AMF in a RESTful enviroment

[rpb]

Hi Guys!

I am scourging the net for article on how one would implement AMF
gateways in a RESTful enviroment. Has anyone come up with or seen a
generic way one could accomplish this regardless of the framework?


Thanks

[Arnar Birgisson]

Hi Akira,

What do you mean by a RESTful environment? The word has kind of lost
its meaning over the past two years.

Arnar

On Dec 2, 2007 7:45 PM, akira <nhytro-python@...> wrote:

> Hi Guys!
>
> I am scourging the net for article on how one would implement AMF
> gateways in a RESTful enviroment. Has anyone come up with or seen a
> generic way one could accomplish this regardless of the framework?
>
>
> Thanks
> _______________________________________________
> PyAMF dev mailing list - dev@...
> http://lists.pyamf.org/mailman/listinfo/dev
>

[rpb]

This was helpful

http://www.xfront.com/REST-Web-Services.html


akira wrote:

> Hi Guys!
>
> I am scourging the net for article on how one would implement AMF
> gateways in a RESTful enviroment. Has anyone come up with or seen a
> generic way one could accomplish this regardless of the framework?
>
>
> Thanks
> _______________________________________________
> PyAMF dev mailing list - dev@...
> http://lists.pyamf.org/mailman/listinfo/dev
>

[Arnar Birgisson]

On Dec 2, 2007 8:17 PM, akira <nhytro-python@...> wrote:
> This was helpful
>
> http://www.xfront.com/REST-Web-Services.html

Ok, I think you are on the wrong track here. I don't see why AMF
remoting should have much to do with REST in its original sense (the
work of Roy Fielding).

There is some huge misunderstanding growing for a while that REST
applies to all web services. The "web services" that Fielding refers
to are not RPC gateways like AMF Remoting, but rather web services
perhaps meant to be used both by humans and computers, exposing and
altering information presented in some semantically sound manner.

AMF remoting (like many RPC solutions) does not at all go by the
principles of REST, and neither should it imo. For RPC, the HTTP
protocol is only being used as a transport medium (and one of many
possible, AMF remoting can also use RTMP and other RPC methods can
even use email). It has very little to do with semantic strucure of
states or data on the web - since it's all about function invocation.

All AMF remoting requests are done with POST, simply because the POST
body is used for the AMF data. There is only one URL per gateway, all
other information is contained in the AMF payload. This is how many
RPC solutions work and is done mostly to abstract away that you are
working over HTTP (because that really doesn't matter).

The question if AMF remoting (or any RPC-over-HTTP) can be implemented
in a RESTful manner is thus a void one imho, there is little to be
gained by doing so (I have tried).

hth,
Arnar

[rpb]

Thanks Arnar! Hmm I felt so stoked up after reading some REST tutorials.
I thought it would be beneficial to do everything in a RESTful manner.
Seems like I was mistaken.
oh well...



Arnar Birgisson wrote:

> On Dec 2, 2007 8:17 PM, akira <nhytro-python@...> wrote:
>  
>> This was helpful
>>
>> http://www.xfront.com/REST-Web-Services.html
>>    
>
> Ok, I think you are on the wrong track here. I don't see why AMF
> remoting should have much to do with REST in its original sense (the
> work of Roy Fielding).
>
> There is some huge misunderstanding growing for a while that REST
> applies to all web services. The "web services" that Fielding refers
> to are not RPC gateways like AMF Remoting, but rather web services
> perhaps meant to be used both by humans and computers, exposing and
> altering information presented in some semantically sound manner.
>
> AMF remoting (like many RPC solutions) does not at all go by the
> principles of REST, and neither should it imo. For RPC, the HTTP
> protocol is only being used as a transport medium (and one of many
> possible, AMF remoting can also use RTMP and other RPC methods can
> even use email). It has very little to do with semantic strucure of
> states or data on the web - since it's all about function invocation.
>
> All AMF remoting requests are done with POST, simply because the POST
> body is used for the AMF data. There is only one URL per gateway, all
> other information is contained in the AMF payload. This is how many
> RPC solutions work and is done mostly to abstract away that you are
> working over HTTP (because that really doesn't matter).
>
> The question if AMF remoting (or any RPC-over-HTTP) can be implemented
> in a RESTful manner is thus a void one imho, there is little to be
> gained by doing so (I have tried).
>
> hth,
> Arnar
> _______________________________________________
> PyAMF dev mailing list - dev@...
> http://lists.pyamf.org/mailman/listinfo/dev
>
>  

[Arnar Birgisson]

Well, that was only my opinion, there are heaps of wiser folks than me :)

Arnar

On Dec 4, 2007 8:09 PM, akira <nhytro-python@...> wrote:

> Thanks Arnar! Hmm I felt so stoked up after reading some REST tutorials.
> I thought it would be beneficial to do everything in a RESTful manner.
> Seems like I was mistaken.
> oh well...
>
>
>
>
> Arnar Birgisson wrote:
> > On Dec 2, 2007 8:17 PM, akira <nhytro-python@...> wrote:
> >
> >> This was helpful
> >>
> >> http://www.xfront.com/REST-Web-Services.html
> >>
> >
> > Ok, I think you are on the wrong track here. I don't see why AMF
> > remoting should have much to do with REST in its original sense (the
> > work of Roy Fielding).
> >
> > There is some huge misunderstanding growing for a while that REST
> > applies to all web services. The "web services" that Fielding refers
> > to are not RPC gateways like AMF Remoting, but rather web services
> > perhaps meant to be used both by humans and computers, exposing and
> > altering information presented in some semantically sound manner.
> >
> > AMF remoting (like many RPC solutions) does not at all go by the
> > principles of REST, and neither should it imo. For RPC, the HTTP
> > protocol is only being used as a transport medium (and one of many
> > possible, AMF remoting can also use RTMP and other RPC methods can
> > even use email). It has very little to do with semantic strucure of
> > states or data on the web - since it's all about function invocation.
> >
> > All AMF remoting requests are done with POST, simply because the POST
> > body is used for the AMF data. There is only one URL per gateway, all
> > other information is contained in the AMF payload. This is how many
> > RPC solutions work and is done mostly to abstract away that you are
> > working over HTTP (because that really doesn't matter).
> >
> > The question if AMF remoting (or any RPC-over-HTTP) can be implemented
> > in a RESTful manner is thus a void one imho, there is little to be
> > gained by doing so (I have tried).
> >
> > hth,
> > Arnar
>
> > _______________________________________________
> > PyAMF dev mailing list - dev@...
> > http://lists.pyamf.org/mailman/listinfo/dev
> >
> >
>
> _______________________________________________
> PyAMF dev mailing list - dev@...
> http://lists.pyamf.org/mailman/listinfo/dev
>