On 16 Feb 2007, at 11:51, Tony Linde wrote:
> Yes, I thought we were making different assumptions here.
>
>> about only how to communicate with an existing (running) hub,
>
> I'd tend to agree with Mark: the messaging standard is about the
> above.
So do I - I was just careless with my choice of words.
> The
> standard should include what a hub does to make known that it is
> running and
> how plastic apps find out how to communicate with the hub. And, for
> now, the
> only solution we all agree might work is that the hub writes a file
> into
> some common area (~/.ivoamsg) and that file contains the mode by which
> communications with the hub take place.
>
> BUT we're still stuck with the problem of embedded hubs and what
> happens if
> they go away. Personally, I think the simplest solution is to
> mandate that
> hubs can only be stand-alone.
>
What do we mean by standalone? That the process that acts as a hub cannot have any other functionality? Is it reasonable to mandate this? (I don't think we could enforce it anyway.)
> THEN we get onto the issue (if we want to address it - could leave
> it to the
> user to sort out) of how a running plastic app can find and invoke
> the hub.
I think we can leave this issue for now as something that is user (and application) specific. Some users might want to set up their machine to start a hub when they log in. Some application authors might want to their application to auto start a hub if it's not running (how it does this need not be defined by the spec). I say we leave this open for now and see how it works in practice.
John
>
> T.
>
>> -----Original Message-----
>> From: owner-apps-at-eso.org [mailto:owner-apps-at-eso.org] On
>> Behalf Of Mark Taylor
>> Sent: 16 February 2007 11:24
>> To: apps-at-ivoa.net
>> Subject: Re: Applications Messaging Standard
>>
>> On Fri, 16 Feb 2007, John Taylor wrote:
>>
>>> On 16 Feb 2007, at 09:25, Tony Linde wrote:
>>>> Can I ask if we all are talking about the same thing here?
>> What are we
>>>> trying to discover at runtime with this file-based mechanism?
>>>>
>>>> Where the hub is and how to invoke it?
>>>> How to communicate with the hub?
>>>> What other plastic apps are running and what they can do?
>>>
>>> Hi Tony,
>>> I believe we are just trying to discover where the hub is
>> and how to invoke
>>> it.
>>
>> Not sure if I'm being pedantic here, but I'd say we're talking here
>> about only how to communicate with an existing (running) hub,
>> not how to invoke it - invoke to me means start up a process which
>> runs a hub, and that's not what the ~/.ivoamsg file would do.
>> Invoking a hub is, to me, an implementation thing.
>>
>>> For instance, if a hub supported xmlrpc as a wire
>> protocol then it would
>>> need to get the URL of its xmlrpc server to the clients,
>> perhaps by writing
>>> this URL into a file in a well-known location. Thereafter,
>> all the discovery
>>> of other applications and their capabilities takes place by
>> talking to the
>>> hub on this URL.
>>
>> .. but I agree with this explanation.
>>
>> Mark
>>
>> --
>> Mark Taylor Astronomical Programmer Physics, Bristol
>> University, UK
>> m.b.taylor-at-bris.ac.uk +44-117-928-8776
>> http://www.star.bris.ac.uk/~mbt/
>>
>
> http://www.Taglocity.com Tags: IVOA, apps
>
Received on 2007-02-16Z13:16:32