IRC log of dig on 2013-09-10
Timestamps are in UTC.
- 01:21:50 [melvster]
- melvster has quit (Ping timeout: 240 seconds)
- 02:45:55 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-226-107.w86-198.abo.wanadoo.fr) has joined #dig
- 02:47:23 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 05:28:33 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-226-107.w86-198.abo.wanadoo.fr) has joined #dig
- 06:05:25 [tyteen4a03]
- tyteen4a03 has quit (Ping timeout: 245 seconds)
- 06:18:21 [bblfish]
- bblfish has quit (Ping timeout: 245 seconds)
- 06:33:48 [bblfish]
- bblfish (~bblfish@AAubervilliers-652-1-89-229.w83-114.abo.wanadoo.fr) has joined #dig
- 06:41:41 [tyteen4a03]
- tyteen4a03 (tyteen4a03@2001:470:e2e4::) has joined #dig
- 07:04:42 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 07:14:50 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-226-107.w86-198.abo.wanadoo.fr) has joined #dig
- 07:16:41 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 07:28:44 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-226-107.w86-198.abo.wanadoo.fr) has joined #dig
- 07:40:40 [deiu]
- deiu (~andrei@157.159.103.113) has joined #dig
- 07:40:40 [deiu]
- deiu has quit (Changing host)
- 07:40:40 [deiu]
- deiu (~andrei@unaffiliated/deiu) has joined #dig
- 07:43:34 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 08:40:47 [jmvanel]
- jmvanel (~jmvanel@199.0.88.79.rev.sfr.net) has joined #dig
- 09:12:13 [Yudai_]
- Yudai_ has quit (*.net *.split)
- 09:12:16 [sandro]
- sandro has quit (*.net *.split)
- 09:20:15 [sandro]
- sandro (~sandro@ssh.w3.org) has joined #dig
- 09:20:15 [Yudai_]
- Yudai_ (~Yudai@nttkyo218001.tkyo.nt.ngn2.ppp.infoweb.ne.jp) has joined #dig
- 09:31:26 [timbl]
- timbl (~timbl@host86-146-148-68.range86-146.btcentralplus.com) has joined #dig
- 09:45:36 [bblfish]
- bblfish (~bblfish@90.24.245.239) has joined #dig
- 09:45:52 [melvster]
- melvster (~melvster@89.176.108.70) has joined #dig
- 10:20:20 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 10:20:53 [bblfish]
- bblfish (~bblfish@90.24.245.239) has joined #dig
- 10:25:54 [bblfish]
- bblfish has quit (Ping timeout: 276 seconds)
- 11:00:14 [deiu]
- deiu has quit (Read error: Connection reset by peer)
- 11:00:42 [deiu]
- deiu (~andrei@unaffiliated/deiu) has joined #dig
- 11:24:18 [RalphS]
- RalphS (rswick@w3cvpn1.w3.org) has joined #dig
- 11:48:44 [bblfish]
- bblfish (~bblfish@90.24.245.239) has joined #dig
- 12:06:36 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 12:08:56 [bblfish_]
- bblfish_ (~bblfish@90.24.245.239) has joined #dig
- 12:25:53 [bblfish_]
- bblfish_ has quit (Remote host closed the connection)
- 12:29:51 [timbl]
- presbrey?
- 12:30:20 [bblfish]
- bblfish (~bblfish@90.24.245.239) has joined #dig
- 12:30:38 [timbl]
- Seems that when spawning a new tracker, I'm getting 200 back from the PUT but no actual data stoted
- 12:32:02 [timbl]
- e.g. http://timbl.data.fm/test/issuetracker.w3.org/1378748362742/track should have bunch of triples in
- 12:33:31 [timbl]
- This happens whether using the code for making a new tracker, or just curl -v --upload-file foo2.n3 -HContent-type:text/turtle http://timbl.data.fm/test/issuetracker.w3.org/1378748362742/track
- 12:42:36 [deiu]
- timbl, about your last question regarding rww.io and data.fm, they should be merged (maybe replace data.fm with rww.io since it brings lots of new features)
- 12:44:40 [deiu]
- rww.io should support the old ACL convention (.meta in the / dir), but I need to check with presbrey and test it
- 12:49:59 [timbl]
- By merge, that would leave the source trees the same. Just get the best of both'
- 12:50:23 [timbl]
- When it comes to the .meta .acl etc
- 12:50:50 [timbl]
- the main thing is that there is a common protocol -- and a common client-ide library -- that works with either way.
