activemq-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hiram Chirino" <hi...@hiramchirino.com>
Subject Re: NMS
Date Mon, 11 Jun 2007 21:04:27 GMT
I agree qpid should have a API that exposes all the features of the
AMQP protocol since AMQP is the focus of the project.

But the point of NMS and CMS is not to have an API that you can access
ALL the features of the messaging provider.  The idea is to have an
abstract API that can semantically be easily implemented by most
messaging providers.  And why would you want to limit yourself to an
API that has a reduced set of features?  So that application can have
a choice when it chooses the vendor for the implementation of the API.

We have seen that technologies like ODBC, JDBC, and JMS made it easier
for folks to stay vendor neutral.  NMS and CMS is just trying bring
the same kind of abstraction for messaging on the .NET and C++ worlds.

I'm going to bet that your average development shop is willing to
sacrifice a little bit of performance of feature access in order to
gain the ability to choose between messaging vendors.  APIs are what
usually lock in an application to a specific vendor.  If the
application uses an abstract API, vendor lock in can potentially be
avoided.

So, while QPID should have a high performance low level API, I think
also supporting a more abstract messaging model like NMS and CMS would
be a big plus for your end users.


On 6/11/07, Colin Crist <colincrist@hermesjms.com> wrote:
>
> As an observer here and user of messaging stuff, all I can say is please
> don't force other languages to have to live with a JMS API varient. I agree
> its very limited and is a success as it's a common demoninator that was
> standardised - the standardisation is really it best asset, not the API per
> se.
>
> I've always thought an AMQP API that maps accurately to the protocol and is
> mostly familiar across languages (in the core abstractions that map to the
> protocol, constants and so on) but uses those languages constructs
> effectively is the way to go.
>
> AMQP != JMS so why go there? ActiveMQ is the Apache JMS implementation is it
> not after all?Build a compliant JMS on the side in Java. Only Java.
> Prefereably layered on top of a more appropriate AMQP API. Then you can
> forget about lawyers. Sighs of relief all round surely?
>
> Hope you don't mind my $0.02 from the the leftfield.
>
> Colin.
>
>
> > -----Original Message-----
> > From: John O'Hara [mailto:john.r.ohara@gmail.com]
> > Sent: 11 June 2007 17:13
> > To: qpid-dev@incubator.apache.org
> > Subject: Re: NMS
> >
> > With JMS/NMS it's more likely to do with Copyright.
> > The question is "Is NMS derived from JMS", not about other
> > aspects of IP, but we need a lawyers comment.
> >
> > If the API looked substantially different from JMS there
> > probably wouldn't be an issue.
> >
> > Since JMS is not the greatest messaging API in the world,
> > imho, why do other languages have to suffer from it?
> > Not everyone using these other languages has seen JMS.
> >
> > Given AMQP (above rant aside) was engineered to be cross
> > platform/language - why not start with a fairly thin AMQP-centric API?
> > Then on top of it build JMS, WCF, etc...
> >
> > JMS by no means the only shape for a messaging API.
> >
> > Just my 2c
> > John
> >
> >
> >
> >
> >
> >
> > On 11/06/07, Hiram Chirino <hiram@hiramchirino.com> wrote:
> > >
> > > I share the same sentiment.  The 'tainting' assumptions
> > that are being
> > > made around the JMS spec need to get resolved ASAP for the sake of
> > > both the projects.
> > >
> > > On 6/11/07, Rupert Smith < rupertlssmith@googlemail.com> wrote:
> > > > You can not seriously tell me that because you've read
> > the JMS spec
> > > > you
> > > may
> > > > not write a spec for a messaging protocol, without breaking the
> > > > rules of
> > > the
> > > > JMS spec. This discussion is bordering on the ridiculous.
> > > >
> > > > On 11/06/07, James Strachan <james.strachan@gmail.com> wrote:
> > > > >
> > > > > On 6/11/07, John O'Hara < john.r.ohara@gmail.com> wrote:
> > > > > > "..please consider that the same folks who developed
> > NMS helped
> > > develop
> > > > > the
> > > > > > AMQP specification so the same alleged 'taint' that is being
> > > > > > claimed
> > >
> > > > > against
> > > > > > NMS would also apply to AMQP."
> > > > > >
> > > > > > To clarify.
> > > > > > James is not the source of any of the content of the AMQP
> > > specification
> > > > > thru
> > > > > > v0-9, to the best of my knowledge.
> > > > >
> > > > > So I've just been through my AMQP spec emails; The
> > first version I
> > > > > was sent was 0.3a4 in August 2004 then I contributed to the
> > > > > specification up to around April 2005 when it had
> > reached around 0.81a.
> > > > >
> > > > > So what happened to the specification I worked on? Was
> > a new group
> > > > > formed excluding all the folks who were on the mailing
> > list I was
> > > > > on (including yourself John you were on the mailing
> > list with me)
> > > > > and were all versions of AMQP destroyed and a new cleanroom
> > > > > version created after 0.81a, ignoring all that previous work?
> > > > >
> > > > > --
> > > > > James
> > > > > -------
> > > > > http://macstrac.blogspot.com/
> > > > >
> > > >
> > >
> > >
> > > --
> > > Regards,
> > > Hiram
> > >
> > > Blog: http://hiramchirino.com
> > >
> >
>
>
>


-- 
Regards,
Hiram

Blog: http://hiramchirino.com

Mime
View raw message