| Time |
S |
Nick |
Message |
| 00:50 |
|
|
Mxyzpltk_ joined #rest |
| 02:09 |
|
|
Mxyzpltk_ joined #rest |
| 02:50 |
|
|
ybit joined #rest |
| 02:50 |
|
ybit |
oh cool |
| 02:50 |
|
ybit |
glad to see this channel is a thing :) |
| 02:50 |
|
ybit |
i've a question for everyone here |
| 02:50 |
|
ybit |
say you have a resource, and that resource has associated with it several other resources |
| 02:51 |
|
ybit |
which is better to do |
| 02:51 |
|
ybit |
{name: "foo", stuff: "http://api.example.com/stuff/343439"} |
| 02:51 |
|
ybit |
or |
| 02:51 |
|
ybit |
{name: "foo", stuff: [{name: bar, baz: "quz"} ... ]} |
| 02:51 |
|
ybit |
i kind of feel that the first one is more in tune with HATEOS |
| 02:52 |
|
ybit |
hateoas rather |
| 02:52 |
|
ybit |
i would love to hear some opinions on this |
| 02:54 |
|
pdurbin |
so in the first example the several other resources can be found by following the link |
| 02:56 |
|
ybit |
pdurbin: that is correct |
| 02:57 |
|
pdurbin |
and the second example has no links |
| 03:00 |
|
ybit |
pdurbin: well.. |
| 03:00 |
|
ybit |
there will be links for both |
| 03:02 |
|
ybit |
_links: {self: { title: "foo", href: "api.example.com/place/34334"}, parent: {... }}, {_items: { ... long list } } |
| 03:02 |
|
ybit |
that's the second example redone |
| 03:03 |
|
ybit |
rather.. {_meta: {}, _links: {self: { title: "foo", href: "api.example.com/place/34334"}, parent: {... }}, {_items: { ... long list } }} |
| 03:03 |
|
ybit |
vs |
| 03:04 |
|
ybit |
{_meta: {}, _links: {self: { title: "foo", href: "api.example.com/place/34334"}, parent: {... }, items: "http://api.example.com/stuff/34343" }} |
| 03:05 |
|
ybit |
i guess now that i put it like this.. |
| 03:05 |
|
ybit |
well... still not sure |
| 03:09 |
|
pdurbin |
do you see examples of both out there? in various APIs |
| 03:13 |
|
ybit |
pdurbin: yeah |
| 03:14 |
|
ybit |
maybe the curlies are confusing everyone |
| 03:14 |
|
ybit |
it's easier to keep track of curlies when you can put them on more than one line :) |
| 03:14 |
|
pdurbin |
yeah, pretty printed in a gist would be easier on the eyes |
| 03:14 |
|
ybit |
in the first _links contains self, parent and items |
| 03:14 |
|
ybit |
in the second, _links only contains title and aprent |
| 03:14 |
|
ybit |
_items should be at the same level as _meta and _links |
| 03:15 |
|
ybit |
..in the second |
| 03:15 |
|
|
systmkor joined #rest |
| 03:16 |
|
ybit |
/win 13 |
| 04:09 |
|
fumanchu |
ybit: embed the attributes the user/agent needs to select the right one(s) from the list. put all the other attributes across a link. In almost every case, the former are more properly attributes of the collection because they categorize the entity in that collection, rather than attributes of the entity itself. |
| 04:09 |
|
|
Mxyzpltk_ joined #rest |
| 04:11 |
|
|
systmkor1 joined #rest |
| 04:44 |
|
|
systmkor2 joined #rest |
| 05:41 |
|
|
lemur joined #rest |
| 05:42 |
|
lemur |
So I'm getting increasingly curious about something. Anyone care for a muse trip on APIs? |
| 05:43 |
|
lemur |
SOA is becoming more and more popular as a means of distributing application concerns, and with it the need to define clients. |
| 05:44 |
|
lemur |
A RESTful API is easy to predict in nature, and if there's a manifest request that can be made to get the sites schema layout, you can effectively define the entire clients connection logic |
| 05:45 |
|
lemur |
