On 3/6/2014 10:49 AM, Santiago Pericas-Geertsen wrote:
>
> On Mar 6, 2014, at 10:21 AM, Bill Burke <bburke_at_redhat.com
> <mailto:bburke_at_redhat.com>> wrote:
>
>>
>>
>> On 3/6/2014 9:18 AM, Santiago Pericas-Geertsen wrote:
>>>
>>> On Mar 6, 2014, at 8:59 AM, Bill Burke <bburke_at_redhat.com
>>> <mailto:bburke_at_redhat.com>> wrote:
>>>
>>>>
>>>>
>>>> On 3/6/2014 8:46 AM, Santiago Pericas-Geertsen wrote:
>>>>>
>>>>> On Mar 5, 2014, at 3:07 PM, Bill Burke <bburke_at_redhat.com
>>>>> <mailto:bburke_at_redhat.com>> wrote:
>>>>>
>>>>>> Honestly, I don't see why we should have support for this in
>>>>>> JAX-RS. Let the servlet spec handle it.
>>>>>
>>>>> Would you mind elaborating on this? I don't know of many
>>>>> developers that would prefer to use servlets over JAX-RS for
>>>>> anything, including new features like SSE.
>>>>>
>>>>
>>>> So, JAX-RS is no longer a REST framework, but now a replacement for
>>>> servlets?
>>>
>>> No, but REST frameworks are becoming increasingly popular with
>>> thin-server architectures and browser clients. As a developer, it
>>> would feel awkward to move from JAX-RS to Servlets just to implement
>>> SSE for those clients.
>>>
>>
>> What would be awkward is having support for a non-HTTP session-based
>> protocol inside of a REST framework.
>
> I would agree. Except SSE _is designed_ to work on HTTP and it is how
> HTML5 clients use it. I think you're mixing up SSE with WebSockets.
> Please take a look at the specification:
>
> http://www.w3.org/TR/eventsource/
>
I did look at the spec. It is session based. Its standardized Comet.
--
Bill Burke
JBoss, a division of Red Hat
http://bill.burkecentral.com