Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
RE: [soa-iwg] ECF remote services

Title: Re: [soa-iwg] ECF remote services

My worry about this thread is that there are two very distinct issues that are getting conflated, and as long as that happens, I doubt there will be any progress and just heated disagreement.

 

1 – The process of working groups, packages, charters and distribution.  I don’t think this group or this thread is the right place for the conversation.  It needs to take place with the board Committer reps, at the Architecture council, at the EMO, etc.  As I’ve noted before, I don’t know if we’ll ever agree about how these work – but the processes and programs have been agreed to and built in past years after much consideration and debate.  If there are things that need to be addressed as you are saying – then the discussion needs to happen elsewhere.

 

2 – You are asking what looks like a very reasonable question to include some well regarded software in the SOA IWG Package.  I still don’t think anyone has said no to that – just that the IWG requires help with it’s integration and inclusion and it sure would be nice if there was some way that it fit with the other components.  The reason is, because as you’ve noted there has been a lot of disjoint SOA projects at Eclipse, and a very important goal of the IWG is to try to decrease that.  Just like the release trains have taken many years to increase the cohesiveness of the core, the SOA IWG is looking to do the same.

 

-          Don

 

 

From: soa-iwg-bounces@xxxxxxxxxxx [mailto:soa-iwg-bounces@xxxxxxxxxxx] On Behalf Of Scott Lewis
Sent: January 12, 2010 6:35 PM
To: SOA Industry Working Group
Subject: Re: [soa-iwg] ECF remote services

 

The SOA charter was poorly constructed.  It doesn't reflect the (obvious) truth that requiring a runtime project to do all it's own tooling (especially when existing tooling works perfectly fine...as in our case...let's not forget that), and requiring a tooling project to do it's own runtime just *doesn't make any sense*...except if the charter is written precisely to enable a single project or company to use the working group to it's benefit.

In other words, I would call this aspect of the SOA charter a load of crap.  *WHY* is the integration of tooling and runtime a fundamental guideline for the working group?  Why does this make sense when there are at least 5 runtime projects that are doing aspects of SOA, and at least 10 tooling projects that are also doing aspects of SOA?  To exclude any/all of these efforts because the SOA charter was poorly conceived doesn't make any sense.

So, for example of the absurdity of this:  would you also say that Equinox can't contribute runtime components to the SOA package because the Equinox team isn't doing tooling?  That EMF can't/couldn't contribute SOA tooling because they are not implementing the OSGi remote services specification (runtime)?  Then why is ECF's implementation of the OSGi remote services runtime standard excluded because of this?  (especially since as I've already pointed out we *do indeed* have tooling available at this very moment...e.g. PDE, DS, and our own).

Scott

Ricco Deutscher wrote:

Hi Scott,

>I would consider that a distinction without a relevant difference

I disagree. The integration of tooling and runtime was a fundamental guideline from the very beginning of this group. The Eclipse SOA charter (http://www.eclipse.org/org/industry-workgroups/soawg.php) says explicitly:

In order to provide incentives to A) form a coherent and integrated SOA platform, and B) to promote adoption, there will be a set of criteria ... This set of criteria will cover as well for the tooling and as well for the runtime. The fulfilment of either tooling only or runtime only will be not sufficient ....

Best,
Ricco


Am 12.01.10 23:17 schrieb "Scott Lewis" unter <slewis@xxxxxxxxxxxxx>:

Donald,

Donald Smith wrote:
> Hi Scott,
>
> The difference being that Virgo is a project, but the inclusions being
> discussed here are for the IWG Package.
>
> That is not a statement of disagreement to your recommendations for
> inclusions -- just pointing out the difference so we can focus the
> discussion.
>  

I would consider that a distinction without a relevant difference (IWG
packages and projects).

Why?  Because

1) projects are the entities that actually produce sw for inclusion in
packages...i.e. IWG doesn't produce any sw for distribution via
packages...they only decide what's included in packages from relevant
projects (in this case...in other cases EPP package contents are decided
in other, arguably even more mysterious ways).
2) projects have to focus in certain areas (runtime, tooling,
etc)...i.e. some things should/must be out of scope (that's only reasonable)
3) cooperation with/dependence on other projects (rather than vertical
project silos) is a good, even necessary thing...for the community and
for the projects

So sure, there's a definitional difference between 'IWG packages' and
'projects'...but WRT these points about sw creation and distribution by
*projects*, it's not a meaningful difference.

Scott

_______________________________________________
soa-iwg mailing list
soa-iwg@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/soa-iwg

 


 
_______________________________________________
soa-iwg mailing list
soa-iwg@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/soa-iwg
  

 


Back to the top