- 12:51:41 [timbl]
- So it shouldn't matter what the filename is, but it should matter that client find s it in a standard way
- 12:56:28 [scor]
- scor (scor@drupal.org/user/52142/view) has joined #dig
- 12:58:20 [bblfish]
- bblfish has quit (Ping timeout: 245 seconds)
- 12:58:47 [bblfish]
- bblfish (~bblfish@90.24.245.239) has joined #dig
- 13:01:50 [deiu]
- timbl, wouldn't it be useful to have a CG/WG to discuss Web ACL stuff?
- 13:02:02 [timbl]
- (If I try it with rww.io I get a 403)
- 13:02:32 [deiu]
- the goal is to standardize WebACL discovery and management
- 13:02:35 [timbl]
- DO you think the Web ACL stuff should be in a separate list, nonpublic-rww ?
- 13:02:50 [deiu]
- I think it's part of the RWW CG now, right?
- 13:04:37 [deiu]
- are you trying to upload foo2.n3 to rww.io?
- 13:05:19 [timbl]
- yes
- 13:05:34 [timbl]
- just s/data.fm/rww.io/ above
- 13:06:28 [timbl]
- How can I set that up for public access -- how can I get at my ffox cert and do webid with curl?
- 13:06:57 [deiu]
- you can export the cert as pem
- 13:07:29 [deiu]
- then use curl --cert or curl -E
- 13:09:30 [deiu]
- BTW, you need to set the acl for /test before writing to it as a public user
- 13:10:09 [deiu]
- open the UI and tick the Write and Default for new checkboxes
- 13:10:28 [deiu]
- might as well give it Read too
- 13:11:08 [deiu]
- hmm or Append instead of Write (can PUT new resources with it)
- 13:12:10 [timbl]
- Somethimng which starts -----BEGIN CERTIFICATE----- is a .pem ?
- 13:12:18 [deiu]
- yes
- 13:13:44 [timbl]
- so curl -v -E timbl.pem
- 13:13:49 [deiu]
- that should work
- 13:14:03 [timbl]
- I still get 403 .. so if I make an ACL on /test/ that will be enough?
- 13:14:25 [deiu]
- did you export the cert with a password?
- 13:14:37 [deiu]
- in that case: curl -v -E timbl.pem:password
- 13:15:01 [deiu]
- or set some default ACL rules for /test/
- 13:15:40 [deiu]
- oh, you're using a http:// url
- 13:15:49 [deiu]
- you might want to use https:// to trigger WebID auth
- 13:16:43 [timbl]
- ooops yes.
- 13:16:47 [timbl]
- Ok now I get unable to set private key file: 'timbl.pem' type PEM
- 13:18:23 [deiu]
- did you convert the cert from .p12 to .pem yourself?
- 13:18:50 [timbl]
- That is just what you get from the Firefox export button
- 13:18:51 [deiu]
- I think FF exports a password protected .p12 by default
- 13:18:55 [timbl]
- I din't change it
- 13:20:03 [deiu]
- I'm not sure how it behaves on OS X
- 13:22:56 [timbl]
- do the thing starting -----BEGIN CERTIFICATE--- might be a p12 not a pem ?
- 13:24:28 [deiu]
- I exported a cert using FF and it doesn't have that line (BEGIN CERT..)
- 13:24:58 [deiu]
- can you use check if this command works? openssl pkcs12 -in file.p12 -out file.pem
- 13:25:08 [deiu]
- replace file.p12 with your cert
- 13:26:23 [timbl]
- No, not happy
- 13:26:24 [timbl]
- 140735183428028:error:0D0680A8:asn1 encoding routines:ASN1_CHECK_TLEN:wrong tag:tasn_dec.c:1319:
- 13:26:25 [timbl]
- 140735183428028:error:0D07803A:asn1 encoding routines:ASN1_ITEM_EX_D2I:nested asn1 error:tasn_dec.c:381:Type=PKCS12
- 13:26:51 [timbl]
- expecting a raw ASN/1 file maybe
- 13:27:10 [cheater__]
- cheater__ (~cheater@p5498A868.dip0.t-ipconnect.de) has joined #dig
- 13:27:44 [deiu]
- try: openssl x509 -text -in timbl.pem
- 13:27:52 [deiu]
- does it print the cert correctly?
