#rest on irc.freenode.net 2012-01-04 (txt)

[01:07] Hakon|mbp (~hakon1@130.82-134-26.bkkb.no) joined #rest.
[02:11] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) joined #rest.
[02:17] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) left #rest.
[02:41] aGHz1 (~Adium@modemcable153.0-178-173.mc.videotron.ca) joined #rest.
[02:41] aGHz (~Adium@modemcable153.0-178-173.mc.videotron.ca) left irc: Read error: Connection reset by peer
[02:47] mephju (~mephju@dslb-188-103-182-202.pools.arcor-ip.net) joined #rest.
[03:26] postmodern (~postmoder@c-71-237-178-98.hsd1.or.comcast.net) left irc: Quit: Leaving
[03:53] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) joined #rest.
[03:53] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) left #rest.
[04:37] <mikekelly> I'm actually quite proud of this post (in a sad way!), anyone got any thoughts: http://tech.groups.yahoo.com/group/rest-discuss/message/18355
[04:40] <mikekelly> mamund: do you know what the defined use of the term 'ontology' means in technology ?
[04:40] <mikekelly> cos I honestly don't have the foggiest.. and I use 'fog' deblierately there :)
[04:54] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Remote host closed the connection
[05:43] mephju (~mephju@dslb-188-103-182-202.pools.arcor-ip.net) left irc: Ping timeout: 252 seconds
[05:58] aGHz1 (~Adium@modemcable153.0-178-173.mc.videotron.ca) left irc: Quit: Leaving.
[06:19] sbanwart_ (~sbanwart@66.6.175.18) joined #rest.
[06:48] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) joined #rest.
[06:48] DracoBlue (~Adium@dslb-088-075-066-254.pools.arcor-ip.net) left #rest.
[07:03] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #rest.
[07:04] mamund slips in
[07:05] <mamund> mikekelly: http://semanticweb.org/wiki/Ontology
[07:12] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Read error: Operation timed out
[07:41] <mikekelly> thanks mamund
[07:41] <mikekelly> so I guess it's technical meaning is just as vague as it's actual meaning..
[07:42] <mikekelly> mamund: did you spot that Jan split off the discussion about forms into another thread ?
[07:42] <mamund> that's what i read
[07:42] <mikekelly> think you should get involved
[07:42] <mikekelly> :)
[07:42] <mamund> yeah. still unlikely to participate<g>
[07:42] <mikekelly> :P
[07:42] <mikekelly> party pooper.
[07:43] <mamund> "LOL, he said 'pooper'!"
[07:43] <mikekelly> ;)
[07:44] <mikekelly> mamund: did you read my post to Alexnader ?
[07:44] <mamund> hmm... last night or this AM?
[07:44] <mikekelly> this am the latest one I linked to earlier
[07:45] <mikekelly> it's nothing particularly new, just interested to know if it was clear or not
[07:45] <mamund> ahh
[07:45] <mamund> no, actually. just found it here
[07:45] <mamund> tl;dnr
[07:45] <mamund> >g?
[07:46] <mamund> yeah, ok, i see your point
[07:47] <mamund> fwiw, i think much of what's been covered on the list in the last fwe days is...
[07:47] <mikekelly> a load of horse shit ?
[07:47] <mamund> stuff about "how to map a problem domain to HTTP using media types" kind of discussion.
[07:48] <mamund> are els sufficient? must all the problem domain details be 'visible'? are forms required?, and so on.
[07:48] <mamund> s/els/rels
[07:48] <mikekelly> right
[07:48] <mikekelly> I think the term 'visible' is not defined well enough
[07:48] <mamund> is it 'best' to use custom media types? only 'generic' ones? etc. etc.
[07:49] <mamund> 'visible' in HTTP, 'self-describing' in REST, 'ontology' in SemWeb
[07:49] <mamund> all clear as mud
[07:50] <mikekelly> right
[07:50] <mikekelly> I'm surprised roy's dissertation didn't have to round those out
[07:50] <mikekelly> or does it ?
[07:50] <mikekelly> :S
[07:50] <mamund> not that i recall
[07:51] <mamund> honestly, the four "sub-constraints" for uniform interface...
