| Time |
S |
Nick |
Message |
| 00:27 |
|
|
vanHoesel joined #rest |
| 01:10 |
|
|
shrink0r joined #rest |
| 02:34 |
|
|
fumanchu_ joined #rest |
| 02:34 |
|
|
shrink0r joined #rest |
| 03:28 |
|
|
lemur joined #rest |
| 04:04 |
|
|
shrink0r_ joined #rest |
| 04:56 |
|
|
mgomezch joined #rest |
| 06:35 |
|
|
vanHoesel joined #rest |
| 07:25 |
|
|
blahdeblah joined #rest |
| 07:36 |
|
|
vanHoesel joined #rest |
| 07:56 |
|
|
adaro joined #rest |
| 07:56 |
|
|
azr joined #rest |
| 08:50 |
|
|
azr joined #rest |
| 09:07 |
|
|
fuzzyhorns joined #rest |
| 09:39 |
|
|
_ollie joined #rest |
| 10:25 |
|
|
vanHoesel joined #rest |
| 10:45 |
|
|
shrink0r joined #rest |
| 10:45 |
|
|
vanHoesel joined #rest |
| 10:54 |
|
|
shrink0r joined #rest |
| 10:55 |
|
|
shrink0r joined #rest |
| 10:57 |
|
|
SupaHam joined #rest |
| 11:04 |
|
|
dEPy joined #rest |
| 11:10 |
|
|
jcromartie joined #rest |
| 11:13 |
|
|
Left_Turn joined #rest |
| 11:26 |
|
|
vanHoesel joined #rest |
| 11:48 |
|
|
azr_ joined #rest |
| 11:59 |
|
|
vanHoesel joined #rest |
| 12:41 |
|
pdurbin |
I think I'm convincing myself we should put a version number for our API right in the URL like Twitter and Trello do: https://github.com/IQSS/dataverse/issues/411 |
| 12:46 |
|
trygvis |
boo! :) |
| 12:47 |
|
pdurbin |
trygvis: you like the way GitHub does it better? "By default, all requests receive the v3 version of the API. We encourage you to explicitly request this version via the Accept header." |
| 12:47 |
|
|
Left_Turn joined #rest |
| 12:47 |
|
pdurbin |
"Important: The default version of the API may change in the future. If you're building an application and care about the stability of the API, be sure to request a specific version in the Accept header as shown in the examples below." |
| 12:50 |
|
trygvis |
no, that is even worse |
| 12:51 |
|
trygvis |
the first thing to do is to specify that the resources can add stuff as they see fit |
| 12:52 |
|
trygvis |
that will often fix most problems, in the most extreme case a v2 consumer can just look at fields starting with "2", or are inside the "2" object |
| 12:53 |
|
trygvis |
they way I see it is that if the data changes so fundamentally that it requires a totally new kind, you probably need a new link to it anyway |
| 12:54 |
|
trygvis |
IMO it is more in line with Postel said back in the days |
| 12:55 |
|
trygvis |
http://tools.ietf.org/html/rfc1122#section-1.2.2 |
| 12:56 |
|
pdurbin |
the ol' Robustness Principle |
| 12:56 |
|
trygvis |
yep |
| 12:56 |
|
trygvis |
I see the appeal with versioned URLs when you don't get clients to follow links |
| 12:57 |
|
|
huckleberry78 joined #rest |
| 12:59 |
|
|
csgeek joined #rest |
| 13:06 |
|
daxim |
where can I find further information about the ISINDEX format (accessible via the query_keyword method in the Perl URI module)? rfc 3986 just says "pchar" for any kind of query parameter, which is not helpful |
| 13:06 |
|
pdurbin |
I see the appeal of versioned URLs too. |
| 13:08 |
|
|
azr joined #rest |
| 13:13 |
|
trygvis |
daxim: I can't find anything about isindex in rfc3986 |
| 13:13 |
|
daxim |
that's what I said |
| 13:15 |
|
trygvis |
uhm, so you're asking about something in a perl module? |
| 13:15 |
|
trygvis |
try #perl |
| 13:15 |
|
|
jsys joined #rest |
| 13:15 |
|
jsys |
_ollie: hi |
| 13:16 |
|
|
jsys joined #rest |
| 13:16 |
|
daxim |
I'm not asking about a perl module |
| 13:17 |
|
trygvis |
then I have no idea what you are asking about :) |
| 13:17 |
|
daxim |
I'm asking about the ISINDEX format |
| 13:17 |
|
daxim |
is it different than www-urlencoded |
| 13:17 |
|
trygvis |
where did you hear about this format? |
| 13:18 |
|
daxim |
I heard about it in the documentation for the Perl URI module |
| 13:18 |
|
trygvis |
can you give me an url so I can read it too? |
| 13:19 |
|
daxim |
https://metacpan.org/pod/URI#uri-query_keywords |
| 13:19 |
|
trygvis |
I can't find any mention of "isindex" on that page |
| 13:19 |
|
daxim |
method query_form processes x-www-form-urlencoded, query_keywords processes isindex |
| 13:20 |
|
jsys |
so I've been reading the HTTP specs like novels. |
| 13:20 |
|
jsys |
Starting from HTTP 0.9 in 1991 |
| 13:20 |
|
jsys |
And I'm not completely disillusioned this thing is a cohesive piece of work with some grand design in it. |
| 13:21 |
|
jsys |
It's patchwork at best, and that's not bad because maybe that's the only way to get work done. |
| 13:21 |
|
jsys |
But it kinda shrinks the REST concept as "the web is REST" more to "Fielding tried to see HTTP as REST, but it's not" |
| 13:21 |
|
jsys |
So it's a design that was meant to last decades as his paper said, but it was stillborn |