- 13:28:35 [timbl]
- Yes
- 13:28:43 [deiu]
- then it's definitely a pem
- 13:29:14 [timbl]
- (includeing URI:https://webid.mit.edu/timbl#)
- 13:29:28 [deiu]
- did you set a password when you exported it?
- 13:29:55 [timbl]
- No
- 13:30:20 [timbl]
- curl would shave prompted I think
- 13:30:41 [deiu]
- I think you might have exported only the public key
- 13:31:54 [deiu]
- I find it strange that FF didn't ask for a password, or that it exported by default as PEM
- 13:33:33 [timbl]
- Ah I'd missed the export format option opn FFox export button
- 13:33:52 [timbl]
- the default is "X509 cert (PEM)"
- 13:34:18 [timbl]
- There is another option X509 cert (PEM) (with chain)"
- 13:36:16 [timbl]
- yes the dump of the cert does not have private key info
- 13:38:06 [deiu]
- does it work now?
- 13:39:29 [timbl]
- I haven't found anything to change. exporting wit the chain gives the same file
- 13:40:27 [deiu]
- what FF version are you using?
- 13:41:37 [timbl]
- 23.0.1
- 13:41:48 [deiu]
- I have the same version
- 13:42:24 [deiu]
- I'm doing Preferences -> Advanced -> Certificates -> View Certificates -> Backup...
- 13:43:47 [timbl]
- I was doing View button to get to one cert, then "detail" tab the "expert" button
- 13:43:52 [timbl]
- export
- 13:44:06 [deiu]
- ah
- 13:44:10 [deiu]
- use the Backup... button
- 13:44:35 [deiu]
- export will only save the public key
- 13:44:45 [timbl]
- That will include al 3 certs and male a p12
- 13:45:20 [deiu]
- then you can convert the p12 to pem using: openssl pkcs12 -in file.p12 -out file.pem
- 13:47:05 [timbl]
- ok
- 13:48:43 [timbl]
- I see .. I though it would backup all 3 but it only does the selected one
- 13:49:53 [deiu]
- I think you have a Backup All for that :-)
- 13:50:03 [timbl]
- Nice. * We are completely uploaded and fine
- 13:50:51 [timbl]
- https://webid.mit.edu/timbl#
- 13:51:15 [timbl]
- from the "User:" field
- 13:51:30 [deiu]
- so it works
- 13:51:39 [timbl]
- Is that field a webid standard as it were … can the tabulator client code look for it?
- 13:51:46 [timbl]
- Yes, it works
- 13:52:07 [deiu]
- it's not a standard but I would REALLY want for it to be
- 13:52:34 [deiu]
- otherwise I don't know how a web app can tell if the user is authenticated or not
- 13:53:08 [deiu]
- and User: should be protocol-independent
- 13:53:20 [timbl]
- Well, if your an program the browser, the bit is easy as yo know which cert you are wielding …. and of course a user can have different certs on different stores.
- 13:53:46 [deiu]
- you can't access the cert info in all browsers
- 13:54:24 [timbl]
- So we have to be careful of jumping through too many hoops to do things you could do easily programming the cb=broswre, and ten eth browser manufacturers saying that the protocol is unncesessailty complicated
- 13:54:27 [timbl]
- :-)
- 13:55:07 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 13:56:18 [deiu]
- but you still need a way to get feedback from the server regarding which identity is currently "logged in"
- 13:56:27 [timbl]
- Yes.
- 13:56:54 [timbl]
- But the client in principle knows. It is just the browser knows, not the webapp
- 13:57:00 [timbl]
- or the extension.
- 13:57:48 [deiu]
- what if the authentication failed? the browser still knows it used the cert for https://webid.mit.edu/timbl#, but it doesn't mean the server has logged the user in
- 13:58:27 [deiu]
- the cert is selected in the browser, but the user has failed to login on the server
- 13:58:50 [timbl]
- Well, I think the "User:" is useful anyway.
- 13:59:29 [deiu]
- it's a hack
- 14:00:14 [timbl]
- why?