[07:51] <mamund> and the things 'around' that (i.e. media types, etc.) are all way under-specified
[07:51] mikekelly nods
[07:52] <mamund> IMO, there is some big 'hand-wavying' in that area
[07:52] <mamund> so, yes, we need some help to clear up 'visible' and 'self-describing' in a way that is helpful to all.
[07:53] <mikekelly> I always try and clarify what I'm talking about with those terms and where the 'value' is in whatever point I'm making with it
[07:53] <mikekelly> I avoid using that terms for an 'ends' to some means
[07:53] <mamund> clarifying is good. what bugs me a bit (at times)...
[07:53] <mikekelly> s/that/those
[07:53] <mamund> is the varying versions
[07:54] <mamund> a consistent (context independent) definition that actually 'helps'
[07:54] <mamund> seems illusive
[07:56] <mikekelly> right
[07:57] <mikekelly> I was most interested in your take on the bit about using link rels
[07:57] <mikekelly> i.e. granularity and control over rels vs rev'ing a media type
[07:58] <mamund> yeah, i've been quiet on that subject for a reason...
[07:59] <mamund> i've not sorted out my thinking there.
[07:59] <mamund> i use this id/name/rel/class design for XHTML quite often
[07:59] <mamund> i, personally, can't get by w/ just rel, tho.
[07:59] <mamund> but even when i do that...
[08:00] <mamund> since i am applying almost always to just HTML, it seems a bit pointless<g>
[08:00] <mamund> since HTML uses the 'human' tokens of strings inside <a></a>, etc.
[08:00] <mamund> maybe if i actually worke dup some implementations using HAL, i'd get a better picture.
[08:01] <mamund> (it's on my list, of course<g>)
[08:02] <mamund> git branch
[08:02] <mamund> RFOL!
[08:03] mamund has too many windows open on hist desktop this AM
[08:08] aGHz (~Adium@modemcable028.248-23-96.mc.videotron.ca) joined #rest.
[08:08] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #rest.
[08:08] <mikekelly> !!
[08:10] <mikekelly> mamund: so assume a really simple app with an entry point a link rel "http://example.com/rels/foo"; from the entry point
[08:14] <mamund> ok
[08:15] <mikekelly> and that link rel tells the client the important info it needs is the innerhtml of div#content
[08:15] <mikekelly> trivial application
[08:15] <mikekelly> but there's no requirement for any kind of profile
[08:15] <mamund> why is that?
[08:16] <mikekelly> the client is following the foo rel - so it's expectations are set
[08:16] <mamund> how are they set?
[08:16] <mikekelly> by following the foo rel
[08:16] <mikekelly> from an entry point
[08:16] <mamund> LOL, possibly met, but not set
[08:16] <mamund> anyway, continue
[08:17] <mikekelly> why not set ?
[08:17] <mamund> 'expecting' is not created by following
[08:17] <mamund> i have no idea (in this convo) _what_ to expect
[08:17] <mamund> that's my Q at this time
[08:18] <mamund> i see that rel has some expectations associated
[08:18] <mikekelly> well you followed a link w/ rel="http://example.com/rels/foo";
[08:18] <mamund> but i don't see where these associations are documented or how that documentation is translated into tangible details in the client code
[08:19] <mamund> following can (possibly) _fulfill_ my expections (hey, i expect X, let's follow this link and see if that's what i get')
[08:19] <mikekelly> they are documented by correspodnign doc at the rel's URI
[08:19] <mamund> aha!
[08:19] <mamund> let's see that
[08:19] <mamund> why is that not a 'profile' in your mind?
[08:19] <mikekelly> *or* you can use tokens and fill a URI template that you're given at the entry point
[08:19] <mamund> what is contained there?
[08:20] <mikekelly> so the entry point could give you link rel="documentation-base-uri" href="http://example.com/rels/";
[08:20] <mikekelly> so you can just use tokens
[08:20] <mamund> ok
[08:21] <mikekelly> so what's contained there is a description of the target resource and the valid forms of request that can be made to it
[08:22] <mamund> yeah, let's see one
[08:22] <mikekelly> you want to see API docs ?
