#rest on irc.freenode.net 2011-11-29 (txt)

[00:07] DracoBlue (~Adium@ip-vlan-obckunde-02-217-66-60-14.pixelpark.net) joined #rest.
[00:21] technoweenie (~technowee@host-86-220-9-69.midco.net) left irc: Remote host closed the connection
[00:27] Split (~split@84.34.147.60) left irc: Ping timeout: 240 seconds
[00:29] Split (~split@84.34.147.60) joined #rest.
[00:49] ivanfi (~ivanfi@62.159.77.167) joined #rest.
[01:17] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) left irc: Ping timeout: 244 seconds
[01:46] DracoBlue (~Adium@ip-vlan-obckunde-02-217-66-60-14.pixelpark.net) left #rest.
[02:12] kennethreitz (~kennethre@c-24-127-96-129.hsd1.va.comcast.net) joined #rest.
[02:49] <mogsie> Time zones suck!
[02:49] <trygvis> true that
[02:49] <mogsie> I think mnot's desire to link stuff is along the same lines as the "convention" of using ISO-8601 to mark up dates.
[02:50] <mogsie> or a convention (by many APIs not based on Atom or HTML) to use a <link rel="xxx" href="yyy"/> construct.
[02:50] <mogsie> It's just a common way to do stuff. It's not a media type, or a mandated thing.
[02:51] <mogsie> It's a bit like we're all in agreement that "It's generally a good idea to use english language names for tags/keys".
[02:51] <trygvis> yeah, that's what I'm reading from his stuff too
[02:53] <mogsie> Face it. Who would coin a JSON message that looks like this: { "a" : "Mr.", "9" : "John", ")" : Smith" }
[02:54] <mogsie> or an XML structure: <dx><a>Mr.</a><q9>John</q9><z7>Smith</z7></dx>.
[02:55] <mogsie> English is used by convention. mnot is simply stating that there _are_ emerging ways of providing links in JSON.
[02:56] <mogsie> I'm starting to agree with him that it would be nice to be able to point to something like http://www.w3.org/TR/NOTE-datetime when I need to document how I mark up links in a JSON media type.
[02:59] <mogsie> Ah, mikekelly is alive :-)
[03:06] KevBurnsJr (~kevburnsj@c-76-126-10-63.hsd1.ca.comcast.net) joined #rest.
[03:08] KevBurnsJr (~kevburnsj@c-76-126-10-63.hsd1.ca.comcast.net) left irc: Client Quit
[03:18] Wombert (~Wombert@dslb-092-075-029-037.pools.arcor-ip.net) joined #rest.
[03:19] Hakon|mbp (~hakon1@84.202.136.151) joined #rest.
[03:21] <mikekelly> mogsie: just about
[03:21] <Ngarthl> mogsie: exactly.
[03:21] <mikekelly> provided I don't have a coranary trying to explain media type identifiers to people.
[03:22] <mikekelly> "leave json alone"
[03:22] <mikekelly> "I have, that's why I used a different media type identifier"
[03:22] <mikekelly> "no. leave json alone"
[03:22] <mikekelly> "right."
[03:23] <Ngarthl> Q: anyone with opensearch experience here?
[03:23] <mikekelly> stupid naming conventions and includeable 'features' are exactly what the problem with XML is
[03:23] <mikekelly> we don't need more documents to reference
[03:23] <mikekelly> we need more positive constraint
[03:24] <mikekelly> positive constriants are stuff like hal or collection+json
[03:24] <mikekelly> not airy-fairy documents
[03:24] <Ngarthl> trying to find a way of modelling facets and taxonomies in opensearch.
[03:25] <mikekelly> mogsie: looks like we're about to repeat the conversation about specific vs generic :|
[03:30] <mogsie> Ngarthl: We're probably going to do something like that soon...
[03:31] <mogsie> mikekelly: ah, I see your POV. I didn't percieve naming conventions as a problem with XML. But perhaps you're right. I hadn't thought of it that way.
