Re: [ubl-dev] cva2xsl script

["G. Ken Holman" <gkholman@CraneSoftwrights.com>]

At 2012-01-09 00:43 +0100, Daniel Orasanu wrote:
> I have (again) a few questions:

All questions are welcome!

> Mr. Holman wrote in an Email:

"Ken" please (though now we have a plethora of Ken's on the TC membership!).

> ??
>
> When thinking last week about an upcoming code 
> list task at UN/CEFACT, I was reminded of 
> arguments raised during a similar development 
> process in the OASIS Code List Representation 
> Technical Committee: that the TC should *not* be 
> in the business of publishing running executable 
> code in support of its standards. The two main arguments are:
>
> (1) - who is going to maintain the code if it 
> breaks and the author is no longer on the committee?
>
> (2) - the TC should not be in competition with 
> any organization that may wish to write code to fulfill the same requirements
>
> One of the arguments for producing code was "but 
> this is a reference implementation" and that was 
> deemed insufficient to address the two concerns.
>
> The case study is the CVA to Schematron 
> conversion tool that I wrote, donated to OASIS, 
> and then repatriated back to Crane because of the committee decision:
>
> <http://www.CraneSoftwrights.com/resources/ubl/#cva2sch>http://www.CraneSoftwrights.com/resources/ubl/#cva2sch 
>
>
>
>
> ..?
>
> Here my Questions:
>
> 1)      He spoke about ?..because of the 
> committee decision..? is there any UBL TC 
> official decision why the script went back to Crane Softwrights Ltd.?

The decision did not happen in the UBL TC but in 
the Code List Representation TC.  I wrote the 
first implemented version of what was then called 
"Schematron-based Value Validation Using 
Genericode" and there were preliminary 
implementations of the process posted to the 
OASIS repository as part of the specification development.

This was the discussion that triggered the change 
from the specification of a validation process to 
the specification of a file format:

http://lists.oasis-open.org/archives/codelist/200708/msg00023.html

Within two months, in the minutes of the October 
9, 2007 meeting, the deemphasis of the process 
and the emphasis of the file format was minuted:

http://www.oasis-open.org/committees/download.php/25622/Meeting-20071009-Minutes.htm

The specification ended up with the name 
"Context/Value Association" for the file format, 
with no implication of any process stuff.

I then repatriated the code and put it in Crane's 
free resources.  It was available to all when in 
the OASIS repository, my updates have also been made available to all.

> 2)       ?..the TC should *not* be in the 
> business of publishing running executable code 
> in support of its standards..?. The script above is such an executable.

Right, which is what the Code List Representation 
Technical Committee did not want to see as part of the specification.

> How do I have to understand this script:
>
> a.       The script is free to use
>
> b.      AND  it is property of Crane Softwrights Ltd.
>
> c.       WHAT IF CraneSoftwrights decides not to 
> update this script on further UBL/CVA/Genericode features ?

The license allows the source code to be modified by others for any purpose.

>  ? this is pure hypothetical-no affront to 
> Crane Softwrights Ltd. J. Isn?t this a kind of 
> ?bottle neck? for the script and for the UBL-Validation?

No-one has come to the Code List Representation 
TC requesting any changes to either genericode or 
CVA file formats, so I don't see a 
bottle-neck.  There is *plenty* of time for 
alternative implementations of the process to be 
created by others if Crane's are deemed insufficient.

> I see  this script as the linking-piece between 
> UBL , CVA ,GC Files and validation process.

These scripts are being used by the committee in 
order to produce the committee deliverables.  A 
snapshot of the scripts are posted for archive in 
the TC repository (admittedly these are old; 
there was discussion recently with the editors 
that I should refresh the snapshot).  If Crane 
ever shuts down unexpectedly, the other TC 
editors have access to everything that is needed.

> Are any other (known | free | not-free) scripts who do the same job?

Not that I know of.

The CLR-Dev list (the equivalent of this list but 
for the Code List Representation TC) is very quiet:

  http://lists.oasis-open.org/archives/clr-dev/

No-one there has expressed any interest in developing an alternative.

> 3)      Can this script information be mentioned 
> in future UBL 2.1 documentation?

The specifications are mentioned, which should be 
sufficient.  There is no requirement to use 
Crane's code and I would not want it implied that it might be a requirement.

I'm not sure where the concerns are.  There 
should be by now enough copies of Crane's 
material out in the wild that if Crane were to 
disappear the TC would not be prevented from 
producing its deliverables.  Is that the nature of your concern?

I hope this reply has helped.

. . . . . . . . . . . . . Ken


--
Contact us for world-wide XML consulting and instructor-led training
Free 5-hour video lecture: XSLT/XPath 1.0 & 2.0 http://ude.my/uoui9h
Crane Softwrights Ltd.            http://www.CraneSoftwrights.com/u/
G. Ken Holman                   mailto:gkholman@CraneSoftwrights.com
Google+ profile: https://plus.google.com/116832879756988317389/about
Legal business disclaimers:    http://www.CraneSoftwrights.com/legal