff (like ownership, provenance, licence etc) and my gut feeling is that one flexible bad is better than trying to defined lots of ridid bags and al algorithm for what triples go in what.
- 19:56:45 [timbl]
- s/ridid/rigid/
- 19:57:04 [timbl]
- s/bad/bag/
- 20:19:21 [melvster]
- DIGlogger, pointer
- 20:19:21 [melvster]
- See http://dig.csail.mit.edu/irc/dig/2013-08-10#T20-19-21
- 20:36:38 [melvster]
- melvster has quit (Ping timeout: 240 seconds)
- 20:38:33 [melvster]
- melvster (~melvster@89.176.108.70) has joined #dig
- 20:41:18 [melvster]
- melvster has quit (Remote host closed the connection)
- 20:43:05 [melvster]
- melvster (~melvster@89.176.108.70) has joined #dig
- 20:48:20 [bblfish]
- bblfish has quit (Read error: Connection reset by peer)
- 20:48:53 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-323-225.w83-200.abo.wanadoo.fr) has joined #dig
- 21:19:35 [danbri]
- danbri has quit (Remote host closed the connection)
- 21:36:25 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 21:41:01 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-323-225.w83-200.abo.wanadoo.fr) has joined #dig
- 21:55:07 [bblfish]
- bblfish has quit (Ping timeout: 246 seconds)
- 23:39:03 [timbl]
- timbl has quit (Quit: timbl)
- 23:54:43 [timbl]
- timbl (~timbl@184.151.63.189) has joined #dig
- 23:57:23 [timbl]
- timbl has quit (Client Quit)