- 14:01:35 [deiu]
- it's not a standardized header
- 14:01:35 [bblfish]
- bblfish (~bblfish@90.24.245.239) has joined #dig
- 14:01:36 [timbl]
- In the webid spec it can be.
- 14:02:17 [timbl]
- ANy architectural reasons its a hack?
- 14:02:26 [deiu]
- not really
- 14:02:44 [deiu]
- I suppose we can mention it in the WebID-TLS spec
- 14:03:16 [deiu]
- I find it really useful for web apps
- 14:03:54 [timbl]
- yes
- 14:04:21 [timbl]
- the tabulator has 'me' preference which it uses but the user has to sit themsleves
- 14:04:47 [timbl]
- I could, anytime a "User" URI is seen, change that.
- 14:06:16 [deiu]
- would it be more useful to have a Link rel=identity header?
- 14:06:45 [timbl]
- I think the link rel= express relationships between the document and other things.
- 14:06:54 [timbl]
- The currently logged in user is to that
- 14:07:03 [timbl]
- is not that.
- 14:07:08 [deiu]
- true
- 14:07:53 [timbl]
- (entity headers and protocol headers or something)
- 14:15:22 [deiu]
- speaking of link rel, rww.io HTTP responses contain a link=acl for the .acl file and a link=meta for the .meta file (if the resource is not an RDF document)
- 14:35:50 [tyteen4a03]
- tyteen4a03 has quit (Ping timeout: 245 seconds)
- 14:46:15 [betehess]
- betehess has quit (Ping timeout: 245 seconds)
- 14:47:23 [tyteen4a03]
- tyteen4a03 (tyteen4a03@2001:470:e2e4::) has joined #dig
- 14:48:04 [betehess]
- betehess (~betehess@2001:470:8b2d:804:4451:91a6:5716:e8cc) has joined #dig
- 14:49:10 [timbl]
- is there any client code for changing ACLs ?
- 14:51:39 [timbl]
- deiu?
- 14:52:09 [deiu]
- no, the web app should handle that
- 14:52:16 [timbl]
- I have made a new tracker now using the tracker ane
- 14:52:20 [timbl]
- pane
- 14:52:25 [timbl]
- at http://timbl.rww.io/test/issuetracker.w3.org/1378824620728/track#TabTracker
- 14:52:43 [timbl]
- but that file gets returned with a syntax error
- 14:53:43 [timbl]
- Extra dots
- 14:54:02 [timbl]
- <#Bug>
- 14:54:03 [timbl]
- a <http://www.w3.org/2000/01/rdf-schema#Class> ;
- 14:54:04 [timbl]
- <http://www.w3.org/2000/01/rdf-schema#label> "bug" ;
- 14:54:05 [timbl]
- <http://www.w3.org/2000/01/rdf-schema#subClassOf> <#TabIssueCategory> ;
- 14:54:06 [timbl]
- <http://www.w3.org/ns/ui#backgroundColor> "#fffed0" ;
- 14:54:07 [timbl]
- <http://www.w3.org/ns/ui#sortOrder> 70. .
- 14:54:21 [timbl]
- An extra trailing dot
- 14:54:30 [timbl]
- Which serializer are you using?
- 14:54:53 [timbl]
- It isn't generaing prefixes
- 14:55:58 [deiu]
- I'm using rdflib
- 14:56:12 [deiu]
- that extra dot is weird
- 14:56:18 [deiu]
- it's the first time I see it
- 14:56:53 [deiu]
- it seems to appear only after an integer value
- 14:56:58 [timbl]
- Oh… I think it may be decimal
- 14:57:07 [timbl]
- missingt trailing 0
- 14:57:16 [deiu]
- yes
- 14:57:21 [timbl]
- maybe syntax ambiguity
- 15:08:20 [betehess]
- betehess has quit (Ping timeout: 245 seconds)
- 15:22:48 [betehess]
- betehess (~betehess@31-35-251.wireless.csail.mit.edu) has joined #dig
- 15:36:17 [deiu]
- deiu has quit (Quit: Leaving)
- 15:59:02 [melvster]
- timbl deiu: User: is on it's way to becoming a standard ... I've done some consensus gathering with mnot, roy, nathan, kingsley, manu, the rww group and the openid foundation
- 15:59:18 [melvster]
- people were asking for use cases, so I'll write some up
- 15:59:42 [melvster]
- the feedback was *not* to reuse the "From: " header as that is for email only
- 16:00:32 [melvster]
- unofficially I think it's OK to use, there were suggestions of calling it UserID too, but I think User is just about winning the naming, and it's already in use in data.fm implemented by presbrey
- 16:00:58 [melvster]
- I plan to make a wiki page, we have text already, then try and submit it to the IETF registry for feedback
- 16:01:20 [timbl]
- What's the best approximation of a webs spec just now?
