Time |
S |
Nick |
Message |
00:12 |
|
|
baweaver joined #rest |
00:27 |
|
|
fumanchu joined #rest |
00:57 |
|
|
shrink0r_ joined #rest |
02:29 |
|
|
sfisque joined #rest |
04:35 |
|
|
baweaver joined #rest |
05:55 |
|
|
angular_mike_ joined #rest |
07:34 |
|
|
ayan joined #rest |
07:38 |
|
|
interop_madness joined #rest |
07:42 |
|
|
Left_Turn joined #rest |
08:13 |
|
|
rosstuck joined #rest |
08:22 |
|
|
graste joined #rest |
09:08 |
|
|
quimrstorres joined #rest |
09:13 |
|
|
quimrstorres joined #rest |
09:21 |
|
|
shrink0r joined #rest |
09:39 |
|
|
dEPy joined #rest |
10:07 |
|
|
mezod joined #rest |
12:00 |
|
|
shrink0r_ joined #rest |
13:35 |
|
|
shrink0r joined #rest |
13:45 |
|
|
saml joined #rest |
13:54 |
|
|
quimrstorres joined #rest |
14:31 |
|
|
sfisque joined #rest |
14:59 |
|
|
quimrstorres joined #rest |
15:08 |
|
|
fladd joined #rest |
15:08 |
|
|
fladd left #rest |
15:38 |
|
|
NOPonPOP joined #rest |
15:39 |
|
NOPonPOP |
i feel like ive read all these links and developed some kind of rudamentary understanding of rest, except for the urls. the data structure i need to represent is a directory tree. inside the parent, there are child directories which contain databases i want to draw charrts for. eahc child directory can contain multiple databases, and there can be an unlimited number of child directories. all children sit beneath the same parent. |
15:40 |
|
NOPonPOP |
when i want to render charts for multiple children, i simply cant figure out what that URL should look like, a single chart should look like /containers/child_dir1/files/file1 |
15:40 |
|
NOPonPOP |
some people have said thats not restful at all |
15:40 |
|
NOPonPOP |
im not even sure how to respond to that, am i missing something completely here? |
15:45 |
|
ironChicken |
NOPonPOP: don't embed implied semantics in the URIs |
15:45 |
|
ironChicken |
URIs are just unique identifiers for things |
15:46 |
|
ironChicken |
if you have a hierarchy, then the response data should encode information about the structure of that hierachy by telling clients where they can go next |
15:46 |
|
ironChicken |
don't make clients guess the hierarchy from the URIs |
15:47 |
|
NOPonPOP |
can you give me an example of ho wyou might represent the same underlying data according to what youve said? |
15:48 |
|
NOPonPOP |
like are you saying the client does GET /containers/child_1/ |
15:48 |
|
NOPonPOP |
and the response tells him abou tfiles/file1,file2,file3 |
15:48 |
|
NOPonPOP |
which then triggers GET /containers/child_1/file1 |
15:49 |
|
NOPonPOP |
if thats the case, then it seems like for a given http request, i can only operate on a single item at a time. so ill have to dispatch a GET for each |
15:49 |
|
NOPonPOP |
and accumulate data on the client side |
15:50 |
|
NOPonPOP |
rather than sending data about multiple things in a single server response, if that makes sense |
15:50 |
|
ironChicken |
even better is the client does: GET /31d88326-f7f5-11e4-b90f-0021ccd2abe3 and the response says that this resource is a 'child', it has a parent called 'containers' (at URI /4cf2f808-f7f5-11e4-b1f5-0021ccd2abe3) and that is has a list of children (with URIs for each) |
15:50 |
|
NOPonPOP |
how about this last bit: 11:16 < NOPonPOP> if thats the case, then it seems like for a given http request, i can only operate on a single item at a time. so ill have to dispatch a GET for each |
15:50 |
|
NOPonPOP |
11:16 < NOPonPOP> and accumulate data on the client side |
15:50 |
|
NOPonPOP |
11:16 < NOPonPOP> rather than sending data about multiple things in a single server response, if that makes sense |
15:51 |
|
ironChicken |
i like the principal that each URI represents exactly one resource, and that each HTTP GET request is for exactly one URI |
15:51 |
|
NOPonPOP |
ok |
15:52 |
|
NOPonPOP |
that helps me actually |
15:52 |
|
ironChicken |
s/principal/principle/ |
15:52 |
|
NOPonPOP |
now back to your previous response |
15:52 |
|
NOPonPOP |
in that example, there is no nesting of those URIs |
15:52 |
|
NOPonPOP |
is that intentional? |
15:52 |
