Title: Incorporating TC Name
Attached is another pass over this. I went
back to look at the issue resolutions for i015, i016 and i017 to get to this. I
think the only deviations from those is that we have, at OASIS staff request,
added the tcShortName to the namespace (and accordingly used ws-rx that is
assigned rather than wsrx).
There are of course the things that were
never decided like locations and the filenames for the wsdl/schema. I think we
should just use the namespace locations. That way we’ll keep the dates
out of the filenames. Here I proposed using short filenames based on product id
and version that do not include the artifact type as I think the file extension
is enough. I’m open to whether or not they need to include the
corresponding CD and version number in the name. I think by placing these files
in a location that matches the namespace with the corresponding spec version it
should be fine. It would be even better with a RDDL doc at the namespace to
explain it all. Finally a RDDL doc at http://docs.oasis-open.org/ws-rx/wsrm
could point to each CD version all the way to the final standard.
Below is basically a short summary of what
is in the doc (it’s getting horrific with track changes already).
Location always = namespace.
Schema:
File name
wsrm-1.1.xsd
Namespace
http://docs.oasis-open.org/ws-rx/wsrm/yyyymm/
WSDL:
File name
wsrm-1.1.wsdl
Namespace
http://docs.oasis-open.org/ws-rx/wsrm/yyyymm/wsdl/
From: Patil, Sanjay
[mailto:]
Sent: Tuesday, November 08, 2005
6:11 PM
To: Marc Goodner; Gilbert Pilz;
Subject: RE: [ws-rx-editors]
Incorporating TC Name
OK. Not using the [tcShortName] as part of
the product identifier/name is fine with me. Besides, this will make
the issue of hyphen as part of the [tcShortName] moot.
Please see attached an updated version for
the metadata, etc. Specifically, please take a look at the identifiers and
names of the schema and wsdl instances. I am not so sure whether you all agree
with including the year-month token as part of the schema and wsdl
identifiers/names.
Thanks,
Sanjay
From: Marc
Goodner [mailto:]
Sent: Tuesday, Nov 08, 2005 15:11
PM
To: Gilbert Pilz; Patil, Sanjay;
Subject: RE: [ws-rx-editors]
Incorporating TC Name
+1
I do not agree that the TC short name
should be part of the artifact identifier. OASIS has a requirement for unique
product names which we have met. That combined with a location that includes
the TC short form name should be sufficient.
Sanjay, to your question of wsrx vs. ws-rx
I think it needs to be ws-rx as it is supposed to be the short form of the TC
name.
Can anyone point me to the committee page
for that is defining the AIR guidelines? The link to the latest for the draft
Gil points to below (which is the only one I’ve ever seen) has had
it’s review period expire and the latest url link they provide returns a
404 (of course).
From: Gilbert Pilz
[mailto:]
Sent: Tuesday, November 08, 2005
1:49 PM
To: Patil, Sanjay;
Subject: RE: [ws-rx-editors]
Incorporating TC Name
I'm not sure why I feel compelled to point
out that the OASIS Artifact Identifciation Requirements 1.0 does not specify
the use of the [tcShortName] in the ArtifactIdentifier. Line 386 of http://www.oasis-open.org/committees/download.php/13344/ArtifactIdentificationRequirements-v1.0-wd-15.pdf lays
out the formula as:
product-productVersion-artifactType-stage-revision-language.form
I really don't care if we include the
[tcShortName] in the ArtifactIdentifier but I am a little miffed that OASIS
would direct us to a set of guidelines that say one thing and latter come back
and say another. At the very least they should take an action item to update
the guidelines so they don't have to keep doing this to every new TC that makes
the mistake of following the current guidelines.
- g
From: Patil,
Sanjay [mailto:]
Sent: Tuesday, November 08, 2005
10:15 AM
To:
Subject: [ws-rx-editors]
Incorporating TC Name
This
is regarding the AI I took to work with the editors and create a new issue for
incorporating TC name in our metadata, artifacts.
Please
take a look at the attached proposal. Consider this only as a starting point if
you may.
<<wsrx-air.doc>>
Some
background comments:
I
was debating with myself whether the TC Name should be wsrx or ws-rx in our
metadata. I remember that on the recent call with Jamie, we decided it to be
ws-rx. However I don’t remember if there was a good reason for this when
we made a decision, or we just didn't pay enough attention to this matter. I
personally think that using "wsrx" would be syntax-friendly for
constructing other structured strings.
I
was also not sure whether we have discussed and decided the location strings.
Are we following the convention for location to be - namespace URI + document
identifier.[format]?
Comments?
We will obviously be presenting this matter for the TC's deliberations by
opening a new issue with a concrete proposal. Before that, it will be good to
brainstorm with this focused group ….
Thanks,
Sanjay
wsrx-air sp 3.doc