[03:32] <mogsie> But wouldn't you agree that ISO-8601 is a useful agreed-upon "standard"?
[03:32] <Ngarthl> mogsie: http://pastie.org/2938140
[03:33] <mogsie> Ngarthl: Ouch :-)
[03:33] <Ngarthl> trying to come up with a better approach
[03:34] <mogsie> Nesting Query objects is that part of OS?
[03:34] <mogsie> s/objects/elements/ <g>
[03:34] <Ngarthl> probably not
[03:35] <mogsie> UAs might fill out filter-value but omit filter-name...
[03:36] <mogsie> Ngarthl: why not just have template="...?filter={f:filter?}
[03:36] <mogsie> and role="subset" f:filter="foo:bar" ?
[03:36] <Ngarthl> that might work
[03:37] <mogsie> You would't get to the "filter-name" part of the filter, but you might not need that.
[03:37] <Ngarthl> i had a reason why i did it like that. but i can't remember anymore
[03:38] <mogsie> In my proposal, a UI to select between different filters wouldn't know that "us" and "no" are in the same "group" as "foo" and "bar". Your pastie would allow that.
[03:38] <mogsie> It would be nice to let the UI know the difference between the two, though. Maybe the filter-name element could be defined but not referenced:
[03:39] <Ngarthl> yeah, sure. that can work
[03:39] <mogsie> <os:Query role="subset" f:filter-name="bleh" f:filter="bleh:foo" title="Foo"/>
[03:39] <mogsie> <os:Query role="subset" f:filter-name="bleh" f:filter="bleh:bar" title="Boo"/>
[03:39] <mogsie> The filter-name would be defined as a grouping mechanism to allow UA to e.g. provide drop-downs to select.
[03:40] <Ngarthl> then I wouldn't have to nest. yes, this will work
[03:42] <mogsie> mikekelly: Completely off topic, but HAL only provides link semantics (and embedding), and doesn't say anything about what <price>11.22</price> means...
[03:49] <mogsie> Ngarthl: To whom do I send my bill? ;-)
[03:49] <Ngarthl> mogsie: ;)
[03:49] <Ngarthl> you will get paid in the Afterlife :P
[03:54] <mikekelly> mogsie: nope it's not supposed to
[03:55] <mikekelly> you're meant to use the relations to figure that out
[03:55] <mikekelly> and no I don't care for schemas, but you could make the rels URIs and represent them with a schema doc
[03:55] <mikekelly> some people are doing that already
[03:57] <mikekelly> that could be part of HAL, but it's meant to be lean
[03:58] <mikekelly> tbh when it comes to generic types it's more about what you leave out than what you bring in
[03:58] <mikekelly> i.e. constraint
[03:58] <mikekelly> which is something atom got wrong, imo
[03:58] <mikekelly> it brought too much in and overspecified in other areas
[04:21] <mikekelly> how to make friends and influence people: https://plus.google.com/118254663306823054612/posts/452VFNUtW8e
[04:22] <mikekelly> /cc mamund darrelmiller ^
[04:22] <trygvis> I'm getting 404
[04:23] <mikekelly> ah balls it's a private thread
[04:26] <mikekelly> trygvis: http://oi43.tinypic.com/wtvssm.jpg
[04:26] <mikekelly> :)
[04:27] <trygvis> awesome
[04:54] bob_sage (~chatzilla@outbound4.ebay.com) left irc: Ping timeout: 258 seconds
[05:06] gchristensen (~gchristen@unaffiliated/grahamc) joined #rest.
[05:31] <mikekelly> mogsie: +1 on the frustration with time zones :)
[05:36] <darrelmiller> Is it just me, or is the G+ interface almost as confusing as facebook!
[05:38] <darrelmiller> mikekelly: I loved this quip "All this snarky 'pragmatism' is very trendy"
[05:47] <mikekelly> thanks :)
[05:48] <mikekelly> that is pretty much the sum total of the argument against hal or wahtver
[05:48] <mikekelly> "stop XML'ing up our JSON and stuff, maaaan"
[05:49] <darrelmiller> The fact that people are trying to use conventions to convey semantics means they really have missed the point of media types.
