(no subject)

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

(no subject)

jlouvel
Administrator


Jérôme

Reply | Threaded
Open this post in threaded view
|

Re: (no subject)

Xavier Méhaut
Bonjour Jérome,

Juste un petit mail pour vous signaler l'existence d'un produit que je trouve bluffant en java :
http://www.ateji.com/px/index.html
associé aux restlets, cela pourrait améliorer les performances j'imagine.

Cordialement
Xavier

2011/6/17 Jerome Louvel <[hidden email]>


Jérôme


Reply | Threaded
Open this post in threaded view
|

Re: (no subject)

Tim Peierls
Are you suggesting using a Java-like language with parallelism constructs to rewrite the Restlet internals? That doesn't sound like a good idea.

Or do you just mean that Restlet users might find such a language helpful in writing Restlet applications? If so, my response is that it might be, but that it's orthogonal to the use of Restlet.

The responsibility of a handler method is to translate from the world of Requests and Representations (and objects converted from Representations) to application-level abstractions, and then to translate the application-level result or exception back to the world of Responses and Representations (and objects to be converted to Representations). There's not much scope for parallelism there. Rather, it's the application-level computation that might benefit from parallel techniques, but that's independent of the use of Restlet.

All of which is a long and (I hope) respectful way of saying that discussions of languages with parallelism constructs aren't really on-topic for this list, buffant though they might be. :-)

--tim


2011/6/21 Xavier Méhaut <[hidden email]>
Bonjour Jérome,

Juste un petit mail pour vous signaler l'existence d'un produit que je trouve bluffant en java :
http://www.ateji.com/px/index.html
associé aux restlets, cela pourrait améliorer les performances j'imagine.

Cordialement
Xavier

2011/6/17 Jerome Louvel <[hidden email]>


Jérôme



Reply | Threaded
Open this post in threaded view
|

Re: (no subject)

Xavier Méhaut
Thanks Tim for your answer...
The REST paradigm is of course orthogonal with parallelism from the user point of view. I've actually thought of using the Ateji features to maybe have a better handle of the misc processing by share out them at best to cores...
But I agree it is a little bit out of scope :-(

regards

2011/6/21 Tim Peierls <[hidden email]>
Are you suggesting using a Java-like language with parallelism constructs to rewrite the Restlet internals? That doesn't sound like a good idea.

Or do you just mean that Restlet users might find such a language helpful in writing Restlet applications? If so, my response is that it might be, but that it's orthogonal to the use of Restlet.

The responsibility of a handler method is to translate from the world of Requests and Representations (and objects converted from Representations) to application-level abstractions, and then to translate the application-level result or exception back to the world of Responses and Representations (and objects to be converted to Representations). There's not much scope for parallelism there. Rather, it's the application-level computation that might benefit from parallel techniques, but that's independent of the use of Restlet.

All of which is a long and (I hope) respectful way of saying that discussions of languages with parallelism constructs aren't really on-topic for this list, buffant though they might be. :-)

--tim


2011/6/21 Xavier Méhaut <[hidden email]>
Bonjour Jérome,

Juste un petit mail pour vous signaler l'existence d'un produit que je trouve bluffant en java :
http://www.ateji.com/px/index.html
associé aux restlets, cela pourrait améliorer les performances j'imagine.

Cordialement
Xavier

2011/6/17 Jerome Louvel <[hidden email]>


Jérôme