Re: [MEDIACTRL] Does a Control Package name include the version number

I say (as an individual) "Nay, nay": Packages do whatever they want.  Backwards compatibility is a nice fairy tale, but not anything we really care about.  I offer packages are arbitrary, opaque names. If they *happen* to end with /1.0 or /9.2993, great.  However, I would offer the protocol should not care or impose any semantics about what the /mumble.foo means.

On Jul 7, 2010, at 3:47 PM, Chris Boulton wrote:

> Hi All,
> 
> As part of our IESG review the following has been identified:
> 
> Section 8.1
> states:
> The package name MUST also register a
> version number for the package which is separated with a '/' symbol
> e.g. package_name/1.0.
> This could be taken to mean that (1) "package_name/1.0" is the complete
> package name or (2) "package_name" is the mere package name and "1.0" is
> the package version number. The ABNF appears to imply (1) because it says:
> 
> Packages = "Packages:" SP package-name *(COMMA package-name)
> package-name = alpha-num-token
> alpha-num-token = ALPHANUM 3*31alpha-num-tokent-char
> alpha-num-tokent-char = ALPHANUM / "." / "-" / "+" / "%" / "=" / "/"
> 
> This appears to allow "/" in the "mere package name". If so, how will
> software parse the complete package name to determine the package
> version number? