greptilian logo

IRC log for #rest, 2015-03-07

https://trygvis.io/rest-wiki/

| Channels | #rest index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

All times shown according to UTC.

Time S Nick Message
00:13 warehouse13 joined #rest
00:17 jsys You know what makes me instantly doubt something. When its supporters can't point out its weak sides.
00:18 jsys Everything needs to have some weak sides, because no solution is best at everything.
00:19 shrink0r joined #rest
00:24 shrink0r_ joined #rest
01:02 pdurbin jsys: you're reminding me of https://sukhbinder.wordpress.com/2015/02/25/python-over-other-languages/
01:06 pdurbin no one likes a fanboy
01:13 shrink0r joined #rest
01:32 jcromartie joined #rest
01:47 Duqu joined #rest
02:09 pgicxplzs joined #rest
03:55 SlippinJimmy joined #rest
04:20 shrink0r_ joined #rest
04:29 lemur joined #rest
04:59 fuzzyhorns joined #rest
05:00 jgornick joined #rest
06:04 shrink0r joined #rest
06:16 fumanchu_ joined #rest
09:22 _ollie joined #rest
09:40 fuzzyhorns joined #rest
09:59 Left_Turn joined #rest
10:13 azr joined #rest
10:57 cartr joined #rest
10:57 cartr Is this channel about Fielding's REST, or
10:57 cartr about pretty-URL-and-CRUD-over-HTTP REST.
11:21 azr joined #rest
11:38 pdurbin cartr: the latter needs a better marketing department
11:40 cartr pdurbin: :D
11:41 cartr I have a question. Why is it considered important to utilize PUT PATCH DELETE if Fielding doesn't even mention any of them? He only mentions PUT once to basically say it's not useful for write through caching.
11:41 cartr The focus is on GET/HEAD it seems, vs the rest (POST)
11:45 pdurbin what would one use instead of DELETE?
11:49 Left_Turn joined #rest
11:54 cartr pdurbin: POST
11:55 cartr pdurbin: POST is just any representation that may have some undefined effect on the server resources
12:00 pdurbin huh. I guess I'm just so used to DELETE. makes sense though
12:12 trygvis DELETE is idempotent, POST is not
12:13 trygvis cartr: stuff has evolved a lot since the thesis was written so you have more options now
12:15 cartr trygvis: it has evolved conceptually but it's unproven.
12:16 trygvis ok
12:16 cartr trygvis: the paper was built upon observable effects from the web.
12:16 cartr Since then it seems the idea has taken precedence over result.
12:16 cartr It happens sometimes.
12:16 cartr It's how XHTML2 failed
12:16 trygvis oh fuck, you're jsys
12:17 * trygvis out
12:18 shrink0r joined #rest
12:22 asdf` err, can someone explain this one plase, from http://roy.gbiv.com/untangled/2008/rest-apis-must-be-hypertext-driven
12:22 asdf` "In general, any protocol element that uses a URI for identification must allow any URI scheme to be used for the sake of that identification."
12:23 asdf` 'for identification'? why would anything about the uri matter, then, as opposed to being just an opaque identifier?
12:23 asdf` or am i just reading that wrong. I've read the sentence like 30 times and i have no idea what it's saying really
12:29 pdurbin asdf`: maybe it means that if you use "http://..." for identification you should also allow "ftp://..." for identification
12:33 asdf` hmm, i'm just not sure why would that ever not be the case
12:33 pdurbin dunno
12:33 pdurbin asdf`: so are you also cartr and jsys?
12:34 pdurbin trygvis: I didn't mind jsys
12:34 asdf` say my client can't use ftp, so then i have to explicitly make it so it rejects ftp://, but this is obviously a ridiculously stupid idea, so i assume saying that you shouldn't do this wasn't what was meant there
12:35 asdf` but maybe it is that simple, huh
12:35 asdf` pdurbin, thanks
12:35 asdf` and no, i don't use any other nick on freenode
12:39 azr joined #rest
13:11 Andre-B joined #rest
13:20 jsys joined #rest
13:20 jsys It's me, jsys
13:21 vanHoesel joined #rest
13:24 _ollie joined #rest
13:27 pdurbin welcome back
13:28 pdurbin asdf`: I'm really not sure.
13:28 jsys So I've been reading about CORBA, and turns out CORBA is RESTful
13:28 jsys go figure.
13:28 jsys :P
13:28 jsys Yet people don't like it.
13:31 pdurbin People seem to hate CORBA with the fire of a thousand suns. I've never used it.
