On 04/02/13 13:18, Bill Burke wrote:
>
>
> On 2/3/2013 1:31 PM, Jan Algermissen wrote:
>>
>> On 03.02.2013, at 19:07, Sergey Beryozkin <sberyozkin_at_talend.com> wrote:
>>
>>> Jan - I keep suffering from not reading the messages carefully - you
>>> are referring to the acceptance within your own project, which I
>>> guess is a good indication of the usefulness of this media type - sorry;
>>
>> Hey, no problem. You are right anyhow.
>>
>>>
>>> I'm with Marek though and I saw you had agreed to his proposal; I'd
>>> also suggest to do the same for WADL media type - whatever people say
>>> we are seeing WADL being useful to users, I'm sure the same is the
>>> case in RI...
>>>
>>
>> Well, yes. I am against WADL as a design time contract, but it is a
>> great 'tool' for server-side stuff, e.g. generating test. It could
>> also be a *runtime* form language (which would be RESTful) ... so I am
>> not against the media type :-)
>>
>
> I'm 100% against the proliferation of WADL in any way, shape, or form...
> :) Users want such a thing so they can implement RPC-like protocols.
> I've seen it first hand.
I've seen that too; and I've also seen users using it for automating the
test processes, or making sure they can design the actual data
representations at the schema level and tying them to @Path (es) instead
of manually doing it all.
That said, I don't care really if we get one less constant in MediaType
class :-), we are much less religious about WADL in CXF, if some users
find it helpful - that is fine for us...
Cheers, Sergey
>
> When users ask for WADL support, I say, well, contribute and maintain it
> yourself. Or, I send them a link to Jersey if they are really persistent
> about it
>