Identifying version of UWS

[Paul Harrison]

> On 2015-09 -02, at 19:18, Patrick Dowler <patrick.dowler at nrc-cnrc.gc.ca> wrote:
> 
> On 02/09/15 01:20 AM, Paul Harrison wrote:
>> I much prefer ivo://ivoa.net/std/UWS#rest-1.1 as the standardID.
> 
> The name of the feature there probably doesn't look too important, but still I would argue that "rest" is essentially meaningless unless you anticipate the standardID is to differentiate bindings (REST vs SOAP (shudder) vs something else). If you envision other capabilities (e.g. standardising a sychronous job execution pattern in UWS) then I would name the current one in a way that both would make sense when you name the new one. On the other hand, if you are sure that UWS will only ever be one capability then don't name it and just use
> 
> ivo://ivoa.net/std/UWS#1.1
> 
> In DataLink we called it #links-1.0 with no particular reason to think there would be other capabilities later, but still naming that one for what it is and allowing for other capabilities to be defined later.

The reason I like “rest” in the standardID is that the standard does talk of future possible alternative bindings - admittedly that was because SOAP was originally the other binding in mind, and I agree that is highly unlikely now. However, there are other bindings that I would do now if I had the time - e.g. JSON over web-sockets that would actually be useful for the latest browser based development environments.

Cheers,
	Paul.