|
NOPonPOP |
you can see im having a hard time seperating the underlying directory structure im really representing |
15:52 |
|
NOPonPOP |
from the abstraction in this api |
15:53 |
|
NOPonPOP |
i can imagine a world in which i represent the directory stucture as one flat area, and then basically recreate the 'file system' using response data as you have described |
15:54 |
|
NOPonPOP |
but if i do that, it seems like im just moving the data structure around, but still using the same file tree data structure |
15:54 |
|
ironChicken |
NOPonPOP: yes, that's right, no nesting of URIs. nesting *implies* a hierarchy; instead you should make your hierarchy explicit in the semantics of the response |
15:57 |
|
NOPonPOP |
ironChicken: heh https://www.npmjs.com/package/restfs |
15:57 |
|
NOPonPOP |
i feel REST is kind of nebulous |
15:57 |
|
NOPonPOP |
i dont mean that as an insult or anything |
15:57 |
|
NOPonPOP |
seems very open to interpretation? |
15:58 |
|
NOPonPOP |
the biggest thing i wanted for my API was to decouple data and its representation |
15:58 |
|
NOPonPOP |
so i just return json data structures now that contain the raw data i want to plot |
15:58 |
|
NOPonPOP |
and let the client do the rendering of that |
15:59 |
|
NOPonPOP |
https://www-01.ibm.com/support/knowledgecenter/SSPT3X_3.0.0/com.ibm.swg.im.infosphere.biginsights.admin.doc/doc/admin_fileupload_rest_apis.html |
15:59 |
|
ironChicken |
nebulous is a good way of putting it |
16:01 |
|
|
alxbl joined #rest |
16:01 |
|
NOPonPOP |
i think since logically the way me and my users will think of this data is a directory tree (this is actually a front end to what used to be command line tools) it still makes sense for me to use that in the url. i think ill apply your principal of 1:1 request/object to solve multiple chart handling |
16:02 |
|
NOPonPOP |
thanks for discussing it with me |
16:02 |
|
fumanchu |
good choice |
16:04 |
|
ironChicken |
np :-) |
16:22 |
|
|
baweaver joined #rest |
16:29 |
|
|
sfisque joined #rest |
16:41 |
|
|
eschmidbauer joined #rest |
16:46 |
|
|
eschmidbauer left #rest |
17:30 |
|
|
quimrstorres joined #rest |
17:30 |
|
|
mezod joined #rest |
17:33 |
|
|
eschmidbauer joined #rest |
18:05 |
|
|
eschmidbauer joined #rest |
18:26 |
|
|
ayan_ joined #rest |
18:26 |
|
|
EnergyCoffee_ joined #rest |
18:26 |
|
|
prologic_ joined #rest |
18:27 |
|
|
pindonga joined #rest |
18:28 |
|
|
EnergyCoffee joined #rest |
18:29 |
|
|
EnergyCoffee joined #rest |
18:30 |
|
|
EnergyCoffee joined #rest |
18:32 |
|
|
EnergyCoffee joined #rest |
18:33 |
|
|
EnergyCoffee joined #rest |
18:34 |
|
|
EnergyCoffee joined #rest |
18:35 |
|
|
EnergyCoffee joined #rest |
18:37 |
|
|
EnergyCoffee joined #rest |
18:38 |
|
|
EnergyCoffee joined #rest |
18:39 |
|
|
EnergyCoffee joined #rest |
18:40 |
|
|
EnergyCoffee joined #rest |
18:42 |
|
|
EnergyCoffee joined #rest |
18:43 |
|
|
EnergyCoffee joined #rest |
18:44 |
|
|
EnergyCoffee joined #rest |
18:45 |
|
|
EnergyCoffee joined #rest |
18:47 |
|
|
EnergyCoffee joined #rest |
18:48 |
|
|
EnergyCoffee joined #rest |
18:49 |
|
|
EnergyCoffee joined #rest |
18:50 |
|
|
EnergyCoffee joined #rest |
18:52 |
|
|
EnergyCoffee joined #rest |
18:53 |
|
|
EnergyCoffee joined #rest |
18:54 |
|
|
EnergyCoffee joined #rest |
18:55 |
|
|
EnergyCoffee joined #rest |
18:57 |
|
|
EnergyCoffee joined #rest |
18:58 |
|
|
EnergyCoffee joined #rest |
18:59 |
|
|
EnergyCoffee joined #rest |
19:00 |
|
|
EnergyCoffee joined #rest |
19:02 |
|
|
EnergyCoffee joined #rest |
19:03 |
|
|
EnergyCoffee joined #rest |
19:34 |
|
|
shrink0r joined #rest |
20:00 |
|
|
quimrstorres joined #rest |
20:20 |
|
|
graste joined #rest |
20:23 |
|
|
quimrstorres joined #rest |
20:28 |
|
|
ayan joined #rest |
20:36 |
|
|
foist joined #rest |
20:46 |
|
|
quimrsto_ joined #rest |
21:06 |
|
|
quimrstorres joined #rest |
21:08 |
|
|
mezod joined #rest |
23:04 |
|
|
sfisque joined #rest |
23:20 |
|
|
prologic joined #rest |
23:39 |
|
|
shrink0r_ joined #rest |
23:40 |
|
|
shrink0r joined #rest |
23:42 |
|
|
huckleberry78 joined #rest |