[iaufwg] convention or standard

[Don Wells]

Doug Tody wrote:
> I would go one step further and suggest that if a convention is deemed
> useful enough to become a standard, it should be possible to consider
> significant revision as it goes through the standardization process.
> The resulting standard might then differ significantly from the original
> convention.  Merely registering a convention on the other hand, does not
> require any revisions other than possibly to the documentation.  - Doug
>   


There is a precedent for the process you describe.  XTENSION='A3DTABLE' 
was the prototype for what ultimately became 'BINTABLE'.  At that time, 
circa 1985, the process for registering names was primitive, but the 
designer (Bill Cotton) knew that extensions with local names were legal, 
even in interchange.   The name 'A3DTABLE' was chosen to be improbable, 
with the goal of avoiding collisions with other local names that might 
be chosen elsewhere.  It was also chosen to be a name that would not be 
used if the design ultimately evolved into a standard (as we hoped and 
expected).   The experience gained by the implementation and several 
years of production operation of the 'A3DTABLE' prototype in Classic 
AIPS was very valuable, and enabled the negotiators to be confident 
during the final negotiations for the BINTABLE standard circa 1991. 

-Don

PS:  As far as I know, all Classic AIPS systems everywhere can still 
read and interpret XTENSION='A3DTABLE' properly, although no FITS files 
containing extensions of that type have been written in almost 15 years.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dwellscho.vcf
Type: text/x-vcard
Size: 197 bytes
Desc: not available
Url : http://listmgr.cv.nrao.edu/pipermail/iaufwg/attachments/20061007/1d0335bf/attachment.vcf