new version of REST in the VO
schaaff at newb6.u-strasbg.fr
schaaff at newb6.u-strasbg.fr
Tue Oct 28 06:52:08 PDT 2008
Quoting Douglas Tody <dtody at nrao.edu>:
> On Tue, 28 Oct 2008, Paul Harrison wrote:
>
>> On 2008-10 -28, at 04:44, Douglas Tody wrote:
>>
>>> I think it is fine to define some recommended practices for interfaces
>>> based upon the pure-REST abstraction, but this should not in any way
>>> invalidate the DAL2 interface pattern(*), which is already standardized
>>> and widely used within the VO (and also based upon a HTTP transport).
>>
>> point of order - the DAL v2 interface pattern is *not* already
>> standardised, we are in the process of trying to create a V2
>> protocol family that is more capable/scalable than the v1 interface
>> pattern - this is why GWS WG is looking at using more advanced
>> techniques than were employed in the v1 protocols.
>
> It is true that the recent DAL2 architecture document per se is not a
> standard, however essentially all the content therein was discussed
> within the DAL WG for several years, leading up to promoting SSA
> (the first of the second generation DAL interfaces, all of which
> were intended to be consistent in terms of interface) to a standard.
> SSA, and the older DAL interfaces which follow a similar pattern,
> are VO standards in wide use within the community.
>
> Note that there is no reason that interfaces such as DAL2 (which
> is based upon OpenGIS standards) and facilities such as UWS (which
> exemplifies the pure-REST approach) cannot coexist. This is already
> being demonstrated in TAP, and the same pattern can carry over
> to future interfaces such as SIAV2. This, plus eventual VOSpace
> integration will provide the scalability we all want.
the aim of the REST work is not to obsolete existing (or future) standards
as sou said the use of (future) standards like UWS may provide de facto
RESTful services
in any case it should be a guideline without any coercive meaning
i think it would be ueseful to get inputs and comments from other WGs,
i will compile all these inputs for the next interop
André
>
> - Doug
>
More information about the grid
mailing list