[08:22] <mamund> this rel document stuff, yeah
[08:23] <mikekelly> I don't have a public example because all the stuff I've worked on so far has been cloud computing companies who don't want to share their API design with the public
[08:23] <mamund> yeah, understood
[08:23] <mikekelly> not how I would do it, btw
[08:23] <mikekelly> but there you go
[08:23] <mamund> this is a stumbler for me
[08:24] <mikekelly> wait I can show you something that's almost there
[08:24] <mamund> that'd be cool
[08:24] <mamund> doesnt have to be today
[08:24] <mamund> my POV here is that what you'll show me is essentially a profile
[08:25] <mamund> but i'd like to see it to clear that up
[08:25] <mamund> if you'd like to share privately, you can email me
[08:25] <mikekelly> http://code.google.com/p/huddle-apis/wiki/Actor
[08:25] mamund checks link
[08:25] <mikekelly> mamund: yeah I'm not sure there is a huge difference between what you call a profile and what I call rel documentation
[08:26] <mamund> !
[08:26] <mikekelly> but the difference is that the rel is in the preceding 'state'
[08:26] <mamund> i suspect this example is incomplete
[08:26] <mikekelly> it's projected by the client, it's not burnt into the resource
[08:26] <mamund> since it only describes the rel, not the '
[08:26] <mamund> 'expections' assocated w/ that rel
[08:27] <mikekelly> well effectively that page is the expectation for a given rel
[08:27] <mamund> what page?
[08:27] <mikekelly> i.e. it's what you would get if you followed a rel="actor" link
[08:27] <mikekelly> that google page
[08:27] <mamund> naw, that's meta
[08:27] <mamund> unless i _totall_ miss the point
[08:28] <mikekelly> yeah I think I'm not talking in terms that make sense
[08:28] <mikekelly> :)
[08:29] <mamund> anyway, my curiosity here is how i can "code for the rel" the way i currently "code for the media type"
[08:29] <mamund> that's all
[08:29] <mamund> this is a 'stumbler' for me because i am not yet clear on the details.
[08:29] <mamund> not because it think it's not possible/valid, etc.
[08:29] <mamund> i just need some tangibles to work w/
[08:30] <mamund> to get real crazy on you here...
[08:30] <mamund> i _suspect_ you are describing an entirely diff way to code hypermedia;
[08:30] <mamund> not via a message format, but just a 'semantic token' set
[08:31] <mamund> this set might apply to all types of message formats
[08:31] <mikekelly> well you could do that
[08:31] <mamund> i might be way overstating the point here, tho.
[08:31] <mikekelly> but it's more complicated, personally I just pick a generic media type and describe the app in terms of that type
[08:31] <mamund> anyway, _if_ you have a format-agnostic way to describe problem domain semantics using a coherent set of URI decorators...
[08:31] <mamund> i'm very much intrigued<g>
[08:32] <mamund> well, see, that's the area i want to learn more about.
[08:32] <mamund> " I just pick a generic
[08:32] <mamund> media type and describe the app in terms of that type"
[08:32] <mikekelly> effectively that is what you get but in practice I always end up mixing the documentation up with a particular generic media type
[08:32] <mamund> where's the details? docs? how can i duplicate that? how can i code my client for that, etc.
[08:32] <mamund> yep - mixing is common
[08:33] <mamund> anyway, it's the distillation of this thing you talk about that interests me right now
[08:33] <mamund> ok, got some billings to get out the door (i.e. i should really _code_ now)<g>
[08:34] <mamund> will background this space for a while as i earn some duckets
[08:34] <mikekelly> "if you follow an xyz relation you will find a resource with <load of properties> and <various other link relations> and <other embedded resources>"
[08:34] <mikekelly> oh look each of those things has a place in HAL
[08:34] <mikekelly> :)
[08:35] <mamund> expand the quote into something that allows me to code my client and i'm cookin'
[08:36] <mikekelly> if it's terms of hal that's pretty trivial right ?
[08:36] <mikekelly> the properties are just the bits of the representation that are not links or embedded resources..
