|Ok, maybe I can put some order into all this:
1) Let's move the new candidate standards with updated schema and adjusted specification (DE 2.0, TEP 1.0) proceed through the approval process;
2) Validate the "Extension" mechanisms and get approval of the concepts by the TC.
3) Add the Extension mechanisms to Common Types and get formal approval from the TC.
4) Publish "How To" / "Best Practices" documentation
5) From here on any standard can use Extensions:
5.1) For new standards, the work done for TEP or DE can serve for guidance.
5.2) To allow for the use Extensions in the spirit of "layers" ("Community Extensions") in existing standards, a "small" addition to the schema is required - an Errata could be sufficient to accomplish this.
5.3) To allow for the full set of extension capabilities (i.e. "list augmentation", "list replacement" and "list reassignment") in existing standards, more substantial changes are needed in the schemas. As
our work with TEP has shown, migrating from WD03 (without Extensions) to WD04 (with extensions), such changes can lead to significant simplifications with repercussions in the schema and the specification. In most cases these will be substantial changes that
can not be accomplished with a simple Errata.
From: email@example.com [firstname.lastname@example.org] on behalf of Rex Brooks [email@example.com]
Sent: Wednesday, April 10, 2013 1:32 PM
Subject: Re: [emergency] Thoughts on EDXL-DE and SitRep
Okay, I'll give it some more thought, but I'm not happy about waiting.
On 4/10/2013 12:11 PM, Wilkins, Brian M wrote:
Extensions should be part of common types so it would be available to any standard using common types. However, it is still up to any individual standard to allow extensions or not. If they are not part of the
schema for the standard, then they are not part of the standard and any adopter trying to add them to a given standard will break interoperability. This in itself explicitly states whether or not Extensions are part of a standard. If this change is implemented
as being suggested, this will not be a minor change to the schema, satisfied by an Errata. This will remove the choice elements and add a whole new element in possibly many places in addition to adding potentially multiple sets of enumerations. Someone that
implements DE2 as it is today will have to make significant changes to their code base to implement DE X with extensions, or SitRep for that matter. As both an implementer and a standards member, I think it is in the best interest of our communities to wait,
make the changes, and help limit that pain/cost of implementation. IMHO.
There is an old saying that if it isn't forbidden, it's mandatory.
That is a deliberately extreme notion, but, to be honest, I don't think we need to expicitly state that extension can apply to SitRep and DE, or HAVE and RM for that matter. If we make the change to the edxl-ct, edxl-gsf or edxl-ciq it applies and all we really
need to do is to provide some guidance.
I absolutely oppose pulling back on SitRep-v1.0 or DE-v2.0.
I could have been equivocal, but where's the fun in that?
On 4/10/2013 11:18 AM, Elysa Jones wrote:
We have worked hard getting SitRep1 and DE2 ready for an organizational vote and we currently have a ballot open for the purpose of approving these works as a Committee Specification. If this passes, the next steps
are obtaining statements of use then agreeing (again by special majority) to begin the 60-day review prior to organizational wide balloting.
Discussion over the past several months regarding a “standard” extension method for all of our specifications has been ongoing. These discussion are best documented in the RIM-SC. TC members have been made aware
of this discussion and invited to participate and stay in touch as it could affect all of our work. I’ll not repeat any of that discussion here but ask you to keep it in mind as you make your decision about SitRep1 and DE2.
You may recall there were two versions of the DE2 being worked – one with the layer-extension concept and one without. We decided to go forward without at this time but all agreed we should keep it in mind and
find some way to incorporate later - perhaps through common types. Some of our members are concerned with releasing DE (as well as SitRep) without this extension concept and then trying to make a change later would be problematic. We all know how long changes
can take and we are anxious to get this work out. We learned in the last couple of weeks that this extension method would solve some issues with TEP as well. So, after considering putting this in TEP, it has re-opened the discussion about DE and SitRep in
The bottom line here is that we have members that are re-thinking whether we need to put either of these standards out without this extension concept. We have sufficient votes today to move the as they are to Committee
Specification. The ballot closes tomorrow. If we wish to add “extensions” to DE2 and SitRep, we will have to make not only changes to the document and back through public review again.
Personally, having a “standardized” extension method across all our standards is very appealing to me. However, delaying SitRep and DE2 concerns me if it is going to take another year. If it can be completed in
6 months, it concerns me a little less - but only if sufficient resources are available to turn this around quickly. Basically DE2 has already been done and there is a WD version with it included. I’m not sure about how much time it would take to add it
Please respond to this note with your thoughts.
Elysa Jones, Chair
Berkeley, CA 94702
Berkeley, CA 94702