[DDI-users] NewVersionType

[Wendy Thomas]

Hi Alerk,
Shall we say this is a victim of German over-engineering? Here is what
I hope is the final decision on this.

Version 3.0
version is NewVersionType which is an infinately long string of
numbers delimited with "."

Version 3.1
In an attempt to more uniformly structure the URN and continuing to
use "." to separate objectType.id.version the version number in 3.1
became a prescribed three section object using numbers and delimiting
with ".". This is the URN structure that Achim provided and that was
agreed upon for 3.2
urn=“urn:ddi:us.mpc:VariableScheme. VarSch01.1.4.0:Variable.V1.1.1.0”

Version 3.2...(the current state of the decision)
This describes the state of play as of Nov 2010 and the use of a : for
separation of objects has been retained in later discussions as has
the added flexibility for the version number.
Tighter constraints on the structure of ID, Version, and Agency in 3.1
were based on the use of a “.” to separate the parts of an object
description.
Changing the “.” separator between the object name, object id, and
object version to “:” allows greater flexibility in the content of
each individual item
The proposed relaxation of ID, Version, and Agency structures better
supports current software and business practice

This describes the state of play as of Nov 2010 and the use of a : for
separation of objects has been retained in later discussions as has
the added flexibility for the version number.

So as it currently looks for 3.2 we will be back to the structure from
version 3.0
At minimum it will no longer REQUIRE a 3 part numeric
It may limit the number of extensions (but it would probably be a
limit unlikely to be used like 10 or something)....I'm waiting on the
final paper. The only reason I can see for that is technical push back
that unlimited is too difficult to handle.

Are we all clear about this moving target? Note that the
"elementVersion" added to GLOBAL in 2.5 is a simple xs:string.

Hope you were sitting down while reading this.

Wendy

On Fri, Mar 9, 2012 at 3:53 AM, Alerk Amin <A.Amin at uvt.nl> wrote:
> Hello,
>
>   I have a question about versions of elements in DDI.  In DDI 3.0, it
> seems like a version could be arbitrarily long (eg. 1.0.0.0.0.0.1).
> However, in DDI 3.1, it looks like a version must always be 3 numbers
> long (eg. 1.0.0).  I am getting this info from NewVersionType, which is
> used by AbstractVersionableType.
>   My question is: am I understanding this correctly?  I recall a
> discussion about this during the DDI Developers meeting in Utrecht in
> December 2010, where the the DDI 3.0 NewVersionType was defended because
> of a use case that required 4 numbers.  Is 3 numbers the new "accepted"
> format, and will it stay that way in DDI 3.2?
>   Just to be clear: I am happy with the 3 number format.  I just want
> to be certain I understand what is going on before I start writing code
> that relies on it.
>   Thank you for your help.
>
> Best,
> Alerk
>
> --
> -------------------------------------------------------------------
> Alerk Amin
> Development Manager
> CentERdata
> Room T418 (Tilburg University, Tias Building)
> Postal address   : PO Box 90153, 5000 LE, Tilburg, The Netherlands
> Visiting address : Warandelaan 2, 5037 AB, Tilburg, The Netherlands
> Telephone        : +31-13-466 2243 / 8325
> Fax              : +31-13-466 2764
> WWW              : www.centerdata.nl
> Disclaimer       : See http://www.centerdata.nl/maildisclaimer.
> _______________________________________________
> DDI-users mailing list
> DDI-users at icpsr.umich.edu
> http://lists.icpsr.umich.edu/mailman/listinfo/ddi-users



-- 
Wendy L. Thomas                              Phone: +1 612.624.4389
Data Access Core Director                 Fax:   +1 612.626.8375
Minnesota Population Center             Email: wlt at umn.edu
University of Minnesota
50 Willey Hall
225 19th Avenue South
Minneapolis, MN 55455