- 16:01:42 [melvster]
- "webs spec" ?
- 16:02:13 [timbl]
- webid spec
- 16:02:15 [deiu]
- deiu (~andrei@2a01:e35:8b67:4160:2d92:4f6a:612a:a1c1) has joined #dig
- 16:02:15 [deiu]
- deiu has quit (Changing host)
- 16:02:15 [deiu]
- deiu (~andrei@unaffiliated/deiu) has joined #dig
- 16:02:19 [timbl]
- stupid autocorrect
- 16:03:08 [melvster]
- timbl: https://dvcs.w3.org/hg/WebID/raw-file/tip/spec/index.html
- 16:03:38 [melvster]
- henry wants to publish it next week to it's home at: http://www.w3.org/2005/Incubator/webid/spec/
- 16:03:42 [melvster]
- cc bblfish
- 16:05:43 [timbl]
- Should User: not be aded as a should to WebID-TLS?
- 16:05:56 [timbl]
- or WebID?
- 16:07:31 [cheater__]
- cheater__ has quit (Ping timeout: 245 seconds)
- 16:07:50 [melvster]
- timbl: in WebID-TLS you get the user from the SubjectAlternativeName ... adding User sounds to me like a good idea, but not sure what the group's opinion on that would be ...
- 16:08:45 [melvster]
- in fact the subject alternative name can contain a list of multiple user URIs
- 16:08:49 [deiu]
- timbl, I'll try to put together and propose a paragraph about User:
- 16:09:06 [melvster]
- deiu: I'd be happy to work with you on that ... I already have text for it
- 16:09:13 [deiu]
- melvster, I think timbl is referring to the User: header
- 16:09:16 [melvster]
- yes
- 16:09:26 [deiu]
- brb
- 16:10:20 [melvster]
- DIGlogger, pointer
- 16:10:20 [melvster]
- See http://dig.csail.mit.edu/irc/dig/2013-09-10#T16-10-20
- 16:12:30 [scor]
- scor has quit (Ping timeout: 245 seconds)
- 16:14:51 [timbl]
- Or should it be added to WebID so that the User: field is given whatever the auth method used?
- 16:15:38 [timbl]
- Yes I'm talking about a User: HTTP header which I notice I get from rww.io
- 16:15:43 [timbl]
- and I like (!)
- 16:16:33 [scor]
- scor (scor@nat/acquia/x-epgddxragswxqrhj) has joined #dig
- 16:16:33 [scor]
- scor has quit (Changing host)
- 16:16:33 [scor]
- scor (scor@drupal.org/user/52142/view) has joined #dig
- 16:19:00 [deiu]
- IMO, the User: header should not be part of the WebID spec
- 16:19:16 [deiu]
- it should be protocol-independent
- 16:19:45 [deiu]
- there should be a "RWW" spec about these things
- 16:19:59 [melvster]
- timbl: User is also useful without auth, I could personalize a page for you, auto filling in many fields without you having to type things on a mobile device, I could also add your name and avatar, and find information from your preferences ... with auth it of course can act as a username
- 16:20:54 [melvster]
- and also when you try and *change* something, you need auth or when you access *protected* information
- 16:21:37 [deiu]
- melvster, I don't understand
- 16:22:15 [deiu]
- personalize a page when you pretend to be someone else?
- 16:22:45 [melvster]
- deiu: im only talking about leveraging public information from your profile here
- 16:23:25 [deiu]
- User: is returned by the server, not sent by the agent
- 16:23:37 [melvster]
- oic
- 16:24:10 [melvster]
- hmmmmm
- 16:24:15 [deiu]
- it's the server's way of indicating that your request was performed based on that particular identity (which is set in the User header)
- 16:24:37 [melvster]
- cant the request send user too?
