These are chat archives for jdubray/sam

7th
Dec 2016
Jean-Jacques Dubray
@jdubray
Dec 07 2016 02:26
Ah... sorry I missed that. I am getting too old.
@inrix-vincent-jo One of the problems with REST is that the URL syntax is not a full fledged query language and sometimes you have to result to a POST to make a query otherwise it would not be possible to express it in a URL path.
My two cents is that when you apply REST to the letter you quickly corner yourself in paths that make no sense at all (because a verb would do a much better job than a noun)
Jean-Jacques Dubray
@jdubray
Dec 07 2016 02:32
Lots of highly visible projects seem to be backing off from REST because of its underlying model is just too constraining and doesn't make much sense outside the trivial samples that you find in tutorials. But hey some people made a lot of money teach how to convert verbs into nouns...
As a rule of thumb, REST works well when the lifecycle of a resource maps to the one of a Web page (Create, Update, Detele, Query). When resources have many "states" (as in SAM states) it becomes increasingly difficult, and frankly plain stupid, to map actions (as in SAM actions) to nouns (ancillary resources).
Edward Mulraney
@edmulraney
Dec 07 2016 19:51
lol... you have a slightly cynical outlook on everything @jdubray
Jean-Jacques Dubray
@jdubray
Dec 07 2016 21:09
Actually, I happened to be right there and know personally the small group of people that started to promote REST heavily for highly political reasons.
My litmus test to cynicism is when people don't want to engage in a discussion about what it is they are promoting. The RESTafarians as they were called back then would simply have one answer "you don't understand". Really?
Jean-Jacques Dubray
@jdubray
Dec 07 2016 21:16
This is my post on the RESTpolitik (in reference to Nixon's Real-Politik). The neo-REST movement started on April 18 2006 with that post from Tim Bray.
Lots of really good work was thrown away, including the ancestor of GraphQL (Service Data Object, for which I was a co-author).
Jean-Jacques Dubray
@jdubray
Dec 07 2016 21:23
Tim Bray et al. made us lose about a decade with his post. We could have substituted SOAP for HTTP and XML for JSON and continue on the rich semantics that had been build over the previous 10 years (1996-2006), I was there too.
Jean-Jacques Dubray
@jdubray
Dec 07 2016 21:30
Now we have the same kind of loosey-goosey semantics being thrown up in the air, consider this, in 2016 our industry has not standardized on the definition of an action, we have been writing billions of lines of codes for decades without ever agreeing on what an action is. Here is my recent post on that question.
Just in the react sphere there are three definitions floating around (Dan's, Lee Byron's and Michel Weststrate's).
Is that really serious?
devin ivy
@devinivy
Dec 07 2016 21:35
whose job is it to make sure everyone agrees what an action is?
language is hard :P
Jean-Jacques Dubray
@jdubray
Dec 07 2016 21:40
it's a community's job
If there are differences, then they should be discussed and resolved
Imagine if there were 4 different definition of what a function is.
devin ivy
@devinivy
Dec 07 2016 21:51
i'd cry?
Jean-Jacques Dubray
@jdubray
Dec 07 2016 22:28
:+1: