00:03:09 our hooks could still adapt 00:03:44 if you know how that's great 00:03:44 when 3, make sure 1 ran, then also 2, then finally do 3 00:04:00 well perhaps it's worth making a function and trying it out 00:04:18 what I have there seems to work 00:04:20 fill in the missed events at each step 00:04:50 firefox still has other probs just as dataskin !ext 00:05:07 how do I enable proxy in tabulator skin? 00:05:23 I don't know :-) 00:05:34 I just tried to get something to work. 00:05:49 amy has left #dig 00:05:53 I have the Safari/Chrome working 00:05:57 there is precedence for globals on proxy settings eg. http_proxy env 00:07:11 ok, so that's something I did not know of. Perhaps adapt the code to use http+rpxy 00:07:46 eg. //data.fm/proxy has tabulator dataskin and should hint to tabulator always use this itself, local-origin data.fm/proxy 00:08:08 your knowledge is deeper than mine very likely. 00:08:34 I'd love it if you find the better solution. 00:08:36 :-) 00:08:52 because I just want to get the social web application going 00:09:17 cool. I'm going to look at firefox brokenness next, then proxy. whats left blocking you? 00:09:38 I think I'd like to have an option so all goes through a proxy 00:10:00 my aim is to have this work on a freedombox 00:10:02 yep just keep the base_uri fixed 00:10:04 the FB will be p2p 00:10:15 browsers are messed up 00:10:25 so we have to work with what we have 00:11:10 RDF is still very helpful even when going through a proxy, because it gives a global datastructure 00:11:12 e 00:11:27 thats ok proxies can helpful to empower our poor, broken, and/or old 00:11:51 If you do that I can write a proxy tomorrow for read-write-web in scala 00:12:14 so it does not all have to go through your proxy 00:12:27 It would be nice if one did not have to parse the RDF in the proxy 00:12:34 linked data is always nice you can conneg down to csv for your poor wget|cut queries 00:12:46 talk about interoperability 00:12:47 ah no 00:12:50 RDF all the way 00:12:54 no CSB 00:12:59 csv 00:13:02 haha of course the source is csv 00:13:21 some embedded linux don't even know what java is henry ;) 00:13:42 ? 00:13:54 *of course the source is RDF* 00:14:06 the javascript should be dealing with rdf 00:14:15 yes 00:14:24 then I can build a structure that will work even when CORS is fixed 00:14:37 already, our proxy currently connegs to text/turtle by default 00:14:47 as long as tabulator doesn't q= itself into xml 00:15:23 yes, but i like the javascript in the browser using rdf 00:15:24 should even convert rdfa if you dont want to wait for the rdflib.js tests to work! 00:16:13 yes. could do that as an intermediate step 00:16:13 you should just use kb in your app 00:16:15 just saying I can build another proxy and put it up on foafssl.org 00:16:26 and let rdflib/tabulator handle the cors/proxy situation 00:16:33 yes 00:16:57 thats great, we can cycle through to the fastest/local proxy on our client 00:17:14 in the end your own freedom box is your only proxy 00:17:20 but firefox has to work too :| 00:17:29 which you connect to with WebID 00:17:37 so there is no danger of overloading a server 00:17:43 yes unless it gets too busy, maybe another on your LAN has some free time 00:18:19 yes. initially we have some proxy services. But in the end your proxy server should have a key it can use to pretend to be you 00:18:39 so it can get to access controlled statements 00:18:54 -> that's why you can't have a global proxy 00:19:13 the proxy will just be a way to get around browser limitations 00:19:38 you can issue a cert to your proxy to let it act as part of you 00:19:43 your FB/proxy is just an extension of your browser 00:19:48 exactly 00:19:51 just add it to your foaf/card/webid 00:19:56 yes 00:20:06 but that means I can't use your proxy (long term) 00:20:14 when you are done, drop the triple :) 00:20:17 every body just uses their FB 00:20:40 so that is where we want to end up at 00:20:46 sure