- 16:24:45 [deiu]
- why?
- 16:25:01 [deiu]
- isn't that what authentication does?
- 16:25:08 [melvster]
- no!
- 16:25:17 [melvster]
- authentication and identification are different things
- 16:26:04 [deiu]
- as a server, I have no incentive to personalize (= spend time/resources) a page if I'm not sure I'm doing it for the right person
- 16:26:31 [deiu]
- besides, identification without proof is wrong
- 16:26:47 [deiu]
- in your use-case at least
- 16:28:35 [melvster]
- deiu: it's how communication works in every field, you identify, and only authenticate on demand ... imagine that it was impossible to write someone a letter without proving your identity?
- 16:28:50 [melvster]
- or phoning someone up
- 16:29:11 [melvster]
- identity and authn are modular concepts
- 16:31:30 [melvster]
- deiu: to put it another way, why would anyone want to identify themselves as the *wrong* person?
- 16:31:47 [deiu]
- melvster, in your example, you are basically saying that you want the server to display a page and personalize it as if you were timbl
- 16:32:28 [melvster]
- deiu: i didnt say I *want* that ... I said it's a possible use case for the server to use public information associated with a URI, e.g. to auto fill forms
- 16:32:46 [deiu]
- like credit card information forms? :-)
- 16:33:05 [deiu]
- or full name / address info?
- 16:33:14 [melvster]
- deiu: do you keep your credit card details public? ;)
- 16:33:28 [deiu]
- no, but the server may save them
- 16:33:41 [deiu]
- anyway
- 16:33:46 [timbl]
- There are two things
- 16:33:52 [deiu]
- I think your use-case is not complete
- 16:34:07 [deiu]
- there's information something missing
- 16:34:35 [deiu]
- besides, browsers already do autocomplete
- 16:34:58 [timbl]
- One is the user id the user has authenticated with tho the server through the client. The other is the person the clinet understands the users to be, which affects certain data, like stuff about people, where the display will be user-related -- like "You are friends with x".
- 16:35:47 [melvster]
- timbl: right, that's used in the social and microblogging panes
- 16:36:51 [deiu]
- who sends the second one?
- 16:40:05 [timbl]
- The second one is set by the user with the "login in" code where it prompts for awe did if you don't have a browsing id
- 16:40:14 [timbl]
- It isn't sent over the net
- 16:40:22 [timbl]
- it is local to the client and the user
- 16:41:04 [timbl]
- The user agent needs to know whop the user is because it has to start with a lot of user preferences
- 16:41:19 [timbl]
- including which workspaces a user has available, fav language, etc etc
- 16:41:56 [timbl]
- Baically like a unit home directory allow you to look up ~/.xxx rc and ~/.xxx.config etc
- 16:42:45 [deiu]
- so the second one is bound to the application
- 16:42:48 [timbl]
- In this experimental workspace ontology here is a pointer from th user's public ID to the private preferences file.
- 16:43:21 [deiu]
- in other words, it's part of the local preferences
- 16:43:30 [melvster]
- timbl: can preferences be public too?
- 16:43:40 [timbl]
- http://www.w3.org/ns/pim/space#preferencesFile
- 16:43:47 [timbl]
- It could be public
- 16:43:55 [timbl]
- but I wouldn't want mine to be
- 16:44:02 [timbl]
- as it points to private worspaces
- 16:44:06 [deiu]
- yes
- 16:44:16 [deiu]
- I'm just trying to see if/how this can be abused
- 16:45:01 [timbl]
- If everyone's pref file is in a similar place then the URI of the file itself should not give much away
- 16:46:40 [timbl]
- Not like <#me> space:preferencesFile <https://private.nsa.gov/groups/snooping/andrewMypreferences.n3>.
- 16:47:22 [timbl]
- At the moment I cheat -- I have one on /localhost
- 16:47:37 [timbl]
- so the link will confuse others
- 16:47:40 [deiu]
- you could have an acl for it though
- 16:47:50 [timbl]
- yes.
- 16:48:05 [timbl]
- definitely have an cal for it.
- 16:48:13 [timbl]
- s/cal/acl/
- 16:49:01 [deiu]
- I see
- 16:49:04 [melvster]
- ontology look great: public / private / shared ... that should cover most use cases ...
