TAPRegExt erratum, Identifiers for Obscore

[Paul Harrison]

On 2013-12 -11, at 09:43, Mark Taylor <m.b.taylor at bristol.ac.uk> wrote:

> On Wed, 11 Dec 2013, Paul Harrison wrote:
> 
>> 
>> On 2013-12 -11, at 07:28, Markus Demleitner <msdemlei at ari.uni-heidelberg.de> wrote:
>> 
>>> Bob,
>>> 
>>> [fixing an IVORN in an example]
>>> 
>>> On Tue, Dec 10, 2013 at 09:45:56PM +0000, Robert J. Hanisch wrote:
>>>> I think it would be more consistent with the versioning system we have
>>>> been using for nearly 10 years to make an incremental (0.01 delta) release
>>>> of the document, note the errors that have been corrected in the change
>>>> log, and fix the problem in the main document.  If people don't bother to
>>>> look at the change log / erratum they will go on implementing the wrong
>>>> thing.
>>> 
>> 
>> The way that I understand section 1.2 of http://www.ivoa.net/documents/DocStd/20100413/REC-DocStd-1.2.pdf - the 1.x number refers to the version of the underlying protocol so for simple typographic style changes that do not change the intended functionality of the described protocol/model etc., the “1.x number” part of the document does not change, but merely the date. In this way there are none of the consequent issues that you outlined.
>> 
>> Ironically REC-DocStd-1.2.pdf does not follow its own recommendation, but for instance http://www.ivoa.net/documents/VODataService/20101202/REC-VODataService-1.1-20101202.pdf does  by formally including the date as part of the document file name.
> 
> You're missing this paragraph:
> 
>   The final published and approved Recommendation retains the date
>   on the title page of the document, but the date is removed from the
>   document filename in order to simplify reference to the document.
> 
> Document names for REC documents, unlike those for other stages in the
> process, do not contain dates.


Ah  - the vast majority of published REC documents since the new naming conventions came into place do actually include the date in the name, so I imagine that people did what they remember being agreed in the TGC rather than what got into the document. 

However, if errata for RECs are to be a separate document, then obviously it is fine not to have the date in REC, but if they are to be put in one document, then it needs the date.

Personally I think that it was a retrograde step to move from HTML to PDF as the mandated IVOA format, as PDF encourages people to take a local copy of the document (and consequently possibly keep referring to an out of date version) - if HTML were the only format for these documents (which is what W3C do) then the “latest” version of the document is always the one retrieved by the URL, and these naming issues become less important.

Cheers,
	Paul.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2774 bytes
Desc: not available
URL: <http://www.ivoa.net/pipermail/dal/attachments/20131211/42103ea6/attachment.bin>