[05:54] <mikekelly> yeah but it's like
[05:54] <mikekelly> JSON and stuff
[05:54] <mikekelly> and like
[05:54] <mikekelly> that's not XML
[05:54] <mikekelly> so like
[05:54] <mikekelly> yeah
[05:54] <mikekelly> and stuff
[05:58] <mikekelly> darrelmiller: http://i.qkme.me/35dil4.jpg
[06:00] <darrelmiller> LOL
[06:11] <mikekelly> wow JJ just layed into HAL
[06:11] <mikekelly> couldn't be bothered to read his post so I just called him a 'gratuitously combative dickleberry'
[06:30] `0660 (olli@oosny.net) got netsplit.
[06:36] `0660 (olli@oosny.net) returned to #rest.
[06:38] <Ngarthl> mogsie: ah, it wasnt quite enough. I have trees i need to display somehow. perhaps adding another attribute will fix that
[06:41] <Ngarthl> although as XML is based on trees. maybe that is what I need to use somehow
[06:46] <trygvis> mmmm, xml
[07:16] <mikekelly> https://plus.google.com/118254663306823054612/posts/gFHBGiaTwmj
[07:17] grove_ (~grove@193.201.9.46.customer.cdi.no) joined #rest.
[07:18] <steveklabnik> wow
[07:18] <steveklabnik> what a fucking dickbag.
[07:18] <trygvis> hot damn
[07:18] <steveklabnik> like, totally disregarding the information he's trying to get across... that is... wow.
[07:20] grove (~grove@aggw006.cappelendamm.no) left irc: Ping timeout: 240 seconds
[07:20] grove_ -> grove
[07:21] <darrelmiller> First rule of REST, don't feed JJ.
[07:24] <steveklabnik> heh
[07:33] <mamund> mikekelly: LOL, you know you've made the big time when JJ decides to piss on your campfire.
[07:45] <mamund> i see steveklabnik is feeling frisky today.
[07:45] <darrelmiller> mikekelly: and with the amount of venom in that post, you must be doing something right.
[07:45] <steveklabnik> mamund: ;)
[07:46] <steveklabnik> i'll give anybody a single one of those kinds of comments.
[07:46] <mamund> you're a nice fellow
[07:46] <steveklabnik> it's due to my years and years of /., digg, reddit, hn, etc...
[07:46] <mamund> :)
[07:46] <steveklabnik> i'm used to people being that kind of asshole
[07:46] <steveklabnik> i dont think it really works often though.
[07:47] <darrelmiller> but JJ is a special member of our community. We love him really.
[07:47] <steveklabnik> heh.
[07:47] <mamund> well, you get good karma points anyway.
[07:48] mamund steps up to the podium, raps the gavel...
[07:48] <mikekelly> well he psoted it privately
[07:48] <mikekelly> her'es why it's public:
[07:48] <mikekelly> Mike Kelly - wow. might help the ongoing conversation if you could try and stop being such a gratuitously combative dickleberry all the time.
[07:48] <mikekelly> just a thought.
[07:48] <mikekelly> 14:11 - Edit
[07:48] <mikekelly> Mike Kelly - If you make this post public, I will happily discuss this with you.
[07:48] <mamund> my book "Building Hypermedia APIs..." is noew released and available.
[07:49] <darrelmiller> mamund: Congrats....
[07:49] <steveklabnik> mamund: grats! it's on my amazon list.
[07:49] <mamund> i just wanted to take a moment to say "thanks" to all here for all the great feedback and convo over the last year or so....
[07:49] <mikekelly> mamund: nice one, I shall be buying :)
[07:49] <mamund> lots of good ideas come from here and you guys help me quite a bit
[07:49] <trygvis> mamund: \o/
[07:49] <mikekelly> mamund: I think you should rebrand yourself Morpheus
[07:50] <mamund> i actually put a shout out for this channel in the aknowledgements, too.