- 16:49:30 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 16:49:50 [deiu]
- BTW, are these workspaces per application or "shared" between apps?
- 16:50:26 [melvster]
- could be either I imagine, the ontology doesnt forbid sharing
- 16:50:43 [deiu]
- would a calendar app be able to GET data from an agenda app?
- 16:51:07 [timbl]
- It will probably be a good idea to shave a bit of code which juts sets up a preferences file and a set of workspaces
- 16:51:25 [timbl]
- A trusted app can get any data a user can access
- 16:51:57 [timbl]
- "shared" for a workspace is
- 16:52:04 [timbl]
- shred between users.
- 16:52:27 [deiu]
- aha
- 16:52:37 [melvster]
- longer term, I think apps should have finer grained permissions, so that it's not necessarily all or nothing
- 16:52:40 [timbl]
- Current thinking about string is between apps is we have spaces which are app-specific and places which are standardizes
- 16:53:01 [timbl]
- do e.g. a map space any mail client can read and write to.
- 16:53:08 [deiu]
- I was thinking about cross-app workspaces
- 16:54:05 [melvster]
- I think we need delegated credentials in that case where the app has it's own identity and you tell it what it can or cant access, much like OAuth does for facebook / google+ etc.
- 16:54:10 [deiu]
- or workspaces dedicated to specific types of resources -- i.e. a photo album app will ask permission to access your "photos" workspace
- 16:54:29 [timbl]
- Two modes, one its you have a cross-app space, the oethr is that one app starts using its own space (like the tracker does) and others just peek into that and join in
- 16:54:33 [timbl]
- extending it.
- 16:54:45 [deiu]
- yes, like the gps app
- 16:55:22 [timbl]
- delegated credentials -- yes -- well, I think I need to be abel to say "I will allow Melvin to access this data only with this application".
- 16:55:26 [deiu]
- "Please indicate where I can save my files" kind of question
- 16:55:53 [timbl]
- Ihave that sort of code written
- 16:57:25 [deiu]
- granting access based on a specific app is difficult
- 16:57:50 [timbl]
- If you make a preferences file like http://pastebin.com/9aExuKW0
- 16:57:56 [deiu]
- it forces users to use app X over Y
- 16:58:41 [timbl]
- well yes but what do you do when someone want to use an app off he web which will steal your data like an iPhone free weather app steals you calendar as a privacy invasion?
- 16:58:57 [timbl]
- An arbitrary app might be maliceious
- 16:59:01 [deiu]
- timbl, that's the million $ question
- 16:59:08 [timbl]
- or it might "just" steak private at a.
- 16:59:19 [timbl]
- steal
- 16:59:24 [timbl]
- A trusted app won't
- 16:59:33 [deiu]
- I only trust apps I write myself :-)
- 16:59:36 [timbl]
- A trusted app isbenificent.
- 17:00:26 [deiu]
- well, I think this problem affects any kind of software system, even outside the Web
- 17:00:57 [timbl]
- Yes.
- 17:00:59 [deiu]
- you can have a trojan that sends data away, even if you use a trusted app
- 17:01:12 [timbl]
- BUttraditionally all software loaded on your computer was beneficent.
- 17:01:53 [deiu]
- BTW, a trust app also means some sort of trusted app store
- 17:02:00 [deiu]
- s/trust/trusted
- 17:02:25 [timbl]
- Beneficent here meaning roughly "would do what the user would normally be expected to want it to do if it had time to ask th user" -- i.e. no ads
- 17:02:57 [timbl]
- I think for me yes useful to have atrusted app store but the installation process is important
- 17:03:30 [deiu]
- being able to install the app locally is very important
- 17:04:03 [timbl]
- Yes. I may be able to e.g. set up a set of githb users which I assume are beneficent nd if they have checked stuff in my client will run it with less of a installation hurdle for example.
- 17:04:03 [deiu]
- you can audit the app yourself and make sure you can trust it, since the code won't change
- 17:04:14 [timbl]
- You can also set up a review process.