| 13:21 |
|
pdurbin |
jsys: um. you're kind of interrupting daxim and trygvis. can you please hold for a few minutes? |
| 13:22 |
|
jsys |
I suppose cookies were a bad sign he himself noted |
| 13:22 |
|
trygvis |
daxim: I still can't find any reference to isindex so I don't understand what you're asking about. I can't find anything about the "isindex" format |
| 13:22 |
|
trygvis |
the query_form stuff seems to be some conventions that perl's URI parser has |
| 13:22 |
|
trygvis |
sorry, query_keywords |
| 13:25 |
|
daxim |
that would run counter to my expectations, the authors don't just to make up conventions, all the rest of the web infrastructure modules follow rfcs/standards |
| 13:26 |
|
trygvis |
well, find me a reference to where where 'ISINDEX' is written and I'll try to find out more |
| 13:27 |
|
daxim |
we can still find plenty of documents talking about the hypermedia control and how it's used, but not the exact format |
| 13:33 |
|
jsys |
daxim, trygvis: if following a standard mattered, it'd have actual tangible impact, like "hundreds of clients who understand isindex won't understand my service if I use it wrong". Is there such an effect? No. So it's an inactive and/or dead standard. Not worth performing exegesis on. |
| 13:37 |
|
jsys |
As a general rule of you need to perform exegesis on an idea or technique or a protocol, you're wasting your time |
| 14:18 |
|
|
Mxyzpltk joined #rest |
| 14:25 |
|
|
gamache joined #rest |
| 14:58 |
|
|
Andre-B joined #rest |
| 15:19 |
|
|
fumanchu joined #rest |
| 15:54 |
|
|
saml joined #rest |
| 16:03 |
|
|
lemur joined #rest |
| 16:27 |
|
whartung |
Pretty sure if you have to use the word 'exegesis' in casual conversation, you're wasting your time. |
| 16:30 |
|
|
vanHoesel joined #rest |
| 16:31 |
|
whartung |
pdurbin: that "containerless" article is interesting, but, I dunno…hard to see anything "container free" today. |
| 16:33 |
|
|
_ollie joined #rest |
| 16:43 |
|
pdurbin |
everything is a container he seems to say. your laptop is a container :) |
| 16:45 |
|
whartung |
yea I didn't read it in detail, but I cite things like DropWizard. |
| 16:46 |
|
whartung |
is DW "lightweight" compared to, say, Glassfish? Sure. |
| 16:46 |
|
whartung |
I've been struggling with this problem, just to "lighten up" my applications. |
| 16:47 |
|
pdurbin |
does this mirror make my war look fat? |
| 16:47 |
|
whartung |
mirror.jar |
| 16:49 |
|
whartung |
size of jars isn't how I measure weight, it's more cognitive weight of the "oh no, what now" factor |
| 16:50 |
|
pdurbin |
yeah. you need moar microservices, apparently |
| 16:51 |
|
whartung |
yea, somehow publishing 100 seperate "applicaitons", all in their own process, is "better". Especially if I have each one wrapped in a container wrapped in a VM... |
| 16:58 |
|
fumanchu |
until performance |
| 17:09 |
|
whartung |
yea I thought we figured out remote calls weren't the bee knees 10 years ago |
| 17:14 |
|
|
shrink0r joined #rest |
| 17:29 |
|
saml |
do you like json? |
| 17:29 |
|
saml |
http://stackoverflow.com/questions/21120999/representing-null-in-json is there good practice for this? |
| 17:29 |
|
saml |
looks like for arrays at least, people prefer {"x":[]} instead of {} |
| 17:30 |
|
saml |
for many other values, they just omit the property all together from json |
| 17:31 |
|
saml |
let's say in xml... <data><title>foo</title><images></images></data> vs. <data><title>foo</title></data> |
| 17:47 |
|
trygvis |
saml: in my DTOs I always initialize arrays to empty arrays so I know that they're always non-null, and let the deserializer handle whatever comes in |
| 17:47 |
|
trygvis |
when integration systems it's often useful to say that missing, null and empty array all mean the same |
| 17:55 |
|
|
tr3online joined #rest |
| 17:58 |
|
|
Mxyzpltk joined #rest |
| 18:06 |
|
|
_ollie joined #rest |
| 18:27 |
|
|
azr joined #rest |
| 18:33 |
|
|
Andre-B joined #rest |
| 19:03 |
|
|
saml joined #rest |
| 19:21 |
|
saml |
trygvis, thanks |
| 19:29 |
|
|
jsys joined #rest |
| 19:29 |
|
|
graste joined #rest |
| 19:30 |
|
jsys |
So, is it RESTful to deprecate a resource. |
| 19:31 |
|
jsys |
X-Deprecated: "This resource is goin doooooooooooooown" |
| 19:39 |
|
saml |
sure |
| 19:39 |
|
|
aGHz joined #rest |
| 19:40 |
|
|
shrink0r joined #rest |
| 19:46 |
|
|
vanHoesel joined #rest |
| 19:50 |
|
|
mezod joined #rest |
| 19:55 |
|
|
_ollie joined #rest |
| 20:04 |
|
|
_ollie joined #rest |
| 20:20 |
|
|
shrink0r_ joined #rest |
| 21:06 |
|
|
_ollie joined #rest |
| 21:56 |
|
|
vanHoesel joined #rest |
| 21:59 |
|
|
_ollie joined #rest |
| 23:19 |
|
|
vanHoesel joined #rest |
