Sanjiva

I think this is an interesting option. I agree it makes sense. On the
other hand its a big change. I'd like to hear some more input on this.
One concern I have is that maybe this needs a fresh email chain to get
peoples attention.

I guess my 10 cents is that there isn't much difference between:

<definitions>
  <xxx>
  <yyy>
</definitions>

<define-xxx>
<define-yyy>

As for <foo-bar> versus <FooBar> or <fooBar> I prefer <foo-bar>.

Paul


On 2/28/07, Sanjiva Weerawarana <[EMAIL PROTECTED]> wrote:
Let me see whether I can recap where we are now .. kinda lost in the
thread. Sorry for being slow to jump in. I'm going to propose some mods
too in what follows:

(Warning: at first this may look very different but all I've done is
renamed elements to use verbs and nouns correctly .. verbs are for actions
and nouns for definitions.)

<synapse>
   <!-- external registry to use to look for stuff when stuff is not
        found in the local registry -->
   <define-registry>
     .. info needed to define the parent registry ..
   </define-registry>?

   <!-- insert a named endpoint to the local registry -->
   <define-endpoint key="string">
     .. content as defined earlier for enhanced endpoints ..
   </define-endpoint>*

   <!-- insert a named sequence into the local registry. The sequence
        names "main" and "fault" are special as proposed in this
        thread. -->
   <define-sequence key="string">
     ..
   </define-sequence>*

   <!-- add an arbitrary thing into the registry -->
   <insert-registry key="string" [src="url"]>
     string or xml
   </insert-registry>*

   <!-- define a proxy service -->
   <define-proxy-service ..>
   </define-proxy-service>

   <!-- mediators to execute as the default set of rules, unless "main"
        sequence is defined in the registry. If "main" is defined then
        these mediators will be ignored. (??) -->
   mediator+
</synapse>

(I'd like to look at all the mediator names and make sure we name them
with verbs but again that's cosmetic changes.)

I also don't see the need to force users to write all the definitions
first and then the mediators. As we parse this XML, all we need to do is
put any mediator statements into the "main" sequence and execute all the
define-* things. Then when messages start flowing we're good to go.

This mail prolly doesn't make much sense yet (its late so it may not make
sense ever!) so fire away.

Ruwan I'm fine with fooBar instead of foo-bar too. Let's just pick one way
for everything!

Sanjiva.

Ruwan Linton wrote:
> Are we agreed on the configuration language changes? Any more
> suggestions, Comments? So that we can go ahead in the implementation....
>
> BTW: I propose the the format *inSequence* over *in-sequence* for both
> attribute names and tag names through out the whole configuration so
> that it will be consistent.
>
> Thanks,
> Ruwan.
>
> On 2/27/07, *Asankha C. Perera* <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED]>> wrote:
>
>     Hi Saminda
>      > Registry definition just provide a URL, impl class  and a  bunch of
>      > parameters to connect to the registry right. So, IMHO it doesn't
>      > matter whether the registry is remote or local, we should be able to
>      > use the same syntax to access them otherwise IMO it would be pretty
>      > complex for a user.
>     Exactly! So the idea is that a user should be able to talk to "a"
>     registry and get everything done. This registry may just be totally
>     local to the configuration, totally remote or a mix of the two. Going on
>     these same lines, think of a registry that is able to talk to more than
>     one other (remote) registry.. like cascading registries! In Synapse
>     this
>     could all be hidden. You will define your registry implementation and
>     configuration parameters, and all "key"s will be resolved by this
>     registry. The fact that some local resource definitions may override a
>     few remote ones would thus not be a problem.
>
>     On your earlier email:
>      > In synapse.xml for configuring proxies or others , for elements, if
>      > the name contains virtually multiple words, it has been written as
>      > <foo-bar/>. When it comes to attributes it's written as carMar. IMHO,
>      > if it can be given as <foo-bar car-mar="something"/> wouldn't this be
>      > easy to user to remember, Or <fooBar carMar="something"/>, rather
>      > mixing the way we write element and attribute names ?
>     Yes, Agreed.. lets get this cleaned up along with the other refactoring
>     that we are doing right now.
>
>     thanks
>     asankha
>
>
>     ---------------------------------------------------------------------
>     To unsubscribe, e-mail: [EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>
>     For additional commands, e-mail: [EMAIL PROTECTED]
>     <mailto:[EMAIL PROTECTED]>
>
>
>
>
> --
> Ruwan Linton
> http://www.wso2.org - "Oxygenating the Web Services Platform"

--
Sanjiva Weerawarana, Ph.D.
Founder & Director; Lanka Software Foundation; http://www.opensource.lk/
Founder, Chairman & CEO; WSO2, Inc.; http://www.wso2.com/
Director; Open Source Initiative; http://www.opensource.org/
Member; Apache Software Foundation; http://www.apache.org/
Visiting Lecturer; University of Moratuwa; http://www.cse.mrt.ac.lk/

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]




--
Paul Fremantle
VP/Technology, WSO2 and OASIS WS-RX TC Co-chair

http://bloglines.com/blog/paulfremantle
[EMAIL PROTECTED]

"Oxygenating the Web Service Platform", www.wso2.com

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to