[07:50] <steveklabnik> :)
[07:50] <mamund> :)
[07:50] <mamund> anyway, thanks guys, reall.
[07:50] <steveklabnik> mamund: did i tell you i'm going to rebrand my book along the 'hypermedia' lines, too?
[07:50] mamund steps down and goes back to his seat
[07:50] <mamund> steveklabnik: hey, that's cool!
[07:50] <steveklabnik> i've actually got a TOC together now... slowly but surely.
[07:50] mikekelly claps and wipes away tears
[07:50] <mamund> ROFL
[07:51] darrelmiller hands mikekelly a tissue
[07:51] <steveklabnik> i decided that I like this whole 'rebranding of rest' bit, plus, it's great marketing. "hey everyone, if you dont know why i'm renaming this book this, you _certainly_ need to buy my book."
[07:51] <mamund> :)
[07:52] <mamund> steveklabnik: i find that talking about the hypermedia details works better in some cases than talking about the REST generalities.
[07:52] <mamund> that's why i started using it
[07:52] steveklabnik nods
[07:52] <mamund> fwiw, i still think there is lots of room for writing about REST in a clear an familiar way
[07:52] <mamund> i've just not stepped up to that work.
[07:52] <steveklabnik> HATEOAS is a nice vector for explaining the difference between REST and RPC
[07:53] <mamund> yep, that's ture
[07:55] mikekelly shrugs
[07:56] <mikekelly> I always find assumign the perspective of an intermediary the best way of making sure you're making the most out of the protocol
[07:56] <mikekelly> at least for the heavy weight scalability stuff
[07:58] <steveklabnik> sure.
[07:58] <steveklabnik> but i know my audience. and they couldn't care less about that.
[07:58] <mikekelly> hypertext is mostly about reducing the surface area of the changeable parts of your application
[07:58] <steveklabnik> ... they will, hopefully.
[07:58] <steveklabnik> but it's not... gripping?
[07:58] <mikekelly> i.e. positive constraint
[07:58] <mamund> mikekelly: that's very interesting (the intermediary part)
[07:59] <mamund> i have been concentrating on thinking of hypermedia from the client side.
[07:59] <mamund> your POV is a good one, too.
[08:00] <mikekelly> yeah, although it's much less applicable to hypertext
[08:00] <mikekelly> other than fruity things like ESI or link cache invlidation or wahtever
[08:00] <mamund> yea, but i appreciate the POV.
[08:00] <mamund> esp. for the protocol-level things.
[08:04] ivanfi (~ivanfi@62.159.77.167) left #rest.
[08:17] <mikekelly> I just went thermonuclear on JJ
[08:17] <mikekelly> https://plus.google.com/118254663306823054612/posts/gFHBGiaTwmj
[08:19] <steveklabnik> heh
[08:22] <mamund> so now mikekelly is feeling frisky, too.
[08:22] <mikekelly> I will eat that chump for breakfast
[08:23] <mikekelly> I'm a fifth-dann internet arguing black-belt
[08:23] <mamund> pretty sure you'll get a bellyache if you do.
[08:23] <steveklabnik> lol
[08:24] <mikekelly> which basically means I don't even have to be right and I can castrate his e-dong in a relatively short space of time
[08:24] <mikekelly> if all else fails I'll just call him a dickleberry again and put him on ignore.
[08:24] <mikekelly> at least I'll be right about that.
[08:24] <mamund> mikekelly: pretty sure JJ feeds on negative energy - be careful<g>
[08:25] <mikekelly> :)
[08:31] Wombert (~Wombert@dslb-092-075-029-037.pools.arcor-ip.net) left irc: Quit: Wombert
[08:42] lmarburger (u1979@gateway/web/irccloud.com/x-qvjnreonyyfucmzi) left irc: Ping timeout: 248 seconds
[08:42] omarkj (u766@gateway/web/irccloud.com/x-xweivxrkcuptsfjf) left irc: Ping timeout: 244 seconds
[08:45] KevBurnsJr (~kevburnsj@c-76-126-10-63.hsd1.ca.comcast.net) joined #rest.