13:35 jsys pdurbin: people invented IP because it was too hard to roll your own network topology and routing all the time.
13:35 jsys pdurbin: then they invented TCP because rolling your own reliability layer over IP was hard
13:35 jsys pdurbin: then they invented HTTP for sharing web pages, that's still not entering the game
13:36 jsys So then someone said it's too hard to roll out your own remoting on TCP so CORBA was born
13:36 jsys Nope SOAP
13:36 jsys Nope RMI
13:36 jsys Nope DCE
13:36 jsys And nope again, and they settled on the *hypertext* protocol??
13:37 jsys It doesn't seem like the final point on the line at all
13:37 jsys It's just a temporary thing.
13:37 jsys joined #rest
13:37 jsys ah damn I got disconnected while explaining my grnd theory
13:38 pdurbin evolution continues, sounds like
13:40 jsys pdurbin: there's a common pattern, I see no evolution here. I just see people piggybacking on whatever is popular at the time for their wire protocols.
13:41 jsys pdurbin: C++ and so on were the norm at the time and CORBA was modeled around that.
13:41 jsys pdurbin: at the time of Java XML was really popular to SOAP used XML for the wire format
13:41 jsys pdurbin: now HTTP is popular because - web sites... so let's put everything on HTTP
13:42 jsys It gets marketed, inexplicably, as a series of silver bullets that should solve everything but turn out the devil
13:42 jsys But in reality... it's neither a silver bullet nor the devil.
13:42 jsys It's piggybacking on something popular for interop.
13:42 jsys So pedestrian and unsophisicated, that it kinda makes all the snobs arguing over the virtues of this or that approach kinda funny
13:43 pdurbin I think a lot of us are just trying to get stuff done. Trying to build useful software. And if people want HTTP and JSON in 2015, that's fine.
13:44 locks getting stuff done considered harmful
13:44 jsys pdurbin: if it was seen like this, it'd be quite refreshing.
13:44 jsys I think in the end as a whole we do get stuff done
13:44 jsys But the internal debate is about stuff that... really doesn't matter.
13:44 jsys It's I think quite interesting we end up doing the practical thing without understanding why.
13:45 jsys In that sense we're not different than animals acting by instinct and just ending up liking whatever seems to work, without knowing why.
13:45 jsys Except unlike animals we're very good at retroactively rationalizing it.
13:46 jsys I mean REST is a bit like retroactive rationalization for the success of the web.
13:46 locks what
13:46 jsys locks: :)
13:47 jsys I'm just sayin' it like it is
13:47 pdurbin People who use APIs I help develop don't scream at me that they are not RESTful enough. They seem happy enough with them. They're able to get work done.
13:47 locks 1:46 PM <jsys> I mean REST is a bit like retroactive rationalization for the success of the web.
13:47 locks what does this even mean
13:47 _ollie the signal to noise ration has dramatically dropped in this channel since a few days ago…
13:47 jsys _ollie: I'm uhmm... sorry :P? Was there any signal before?
13:47 _ollie yes
13:48 pdurbin It's a nice channel. Good signal, generally.
13:48 jsys locks: web is successful, paper gets written why; It gets accepted as truth, but might as well not be.
13:48 _ollie so: who cares?
13:49 jsys _ollie: I won't come here anymore after I quit in a few mins, allowing you all to swim in your signal all day unpolluted by my noise
13:49 jsys :D
13:49 _ollie seems you're the only one complaining about too much meta-discussion in an endless meta-discussion
13:49 pdurbin _ollie: are you more interested in the nuts and bolts of getting stuff done than the philosophy behind it all?
13:49 locks that's… wow
13:50 pdurbin _ollie: heh. touche
13:50 _ollie pdurbin: you can only get the former having understood he latter
13:51 pdurbin yeah
13:51 jsys locks: so are you just inexplicably shocked I said this or you have some argument against it
13:52 locks jsys: http://en.wikipedia.org/wiki/Not_even_wrong
13:52 jsys I feel slightly uncomfortable, like I feel when I'm telling a religious joke and realize someone in the group is religious and staring at me with this expression of shock