you can do your own authentication on your freedombox 00:20:51 long term: working browsers with no proxy 00:20:57 like allow 192.168/16 is pretty common 00:21:06 medium term: person FB with personal proxy 00:21:12 short term: any proxy will do 00:21:32 ah yes, you can have local proxy on machine too 00:22:00 but that's more work for the user 00:22:28 in any case to start off with we should get it to work with any proxy 00:22:36 "proxy service" 00:22:41 should be the right name 00:22:50 because these proxies are not real http proxies 00:23:30 I'd refer to it as an LD proxy 00:23:49 so if I can set #rdf.proxysvc = "http://data.fm/srv/cors?{url} 00:23:52 that would be good 00:23:55 melvster has quit (Ping timeout: 276 seconds) 00:24:31 I mean $rdf.proxysvc = "http://data.fm/srv/cors?{url} 00:25:11 the pitty about LD proxy is that it has to parse the rdf before spitting it out again: that uses memory 00:25:33 a pure representation based proxy can be 1000 more efficient 00:25:42 using new IO 00:25:53 also known as the select() call in C 00:26:42 the problem is that with a proxy service like that your relative URLs don't get rewritten 00:26:55 and so you end up creating new URLs 00:27:09 there are tradeoffs with any of this 00:27:36 yes a pure HTTP proxy would be nice, but it won't work for people who are behind a proxy I think 00:27:46 (unless one can chain proxies?) 00:27:54 melvster started without rdflib.js 00:28:07 just using JSON 00:28:43 to merge with FB graph JSON 00:28:44 I don't want to use JSON 00:28:49 I think it's mixing up syntax and semantics 00:28:56 I want to use RDF 00:29:01 if you don't want to do connect, yes you can just retransmit packets 00:29:21 question: can one chain pure HTTP proxies 00:29:35 if you work with lots of types of clients, conneg routing is worth the cost I think 00:30:17 also you get extra assurances and data conditioning 00:30:50 in reality when LD takes over all your interfaces, you will just run your own localhost proxy 00:31:05 can be 00:31:05 that way every app doesn't need your foaf in its own memory 00:31:19 they can all just query their shared LD memory / proxy 00:32:59 ... 00:33:02 yes, but still you are republishing rdf you may not want to republish 00:33:15 I think the Via: is used to inidicate chaining? 00:33:25 was just looking at that 00:33:33 is this the thing? http://www.ietf.org/rfc/rfc2068.txt 00:33:59 access to (re)published RDF is always assumed subject to WebID / ACL 00:34:59 well anyway. see what you can do. I was just trying to see if one could make this very efficinet 00:35:45 what were the major slowdowns when you were testing? 00:37:31 do you need a way to specify http-proxy to eg. data.fm/proxy? 00:37:43 I don't have any yet. But I know that if someone sends a 1 MB foaf profile, parsing that in memory is going to take a lot of space 00:38:07 yes, please be nice to our proxy :) 00:38:10 with a simple HTTP proxy you could send any size of file I would not need more than 1k to pass it along 00:38:16 it can probably be easily DoSd 00:38:36 I am nice. The problem is people pointing to files that are 1MB large 00:38:44 or 10MB large 00:38:48 or 1000MB large 00:39:02 we can't tell in advance 00:39:17 ( Kinglsey had 1MB files he sent me some time ago) 00:39:58 thats ok the proxy might still do it for friends or localhost 00:40:19 then there are denial of service attacks on hashmaps 00:40:53 http://www.youtube.com/watch?v=R2Cq3CLI6H8 00:41:14 "28c3: Effective Denial of Service attacks against web application platforms" 00:41:16 those are things you need to take into consideration 00:41:43 people could design xml structures that use weakness in your hashmaps 00:42:07 sure Ruby had a bad one a bit ago 00:42:55 Anyway, we can get going with something that's vulnerable to start off with just to get going 00:43:05 but keep those in the back of your mind 00:44:09 yes there's basically unlimited fun ahead of us 00:44:16 :-) 00:44:51 anyway I am