[08:56] KevBurnsJr (~kevburnsj@c-76-126-10-63.hsd1.ca.comcast.net) left irc:
[08:57] omarkj (u766@gateway/web/irccloud.com/x-auxevtmmdbozbrdf) joined #rest.
[09:09] <mikekelly> mamund: are you following this epic thread on google+ or have you given up by now ?
[09:10] <steveklabnik> lol
[09:11] <mikekelly> I'm now discussing with Julian thingy-maboby about media type extension
[09:11] <mikekelly> and he is confusing me.
[09:13] <mamund> i keep an eye on it, but haven't really been "following" it<g>
[09:16] lmarburger (u1979@gateway/web/irccloud.com/x-xvylkylwsdwrrjfw) joined #rest.
[09:27] KevBurnsJr (~KevBurnsJ@50.0.103.39) joined #rest.
[10:07] apok (~apok@c-107-3-149-167.hsd1.ca.comcast.net) joined #rest.
[10:12] sbanwart_ (~sbanwart@66.6.175.18) joined #rest.
[10:35] sbanwart_ (~sbanwart@66.6.175.18) left irc: Remote host closed the connection
[10:36] sbanwart (~sbanwart@66.6.175.18) joined #rest.
[11:02] <mamund> mikekelly: goes in for the kill on JJ
[11:07] <steveklabnik> heh
[11:22] apok_ (~apok@c-107-3-149-167.hsd1.ca.comcast.net) joined #rest.
[11:23] apok (~apok@c-107-3-149-167.hsd1.ca.comcast.net) left irc: Ping timeout: 248 seconds
[11:23] apok_ -> apok
[11:50] lmarburger (u1979@gateway/web/irccloud.com/x-xvylkylwsdwrrjfw) left irc: Ping timeout: 260 seconds
[11:59] lmarburger (u1979@gateway/web/irccloud.com/x-qfpnbszosuopadak) joined #rest.
[12:13] gchristensen (~gchristen@unaffiliated/grahamc) left irc: Ping timeout: 244 seconds
[12:24] kke (~kimmo@kuuma.ho.ro) joined #rest.
[12:29] kke (~kimmo@kuuma.ho.ro) left #rest.
[12:34] <mikekelly> mamund: I just went out with a bang on the JJ thread
[12:34] <mikekelly> https://plus.google.com/u/0/118254663306823054612/posts/gFHBGiaTwmj
[12:35] gchristensen (~gchristen@unaffiliated/grahamc) joined #rest.
[12:39] <mikekelly> "Mike, I don't believe anything, I look at what works and what doesn't. I have goals to make certain things simpler, in a given context, like building enterprise mobile applications."
[12:39] <mikekelly> EEHHH UUUHHHH. wrong again! You believe in a particular set of criteria for determining if something "works" or not. My hypothesis is that your criteria are somewhere in the region of "crap" and "astronomically crap".
[12:39] sbanwart (~sbanwart@66.6.175.18) left irc: Remote host closed the connection
[12:39] <Ngarthl> people really need to relax :)
[12:40] <mikekelly> btw that post is not at all serious, read the concluding paragraph -_-
[12:40] <mikekelly> Ngarthl: but yeah, you're right
[12:40] sbanwart (~sbanwart@66.6.175.18) joined #rest.
[12:45] Split (~split@84.34.147.60) left irc: Ping timeout: 240 seconds
[12:46] Split (~split@84.34.147.60) joined #rest.
[12:48] sbanwart (~sbanwart@66.6.175.18) left irc: Ping timeout: 260 seconds
[12:48] sbanwart (~sbanwart@66.6.175.18) joined #rest.
[12:53] sbanwart (~sbanwart@66.6.175.18) left irc: Ping timeout: 240 seconds
[13:01] mamund was away
[13:03] <mamund> mikekelly: yeah, it went pretty much as i figured.
[13:06] <mikekelly> :)
[13:07] <mikekelly> well it was fun, granted a complete waste of time, but still fun
[13:07] <mamund> LOL, i was about to post a similar line ('it was fun while it lasted')
[13:07] <mikekelly> and probably coloured my name with of bunch of people I really could not give a shit about
[13:07] <mamund> meh
[13:07] <mikekelly> c'est la vie, as they say
[13:07] <mamund> whatever
[13:08] bob_sage (~chatzilla@outbound4.ebay.com) joined #rest.
[13:08] <mikekelly> interesting comment on the HAL spec page
[13:08] <mikekelly> Hi Mike, a nit observation and a request. First up the nit observation: HTML is really not GUI based and is itself designed to be consumed by agents and not people. CSS (rendering info) + HTML (content+format) + agent -> gets consumed by people. But that's just a distraction. The request is that you relate this to OSLC for me.
[13:09] <mikekelly> I dont' agree with that nit observation
[13:10] <mikekelly> I know nothing about OSLC, can anyone fill me in on that ?
[13:10] <mamund> not a clue
[13:11] <mikekelly> mamund: any thoughts on that nit ?
[13:12] <mikekelly> html's interface is presented by agents, it's not consumed by them
[13:13] mikekelly shrugs at least not the application part of the interface anyway
[13:16] <mikekelly> bleh, is the comparison to html confusing to everyone else as well ?
[13:16] <mamund> LOL, i really don't know what that's about
[13:34] <trygvis> funny: http://s3.amundsen.com/amundsen-logo.tn.gif
[13:35] <mamund> LOL, how did you find that?
[13:35] <mamund> that's my upper right logo for branded gmail<g>
[13:36] <mamund> the "tag line" is a common phrase in our house; usually when someone does something dumb<g>
[13:38] <trygvis> I found it through your custom search. chrome shows it inline in the address bar
[13:38] <mamund> ha!
[13:38] <mamund> i think there's a larger one out there, too.
[13:38] <mamund> a bit pretentious, but we laugh
[13:39] <trygvis> are you related to him?
[13:39] <trygvis> it's still funny :)
[13:39] <mamund> yes, distant, tho.
[13:39] <mamund> uncle-kinda thing
[13:39] <mamund> my grandparents explained it to me when i was a kid and i forgot the deatils.
[13:39] <mamund> they used to have some dog traces and snowshoes in the house
[13:40] <mamund> and a few pics
[13:40] <mamund> tried doing a fmaily tree search on my own, but came up short.
[14:01] Wombert (~Wombert@dslb-092-075-029-037.pools.arcor-ip.net) joined #rest.
[14:04] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[14:11] apok_ (~apok@c-107-3-149-167.hsd1.ca.comcast.net) joined #rest.
[14:15] apok (~apok@c-107-3-149-167.hsd1.ca.comcast.net) left irc: Ping timeout: 260 seconds
[14:15] apok_ -> apok
[14:15] Hakon|mbp (~hakon1@84.202.136.151) got netsplit.
[14:15] Ngarthl (erlenha@158.38.48.10) got netsplit.
[14:17] Ngarthl (erlenha@158.38.48.10) returned to #rest.
[14:20] <bob_sage> hi I am writing a rest client on Jersey
[14:21] <bob_sage> and trying to access a (MediaType.APPLICATION_JSON)
[14:21] <bob_sage> the return type for the method is ArrayList<XYZ-Object>
[14:22] <bob_sage> I get a 500 error when trying to access the resource, defined like above\
[14:23] <bob_sage> "A message body writer for Java class java.util.ArrayList, and Java type java.util.List<>, and MIME media type application/json was not found"
[14:23] <bob_sage> is there somewhere I have to declare ArrayList<XYZ-Object> with an xml annotation?
[14:23] <bob_sage> and if so where?
[14:26] Hakon|mbp (~hakon1@84.202.136.151) got lost in the net-split.