- 17:04:28 [timbl]
- You can sign a cert for a given hit hub hash
- 17:04:50 [timbl]
- (which is feet in that it can be verified locally to still give the same hash I assume)
- 17:05:00 [deiu]
- I need to find a company to found me so I can work on it :-)
- 17:05:01 [timbl]
- s/feet/sweet/
- 17:05:11 [deiu]
- s/found/fund/
- 17:06:13 [deiu]
- right now it would be great to have an app manifest vocabulary
- 17:06:30 [deiu]
- otherwise you can't know what to expect from the app
- 17:07:58 [deiu]
- it's great that we have data.fm / rww.io as a base where people can install apps
- 17:10:01 [deiu]
- maybe I can package rww.io for debian, so people can install it everywhere
- 17:12:36 [timbl]
- http://www.hhs.gov/ohrp/policy/belmont.html#xbenefit
- 17:13:06 [timbl]
- where I came across the word beneficent
- 17:13:37 [timbl]
- I don't want data.fm code and rww.io code to get too far apart
- 17:14:16 [deiu]
- isn't presbrey working on a python implementation of data.fm?
- 17:14:46 [melvster]
- it may be possible to bootstrap the mozilla marketplace : https://marketplace.firefox.com/
- 17:15:46 [deiu]
- they have a really nice manifest system
- 17:16:01 [deiu]
- it can easily be transformed into an ontology
- 17:17:35 [timbl]
- I think presbrey might prefer everyone to switch to python
- 17:18:46 [timbl]
- https://github.com/linkeddata -> https://github.com/linkeddata/ldpy
- 17:19:07 [deiu]
- I would like that too
- 17:19:27 [melvster]
- deiu: there's a community group working on manifests, w3c has some work in this area e.g. widgets
- 17:19:57 [timbl]
- YEs, a whole spec which wasn't adopted.
- 17:20:30 [melvster]
- timbl: it's a pity, w3c widgets looked like a great spec
- 17:22:26 [melvster]
- perhaps it will be a good excercise to bootstrap the mozilla market place so that it's possible to install apps into a framework like tabulator
- 17:23:27 [melvster]
- most apps are free and dont require auth, which is nice ... and on unity they can even be downloaded as standalone desktop apps
- 17:24:57 [timbl]
- Hmmm the code in https://github.com/linkeddata/ldpy/blob/master/ld.py may be a one-file solution
- 18:08:24 [zuzak]
- zuzak has quit (Quit: Reconnecting)
- 18:08:31 [Zuzak]
- Zuzak (~zu@2001:ba8:1f1:f2f5::2) has joined #dig
- 18:08:31 [Zuzak]
- Zuzak has quit (Changing host)
- 18:08:31 [Zuzak]
- Zuzak (~zu@pdpc/supporter/professional/zuzak) has joined #dig
- 18:09:07 [Zuzak]
- Zuzak is now known as zuzak
- 19:01:21 [scor]
- scor has quit (Ping timeout: 276 seconds)
- 19:04:32 [jmvanel]
- jmvanel has quit (Ping timeout: 260 seconds)
- 19:05:05 [scor]
- scor (scor@drupal.org/user/52142/view) has joined #dig
- 19:08:23 [jmvanel]
- jmvanel (~jmvanel@199.0.88.79.rev.sfr.net) has joined #dig
- 19:48:02 [bblfish]
- bblfish (~bblfish@AAubervilliers-651-1-226-107.w86-198.abo.wanadoo.fr) has joined #dig
- 20:05:36 [cheater__]
- cheater__ (~cheater@p57AEA1E6.dip0.t-ipconnect.de) has joined #dig
- 20:17:28 [timbl]
- timbl has quit (Quit: timbl)
- 20:18:01 [RalphS]
- RalphS has quit ()
- 20:24:56 [deiu]
- deiu has quit (Quit: Leaving)
- 20:58:23 [timbl]
- timbl (~timbl@host86-146-148-68.range86-146.btcentralplus.com) has joined #dig
- 21:08:45 [jmvanel]
- jmvanel has quit (Ping timeout: 276 seconds)
- 21:16:43 [scor]
- scor has quit (Quit: scor)
- 21:24:09 [scor]
- scor (scor@drupal.org/user/52142/view) has joined #dig
- 21:24:31 [scor]
- scor has quit (Client Quit)
- 23:12:46 [bblfish]
- bblfish has quit (Remote host closed the connection)
- 23:21:15 [betehess]
- betehess has quit (Ping timeout: 245 seconds)