sure you can improve my code. Just let me know where you are heading so I know what I can work on next 00:45:09 send me a mail at henry.story@bblfish.net 00:45:14 time for me to go to sleep 00:47:29 betehess has quit (Quit: Leaving) 00:48:15 bblfish has quit (Quit: bblfish) 00:54:54 bblfish (~bblfish@AAubervilliers-651-1-229-62.w86-198.abo.wanadoo.fr) has joined #dig 01:12:08 bblfish has quit (Quit: bblfish) 03:21:52 presbrey has quit (Quit: Terminated with extreme prejudice - dircproxy 1.2.0) 04:46:45 cheater has quit (Ping timeout: 248 seconds) 04:49:28 presbrey (~presbrey@2001:4830:2446:b5:aede:48ff:fe00:2001) has joined #dig 05:21:15 cheater (~cheater@dslb-084-057-001-153.pools.arcor-ip.net) has joined #dig 05:28:51 bblfish (~bblfish@AAubervilliers-651-1-229-62.w86-198.abo.wanadoo.fr) has joined #dig 06:00:44 bblfish has quit (Ping timeout: 240 seconds) 06:11:55 bblfish (~bblfish@ALagny-551-1-33-93.w90-3.abo.wanadoo.fr) has joined #dig 06:58:20 tlr has quit (Quit: tlr) 07:27:10 nunnun_away is now known as nunnun 07:27:39 nunnun is now known as nunnun_away 08:16:33 tlr (~tlr@89.204.139.169) has joined #dig 08:29:10 tlr has quit (Quit: tlr) 10:22:08 melvster (~melvin@p4FF97F1C.dip.t-dialin.net) has joined #dig 10:27:18 github (~github@sh1-ext.rs.github.com) has joined #dig 10:27:18 [rdflib.js] timbl pushed 2 new commits to master: http://git.io/G1qcUw 10:27:18 [rdflib.js/master] small fixes - Tim Berners-Lee 10:27:18 [rdflib.js/master] sigh - Tim Berners-Lee 10:27:18 github has left #dig 11:40:59 timbl has quit (Quit: timbl) 12:15:04 rszeno has quit (Quit: Leaving.) 12:19:31 rszeno (~rszeno@79.114.101.170) has joined #dig 12:39:11 danbri has quit (Read error: Connection reset by peer) 12:40:03 danbri (~danbri@cable-146-255-156-245.dynamic.telemach.ba) has joined #dig 12:56:14 danbri has quit (Read error: Connection reset by peer) 12:56:23 danbri (~danbri@cable-146-255-156-245.dynamic.telemach.ba) has joined #dig 14:53:26 melvster1 (~melvin@p4FF96D30.dip.t-dialin.net) has joined #dig 14:55:07 melvster has quit (Ping timeout: 252 seconds) 15:26:45 timbl (~timbl@host109-156-61-106.range109-156.btcentralplus.com) has joined #dig 17:36:28 danbri has quit (Read error: Connection reset by peer) 17:36:44 danbri (~danbri@cable-146-255-156-245.dynamic.telemach.ba) has joined #dig 18:02:09 hey does anyone have multiple SANs in their WebID? 18:04:21 please send me [extensions] section: https://data.fm/test/x509 18:25:29 you should ask on #webid list 18:25:34 presbrey 18:25:45 there's bound to be someone there 18:42:10 danbri has quit (Remote host closed the connection) 18:51:19 presbrey: i think kingsley has multiple sans 18:52:37 yes 18:57:57 github (~github@sh1-ext.rs.github.com) has joined #dig 18:57:57 [rdflib.js] timbl pushed 1 new commit to master: http://git.io/v_6zcw 18:57:57 [rdflib.js/master] moving to ttl manifest - Tim Berners-Lee 18:57:57 github has left #dig 19:31:14 github (~github@sh1-ext.rs.github.com) has joined #dig 19:31:14 [data.fm] presbrey pushed 3 new commits to master: http://git.io/Gzg2xA 19:31:14 [data.fm/master] header: optimize sess deps - Joe Presbrey 19:31:14 [data.fm/master] WebID: add SSL_CLIENT_CERT implementation - Joe Presbrey 19:31:14 [data.fm/master] s/debug: add SSL_ - Joe Presbrey 19:31:14 github has left #dig 19:33:01 * now WebID works on the new data.fm 19:33:36 we'll have to update that after kingsley brings his multi-san webid around 19:37:18 tested it with my multi-key just fine 19:43:21 melvster1, yours work? 19:55:48 let me check 19:56:01 yes it works 19:56:07 , presbrey :) 19:58:10 presbrey ... im not sure if timbl's webid works on data.fm 19:58:24 failed last time we checked 20:10:45 bblfish, http://bblfish.com/ looks down from here 20:11:04 it's bblfish.net 20:12:52 oops hehe 20:12:59 timbl had a lot of spaces in his xsd:hexBinary 20:13:02 ok try https://data.fm/test/webid 20:13:20 so I brought that up on the xml-schema mailing list