Another reason to like WAMP

#1

People always end up coding something similar, but not standard, not compatible with several stacks/languages, and will less features :

GO : https://t.co/aSFkhCYVCG
JS : https://github.com/capaj/socket.io-rpc
Python : http://pubsub.sourceforge.net/installation.html#system-requirements
Ruby : http://www.rabbitmq.com/tutorials/tutorial-six-ruby.html

Of course, none of that implement both RPC and Pub/Sub, none of them are compaible with each others, and none of them are standard.

0 Likes

#2

Hello Michel,

found time to read up on this now ..

People always end up coding something similar, but not standard, not
compatible with several stacks/languages, and will less features :

GO : https://t.co/aSFkhCYVCG

The Docker thing

https://github.com/docker/libchan

and

http://de.slideshare.net/shykes/docker-the-road-ahead

is interesting. The goal of "scaling an application by breaking it down into loosely coupled concurrent services" is similar to WAMP.

The ability to use different transports .. btw: WAMP does "in-memory" transport also - this kind of transport is used when you run app components embedded in a router ("side-by-side"). Pluggable serializations.

However, the RPC example

https://github.com/docker/libchan#example-usage

looks quite tedious, and I can't find an example of PubSub.

I'd also say that WAMP has an edge design wise: it's the result of quite some iterations, with practical experience in between, all fused into a coherent design.

JS : https://github.com/capaj/socket.io-rpc
Python : http://pubsub.sourceforge.net/installation.html#system-requirements
Ruby : http://www.rabbitmq.com/tutorials/tutorial-six-ruby.html

Of course, none of that implement both RPC _and_ Pub/Sub, none of them
are compaible with each others, and none of them are standard.

Yes. E.g. having a public, open specification with multiple interoperable implementations is a different level than a single ad-hoc implementation.

The new WAMP v2 spec (basic profile) is - I think - not bad: quite precise and exhaustive. For an RFC, it wouldn't be sufficient yet - but we could do an RFC for WAMP (if time permits).

Another thing that bothers me: Autobahn|Testsuite had testing facilities for WAMP v1 - I started with WAMP v2 test cases, but got distracted by other stuff again.

If we had an automated test suite for WAMP v2 implementations, that would be another big advantage. There are only very few such things on the street, and it is incredible useful for maintaining the integrity of a protocol out in the wild (for implementors).

/Tobias

ยทยทยท

Am 11.06.2014 04:01, schrieb Michel Desmoulin:

--
You received this message because you are subscribed to the Google
Groups "Autobahn" group.
To unsubscribe from this group and stop receiving emails from it, send
an email to autobahnws+...@googlegroups.com
<mailto:autobahnws+...@googlegroups.com>.
For more options, visit https://groups.google.com/d/optout.

0 Likes