OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

ubl message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Candidate decision regarding code list support publishing

In today's face-to-face we spent a good part of the afternoon talking 
about the new additional code list resources and the code list revision policy.

To prepare to make decisions, this post summarizes where I think we 
reached in our discussions.  We will decide on this later on during the week.

(1) - I have prepared context/value association files used to 
recreate the and I have published these to the Kavi directory, but 
during the meeting I did not remember that these actually have been 
published; they can be found here:


(2) - the code list support files will be more useful if they are 
posted to the UBL docs repository, since their relative XPath 
addresses can point to the actual code lists
  - these support files are not for all users of UBL, and in fact are 
not for many users of UBL ... most users will be using either 
defaultCodeList.xsl or the XSLT stylesheet that users of the support 
files create for their use

(3) - as new code lists are produced, it makes sense to publish these 
lists and their support files as subdirectories of the UBL docs 
repository so that at any time a developer can download from the 
repository *all* versions of *all* code lists in the UBL history so 
that a developer can create their own context/value association files
  - it is not necessary to duplicate the publishing of unchanged 
lists since all of the subdirectories are in the UBL docs repository
  - for any support release of new code lists there will be a new 
version of defaultCodeList.xsl published so that UBL users can always 
obtain the latest default second-pass validation without having to 
use the methodology to make it

- there is currently a UBL 2.0 code list directory:


- two of those code list files are the country code list and the 
transportation status code list:


- there is a fault in the country identification code list for 
Bosnia, so we should publish a repaired version as an erratum
- say we decided to publish a new code list support package in 
February 2007 (a past date for illustration purposes) for the erratum 
... we would have the new code list, the new context/value 
association file that overrides the old context-value association 
file, and the new resulting defaultCodeList.xsl in a revision subdirectory:


- following that date there might be a new transportation status list 
(the current version is "Third Revision", the next version we might 
want to call "4" to use numbers instead of words for the meta data) 
that has new values
- say we decided to publish a new code list support package in March 
2007 for the addition ... this would again have the required support 
files, a new context/value association file that *supplements* (not 
overrides) the old context-value association file, and the new 
resulting defaultCodeList.xsl in a revision subdirectory:


- note that the cl-200703.zip file would contain the cl-200702 and 
cl-200703 directories, such that a developer need only download the 
latest ZIP file to get *all* of the code list directories since the 
publishing of os-UBL-2.0 ... this way each dated directory need only 
have the changed files and not all files
- in the above examples we would not be limited to only changing one 
list at a time:  any time we publish a new code list dated directory, 
it would contain all of the changes for that date
- users wanting the latest defaultCodeList.xsl just goes to the 
latest dated subdirectory and obtain their file ... nothing else is 
needed for them to get default second-pass validation
- developers wanting the latest materials to make the latest 
defaultCodeList.xsl would download the latest ZIP file and the 
os-UBL-2.0.zip package ... unzipping both of these in their 
development directories ensures the relative paths are all correct

In preparing this I have these questions we need to answer:

(Q1) - when do we set this up and get started?
      - genericode 1.0 is about to go to Public Review
      - the methodology will be finalized at 1.0 only after 
genericode 1.0 is finalized
      - should we wait to set all this support stuff up once both are 
finalized at 1.0?

(Q2) - how often do we publish the code list directory?
      - perhaps we can schedule new code list directories every 6 
months (4 months?) from December 2007 which should give us the time 
to shake out the approach before then and it is on the first 
anniversary of the specification's release

(Q3) - does it make sense for a developer to find everything in the 
latest ZIP file, or should the above be changed such that there is no 
duplication and each ZIP directory has only the files in that 
directory:  this makes more downloads for the developer, but for the 
developer who is following development step by step they only 
download the latest and unpacking it won't overwrite any of their 
existing files ... the more I think about it, the more I think each 
ZIP file should only have the files of that directory

(Q4) - note that the genericode filenames used above are ambiguous 
... is this a problem?  The files are distinguished by the 
subdirectories in which they are found, and by the meta data found 
therein ... do we need to go the extra step of distinguishing the 
file names, and if so, how do we do this in a consistent fashion?  My 
gut feel is that we do not and that directory context is enough ... 
if a file is found out of context, the contents reveal the meta data

My initial assessment is that all of the above is achievable and will 
mechanically work, but is it too complicated?  I think not since we 
will always have a simple defaultCodeList.xsl that anyone can 
download with all of the hard work done.  Anyone who needs to use the 
validation methodology will already be familiar with how the 
methodology works.  Having these files in multiple directories 
shouldn't scare them off since all of the relative directories work.

Thanks to everyone for their input at today's meeting ... it was very 
productive and the questions were very astute.

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

Upcoming hands-on training(Europe 2007): XSL-FO Jun 11; UBL Oct 01
World-wide corporate, govt. & user group XML, XSL and UBL training
RSS feeds:     publicly-available developer resources and training
G. Ken Holman                 mailto:gkholman@CraneSoftwrights.com
Crane Softwrights Ltd.          http://www.CraneSoftwrights.com/o/
Box 266, Kars, Ontario CANADA K0A-2E0    +1(613)489-0999 (F:-0995)
Male Cancer Awareness Aug'05  http://www.CraneSoftwrights.com/o/bc
Legal business disclaimers:  http://www.CraneSoftwrights.com/legal

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]