[08:36] <mikekelly> i.e. it's just an xml or json document
[08:37] <mamund> here's a suggestion (my, i'm pushy today...)
[08:37] <mikekelly> the maze thing ?
[08:38] <mamund> whip up a 'trivial' app (server) that emits HAL
[08:38] <mamund> then send me the 'docs' for creating a client for that server
[08:38] <mikekelly> alright
[08:38] <mamund> that will force me to write a client using HAL and i bet that clears the cobwebs
[08:38] <mamund> (or gives me a chance to badger you w/ inane Qs!)
[08:42] <mikekelly> sounds awesome :D
[08:42] <mamund> kewl
[08:42] <whartung> o/
[08:43] <mamund> btw, i tried out c9/vim last night...
[08:43] <mamund> 10dd worked for me in Chromium
[08:43] <mamund> i could cut a block, move, then 'p' it in the new location
[08:44] <mamund> i didn't try _named_ blocks, tho
[08:45] <mamund> whartung: \o
[10:41] Wombert (~Wombert@dslb-092-075-003-110.pools.arcor-ip.net) joined #rest.
[10:56] mamund tosses a couple logs on the "forms are evil" meme/thread
[13:28] KevBurnsJr (~KevBurnsJ@50.0.103.39) joined #rest.
[13:43] sbanwart_ (~sbanwart@66.6.175.18) left irc: Ping timeout: 252 seconds
[14:01] KevBurnsJr (~KevBurnsJ@50.0.103.39) left irc: Ping timeout: 240 seconds
[14:05] Hakon|mbp (~hakon1@130.82-134-26.bkkb.no) left irc: Quit: Leaving...
[14:10] KevBurnsJr (~KevBurnsJ@50.0.103.39) joined #rest.
[14:26] aGHz (~Adium@modemcable028.248-23-96.mc.videotron.ca) left irc: Quit: Leaving.
[15:06] mamund is done now; peace-out!
[15:08] aGHz (~Adium@74.198.165.35) joined #rest.
[15:32] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Quit: Textual IRC Client: http://www.textualapp.com/
[15:53] KevBurnsJr (~KevBurnsJ@50.0.103.39) left irc:
[16:53] sbanwart_ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[17:39] sbanwart__ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[17:40] sbanwart_ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Ping timeout: 252 seconds
[17:40] twilliams (~twilliams@apache/committer/twilliams) joined #rest.
[17:42] aGHz (~Adium@74.198.165.35) left irc: Quit: Leaving.
[17:46] Jarda (~jarda@o178.nor.fi) left irc: Ping timeout: 268 seconds
[17:50] Jarda (~jarda@o178.nor.fi) joined #rest.
[17:56] aGHz (~Adium@74.198.165.75) joined #rest.
[18:03] Jarda (~jarda@o178.nor.fi) left irc: Ping timeout: 268 seconds
[18:05] sbanwart__ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Remote host closed the connection
[18:07] sbanwart__ (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[18:40] sbanwart__ -> sbanwart
[18:50] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Remote host closed the connection
[18:50] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[19:03] <twilliams> geez... where is bigbluehat when i need 'em?
[19:04] <twilliams> good job, mikekelly: btw, for stayin on message:)
[19:06] <twilliams> that was, https://twitter.com/#!/mikekelly85/status/154503526104563712 btw:)
[19:06] <twilliams> love you guys, happy new year too!
[19:22] Jarda (~jarda@o178.nor.fi) joined #rest.
[19:28] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Read error: Operation timed out
[19:34] Jarda (~jarda@o178.nor.fi) left irc: Ping timeout: 268 seconds
[19:41] Jarda (~jarda@o178.nor.fi) joined #rest.
[19:57] Wombert (~Wombert@dslb-092-075-003-110.pools.arcor-ip.net) left irc: Quit: Wombert
[20:57] aGHz (~Adium@74.198.165.75) left irc: Quit: Leaving.
[23:52] Wombert (~Wombert@dslb-092-075-003-110.pools.arcor-ip.net) joined #rest.
[00:00] --- Thu Jan 5 2012