http://lists.w3.org/Archives/Public/www-xml-schema-comments/2012JanMar/0011.html 20:13:35 some seem to think xml schema could be made more flexible 20:13:49 others say that as things stand it is not 20:15:52 for example there was this in support http://lists.w3.org/Archives/Public/www-xml-schema-comments/2012JanMar/0018.html 20:16:25 an Henry Thompson also was in support for a more flexible xml schema 20:17:50 SPARQL on timbl's card returns 0 bindings 20:18:01 but we can see 24 triples on the graph 20:18:03 https://data.fm/test/webid?uri=http://www.w3.org/People/Berners-Lee/card%23i 20:19:39 <#i> cert:key [ a cert:RSAPublicKey; 20:19:40 cert:modulus """d7 a0 e9 1e ed dd cc 90 5d 5e cc d1 e4 12 ab 0c 20:19:40 5b db e1 18 fa 99 b7 13 2d 91 54 52 f0 b0 9a f5 ... 20:19:51 do you remove the white space? 20:20:23 If timbl thinks that should be legal then it would be good if he added a small note in support of my request to the xml schema crowd 20:20:48 github (~github@sh1-ext.rs.github.com) has joined #dig 20:20:48 [data.fm] presbrey pushed 2 new commits to master: http://git.io/NBZwXQ 20:20:48 [data.fm/master] webid: separate query function - Joe Presbrey 20:20:48 [data.fm/master] root: add test - Joe Presbrey 20:20:48 github has left #dig 20:21:38 I'll write a summary to the webid mailing list and semweb mailing list 20:25:22 timbl would work if he didn't have syntax errors 20:25:25 rapper -q -c -i turtle -o turtle 'http://www.w3.org/People/Berners-Lee/card' 20:25:25 rapper: Error - URI http://www.w3.org/People/Berners-Lee/card:47 - syntax error 20:25:25 rapper: Failed to parse URI http://www.w3.org/People/Berners-Lee/card turtle content 20:26:09 47: is rdfs:seeAlso of card:i; # Go fetch it 20:26:14 that's probably because it's n3 20:26:19 not turtle 20:26:28 and few tools parse n3 20:26:35 arg 20:34:34 curl -vH 'Accept: text/turtle, application/rdf+xml' 'http://www.w3.org/People/Berners-Lee/card' 2>&1 >/dev/null | grep Content-Type 20:34:35 < Content-Type: text/n3; qs=0.89 20:35:01 seems bad 20:35:04 kidehen (18daf0de@gateway/web/freenode/ip.24.218.240.222) has joined #dig 20:35:34 hey Kingsley 20:35:56 remind me how to find your webID? 20:35:57 @presbrey : still getting hang when I present my WebID to data.fm . This time its at the manage stage. I am testing from Chrome 20:36:23 try: http://data.fm/test/webid 20:36:27 oops 20:36:30 https://data.fm/test/webid 20:36:39 yes that will need https: 20:36:58 id.openlink...? 20:37:31 @presbrey: http://id.myopenlink.net/dataspace/person/KingsleyUyiIdehen#this 20:39:35 what does your [claim] say? 20:40:07 you mean: what's the relation connecting WebID to public key in my graph? 20:40:50 : what are you actually looking up? 20:42:12 if you go to https://data.fm/test/webid, you should see a [claim] section 20:43:56 there is also a [lookup] section and [verified] 20:44:04 : I see that 20:44:21 the claim section should say: 20:44:23 [uri] => http://id.myopenlink.net/dataspace/person/KingsleyUyiIdehen#this 20:44:28 I see: [verified] => http://id.myopenlink.net/dataspace/person/KingsleyUyiIdehen#this 20:44:42 great so yours works! :) 20:45:57 : yes, but I can't get going with setting up my space in data.fm. Eons ago (first cut of this) it all worked fine 20:46:19 thanks Kinglsey! 20:46:24 just go to any new cloud and PUT 20:46:35 https://kingsleysnewone.data.fm 20:47:14 kidehen: try http://vote.data.fm/ 20:47:40 , : ok 20:48:17 : https://kingsleysnewone.data.fm/, works 20:50:18 : as you can see from ++ on http://vote.data.fm/, it worked :-) 20:50:27 yay 20:52:05 : next is getting the simple Resource CRUD pattern working properly for all RDF formats. Early next week (latest) I'll demonstrate how ODS-Briefcase does this with its old circa. 2000 UI (we'll update this with HTML5 + JS soon) 20:57:24 ok by for now 20:57:57 kidehen has quit (Quit: Page closed) 21:09:10 ok, I sent a message out regarding hexBinary to the list 23:26:41 melvster1 has quit (Ping timeout: 252 seconds)