Fwd: Applications Messaging Standard

[John Taylor]

Whoops.  Forgot to hit reply-all.

Begin forwarded message:

> From: John Taylor <jontayler at gmail.com>
> Date: 8 February 2007 12:42:47 GMT
> To: "Tony Linde" <Tony.Linde at leicester.ac.uk>
> Subject: Re: Applications Messaging Standard
>
>
> On 8 Feb 2007, at 11:48, Tony Linde wrote:
>
>> Thanks all. I agree that, as it stands, plastic is not supposed to  
>> be any
>> sort of message queuing system: if it is just passing messages  
>> around apps
>> on a desktop, the user can plainly see if anything has happened  
>> and click
>> the button again if the message didn't arrive.
>
> My experience has been that if the message doesn't arrive then it  
> never will - something has screwed up and you need to try that time- 
> honoured computer fix of shutting the app down and starting it up  
> again (or unregistering and reregistering).
>
>>
>> It may be that Al's hack to route messages between different  
>> machines could
>> stray into Doug's distributed apps scenario and agree with Al that  
>> it ought
>> not to be part of any initial standard.
>
> That's right.  I don't see any reason for this to hold back  
> standardizing the protocol for messaging within the desktop.  We  
> need interested parties to work with Al to experiment with  
> connecting desktops together, write down what works and call that a  
> different standard.
>
>>
>>> PLASTIC did right was leaving the messages alone. My application can
>>
>> I think you do need some standard messages but agree that the  
>> messages need
>> to be extendable, much as we did with the registry types.
>
> That's pretty much what we've got.  Our process so far has been to  
> propose messages, incubate them for a while, then once they're  
> widely adopted call them a standard.  I'm afraid our documentation  
> in this area is poor though.
>
>
>>
>> T.
>>
>>> -----Original Message-----
>>> From: Alasdair Allan [mailto:aa at astro.ex.ac.uk]
>>> Sent: 08 February 2007 10:20
>>> To: apps at ivoa.net
>>> Cc: Tony Linde; Noel Winstanley
>>> Subject: Re: Applications Messaging Standard
>>>
>>>
>>> Noel Winstanley wrote:
>>>> Plastic has a lot less support for security, transactions, and
>>>> other things that typically concern enterprise software. If a
>>>> plastic message is dropped, or even spoofed or intercepted, it's
>>>> assumed that this isn't the end of the world.
>>>
>>> I think this is a good assumption, we shouldn't over-engineer our
>>> solution. Things like quality of service and guaranteed delivery add
>>> a whole layer of complexity we don't really need for the task at  
>>> hand.
>>>
>>> I also don't think we need support for (much) security, and I don't
>>> really want to go near logging or transactions, again I don't think
>>> its needed.
>>>
>>>> Similar to a messaging system, a plastic hub doesn't really care
>>>> about the content of the messages - the definition and format of
>>>> mesages are left to the message producer and consumer to agree,
>>>> whilst the hub just takes care of routing them.
>>>
>>> However this is vital I think, one thing I really (really) think
>>> PLASTIC did right was leaving the messages alone. My application can
>>> get a showObjects message (for instance) and do whatever it likes
>>> with it. It isn't constrained by the standard to do the expected
>>> thing. That leads to innovative uses for the messages, you can't
>>> think of everything when you build a standard...
>>>
>>>> Plastic is good for quickly exchanging control info between
>>> desktop
>>>> apps. Message Brokers are good for, eg,  processing banking
>>>> transactions.
>>>
>>> Yup!
>>>
>>> Al.
>>>
>>>
>>>
>>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.ivoa.net/pipermail/apps/attachments/20070208/5802b69f/attachment.html>