13:53 pdurbin jsys: maybe you should write up all your thoughts into a blog post
13:53 pdurbin lay out your argument
13:53 jsys locks: everything I said is factual. The web was there, the paper got written, and it used REST to implicitly explain the success of the web due to its architectural style.
13:54 * _ollie yawns
13:54 pdurbin jsys: what would the title of your blog post be?
13:54 locks jsys: ok
13:54 _ollie "parroting the obvious"?
13:54 jsys pdurbin: "Correct architecture by mistake"
13:55 jsys pdurbin: it'll be about our history of making simple, practical choices, while having no clue in our internal debates why
13:55 _ollie that's better than wrong architecture by purpose
13:57 jsys The cumulative natural selection effect or making choices for the wrong reason is that some of the choices turn out right by accident, and then we figure out reasons retroactively why, so we can feel good about understanding our choices.
13:57 pdurbin jsys: I'd read your blog post.
13:59 jsys And putting things on URL resources can be explained as simple as "it's popular right now, everyone has/can write an HTTP client" instead of writing a big paper about it. Because the other effects (caching, distribution, intermediaries) occur naturally in any such architecture over time, without predetermined intent. It happens naturally due to for
13:59 jsys ces in the system pressuring the parties into implementing solutions like that.
13:59 jsys It occurs again and again every time a new system like this is developed.
14:00 locks it's kinda like cooking
14:00 jsys pdurbin: I suspect you just want me to feel better about my rants ;) But it's fine
14:01 locks we know a tomato is a fruit
14:01 locks but we put it in salads
14:01 _ollie oh, jeez… I need to go…
14:01 jsys _ollie: when you're back I won't be here.
14:01 jsys _ollie: get some beer to celebrate
14:01 pdurbin jsys: I love a good rant but I like reading it as a single cogent argument. It's less appealing spread over multiple days in chat logs.
14:01 _ollie that's very kind of you, thanks
14:01 locks don't make promises you can't keep
14:02 jsys locks: nah this is my pattern of IRC usage. I come like a storm, last few days and then disappear for months
14:02 _ollie that's how hot air usually works, yes…
14:03 jsys Anyway I'm goin :/
14:03 jsys Forevah
14:03 jsys :P
14:18 _ollie1 joined #rest
14:53 fuzzyhorns jsys: one thing that always bugs me about natural selection metaphors is we ignore how wasteful such a strategy really is
14:54 fuzzyhorns also lol _ollie
14:55 fuzzyhorns jsys: im curious what you think hypermedia actually _is_
15:06 jcromartie joined #rest
15:33 jcromartie joined #rest
15:43 graste joined #rest
16:34 SlippinJimmy joined #rest
16:37 _ollie joined #rest
17:56 begriffs joined #rest
17:57 azr joined #rest
18:22 _ollie joined #rest
18:37 SlippinJimmy joined #rest
18:50 lemur joined #rest
19:15 vanHoesel joined #rest
19:35 ||Zz|| joined #rest
19:37 ||Zz|| Hi all. I was looking for some API REST cool for demonstrating/teaching purposes. I used to use Spotify MetaData API, and Twitter's one, but both migrated to new APIs which require registration and OAuth2. This adds unnecessary extra complexity to my teaching goal
19:37 ||Zz|| I know about Google's Geocoding API, which is open and do not require any kind of auth or even API_KEY
19:38 ||Zz|| do you know some other in this line?
19:41 pdurbin |Zz|: parts of github's api don't require auth
19:44 |Zz| pdurbin: thank you, I'll check that
19:45 pdurbin oh sure
20:00 vanHoesel joined #rest
20:28 SlippinJimmy So I need an ability to modify an existing resource by setting a rejected state to true. I also need the ability to optionally specify a reason for the rejection which is a one-time message that will not be stored with the resource. Normally I’ve been using a PATCH to updating an existing resource with partial data, but is this a case for using POST?
20:54 azr joined #rest
20:59 shrink0r joined #rest
21:18 Andre-B joined #rest
21:50 vanHoesel joined #rest
22:34 lemur joined #rest
22:41 huckleberry78 joined #rest

| Channels | #rest index | Today | | Search | Google Search | Plain-Text | plain, newest first | summary

https://trygvis.io/rest-wiki/