[14:27] grove (~grove@193.201.9.46.customer.cdi.no) left irc: Quit: grove
[14:31] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Ping timeout: 248 seconds
[14:43] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) joined #rest.
[14:45] Wombert (~Wombert@dslb-092-075-029-037.pools.arcor-ip.net) left irc: Quit: Wombert
[15:04] gchristensen (~gchristen@unaffiliated/grahamc) left irc: Quit: Leaving...
[16:03] <fu-manchu> so I'm making a new media type, and it includes a reserved syntax for linking to a different resource which defines datatypes for it (somewhat like a DTD does). Q: is it legal/desirable to use SHOULD in "the representation obtained by dereferencing the URI *SHOULD* be a definition of the types..."?
[16:03] <fu-manchu> or is that too strong of a guarantee, or not strictly required for interop, or...?
[16:04] <fu-manchu> I should say "the representation (if any)" to allow for URI's that are not dereferencable
[16:05] <fu-manchu> I'm not hard-coding the URI to return a particular media type, but I still want to say what that other resource should be about
[16:06] <fu-manchu> should or (BCP 14) SHOULD?
[16:14] sbanwart (~sbanwart@99-177-126-136.lightspeed.bcvloh.sbcglobal.net) left irc: Ping timeout: 252 seconds
[16:59] <mikekelly> mamund: so, are you clear on what the big deal is with a media type identifier ?
[17:01] <mikekelly> because that's still pretty unclear to me right now
[17:14] darrelmiller (~darrelmil@70.24.176.11) left #rest.
[18:22] darrelmiller (~darrelmil@70.24.176.11) joined #rest.
[18:41] nuclearsandwich (~nuclearsa@173.247.193.198) joined #rest.
[18:55] apok (~apok@c-107-3-149-167.hsd1.ca.comcast.net) left irc: Quit: apok
[19:09] nuclearsandwich (~nuclearsa@173.247.193.198) left irc: Remote host closed the connection
[19:30] KevBurns (~kevburnsj@c-76-126-10-63.hsd1.ca.comcast.net) joined #rest.
[19:45] nuclearsandwich (~nuclearsa@99.22.229.14) joined #rest.
[20:35] nuclearsandwich (~nuclearsa@99.22.229.14) left irc: Remote host closed the connection
[20:48] apok (~apok@96.24.67.22) joined #rest.
[22:32] <Ngarthl> fu-manchu: I am doing something similar. although my definition is that it MUST be a definition, and it MUST have a certain media type
[22:33] <Ngarthl> bob_sage: try the jersey forum
[23:00] grove (~grove@193.201.9.46.customer.cdi.no) joined #rest.
[23:04] grove (~grove@193.201.9.46.customer.cdi.no) left irc: Client Quit
[23:08] <fu-manchu> Ngarthl: yeah, it just seems a bit too tightly coupled...if I update or replace the DTD media type, am I then forced to release a new version of the document media type? seems needlessly brittle
[23:09] <fu-manchu> perhaps I'm in a bit of a unique position, though, because my original media type can be seen itself as metadata for a tabular format
[23:10] <Ngarthl> well, in my design this won't be a problem. a UA might never know about the new fields, and that is quite allright
[23:10] <Ngarthl> there is a contract of MUST ignore of unknown fields
[23:11] <Ngarthl> updating the model type is then async of the document type
[23:12] <fu-manchu> yeah here too--we want to deploy new types without breaking existing consumers
[23:15] <Ngarthl> one thing about my design is that I have different model documents for different "instance" documents
[23:16] <Ngarthl> say you have a model Foo and an "instance" foo, that is different from the model of "Bar" and the "instance" of bar
[23:17] <Ngarthl> this especially works if most of the API is read only. If read-write, you will need to handle out-of-date UAs somehow
[23:18] <Ngarthl> defining default values might help. either in the service, or in the "model" document
[23:19] <fu-manchu> sounds fun
[23:19] fu-manchu toddles off to bed now
[23:19] <fu-manchu> thanks!
[23:19] <Ngarthl> np
[00:00] --- Wed Nov 30 2011