So effectively I'm musing on taking say a Rails API and having its manifest used to automatically generate the hooks for Angular, iOS, Droid, and any other framework at request. |
| 05:46 |
|
lemur |
Does this seem feasible? It just seems like so much of this is just straight boilerplate code that could be abstracted out. |
| 05:49 |
|
|
ramsey joined #rest |
| 06:10 |
|
|
Mxyzpltk_ joined #rest |
| 06:38 |
|
|
_ollie joined #rest |
| 06:51 |
|
|
rosstuck joined #rest |
| 07:30 |
|
|
_ollie left #rest |
| 07:30 |
|
|
_ollie joined #rest |
| 08:10 |
|
|
graste joined #rest |
| 08:11 |
|
|
Mxyzpltk_ joined #rest |
| 08:43 |
|
|
proteusguy joined #rest |
| 08:43 |
|
|
martinfilliau joined #rest |
| 08:51 |
|
|
interop_madness joined #rest |
| 09:08 |
|
|
proteusguy joined #rest |
| 09:16 |
|
|
interop_madness joined #rest |
| 09:33 |
|
|
Left_Turn joined #rest |
| 10:12 |
|
|
Mxyzpltk_ joined #rest |
| 10:27 |
|
|
shrink0r joined #rest |
| 10:37 |
|
|
_ollie joined #rest |
| 10:50 |
|
pdurbin |
lemur: give it a try! |
| 10:59 |
|
|
Andre-B joined #rest |
| 12:10 |
|
|
rosstuck joined #rest |
| 12:13 |
|
|
Mxyzpltk_ joined #rest |
| 12:50 |
|
Andre-B |
just saved $65 buying this video http://player.oreilly.com/videos/9781491900130 with coupon code: WCYAZ |
| 13:39 |
|
|
Mxyzpltk_ joined #rest |
| 13:39 |
|
|
foist joined #rest |
| 14:10 |
|
|
saml joined #rest |
| 14:44 |
|
|
fumanchu joined #rest |
| 16:00 |
|
|
Mxyzpltk joined #rest |
| 16:02 |
|
|
shrink0r1 joined #rest |
| 16:14 |
|
ybit |
thanks for the feedback pdurbin and fumanchu |
| 16:22 |
|
|
Mxyzpltk joined #rest |
| 16:27 |
|
Andre-B |
there’s a live stream right now on http://www.apistrategyconference.com/2014Chicago/index.php |
| 16:27 |
|
Andre-B |
in case you missed it. they have a break right now |
| 16:50 |
|
|
ph88_ joined #rest |
| 17:18 |
|
|
foist joined #rest |
| 17:27 |
|
Mxyzpltk |
thanks, watching |
| 17:44 |
|
pdurbin |
oh sure |
| 17:51 |
|
|
Left_Turn joined #rest |
| 17:54 |
|
|
systmkor joined #rest |
| 17:55 |
|
Andre-B |
I hope that they’ll put up the recordings later online |
| 19:06 |
|
|
shrink0r joined #rest |
| 19:15 |
|
|
graste joined #rest |
| 19:30 |
|
|
shrink0r joined #rest |
| 19:35 |
|
|
graste joined #rest |
| 19:46 |
|
|
systmkor1 joined #rest |
| 20:52 |
|
|
systmkor joined #rest |
| 21:12 |
|
|
shrink0r joined #rest |
| 21:14 |
|
ybit |
yet another question |
| 21:15 |
|
ybit |
when creating a resource, which should you prefer, numbers for ids or hex strings? |
| 21:15 |
|
ybit |
1 vs 5424593c4d04c1dce0fd157d |
| 21:16 |
|
ybit |
api.example.com/resource/5424593c4d04c1dce0fd157d seems a bit much, but most likely i won't have two duplicate inserts |
| 21:16 |
|
ybit |
s/two// |
| 21:42 |
|
|
tr3online joined #rest |
| 21:53 |
|
|
begriffs joined #rest |
| 21:55 |
|
|
begriffs_ joined #rest |
| 22:02 |
|
whartung |
not really germance ybit -- computers really don't care what the ID is. Long ids like that I end up cut and pasting anywhere I need to use them, so, as long as they're "double click" friendly (i.e. I double click on the id, and I get all, and only, the ID), I'm content |
| 22:04 |
|
whartung |
*germane |
| 22:06 |
|
|
begriffs joined #rest |
| 22:08 |
|
|
_ollie joined #rest |
| 22:09 |
|
|
begriffs joined #rest |
| 22:36 |
|
|
begriffs joined #rest |
| 22:46 |
|
|
begriffs joined #rest |
| 23:10 |
|
|
begriffs joined #rest |
| 23:28 |
|
|
begriffs joined #rest |
| 23:48 |
|
|
begriffs joined #rest |
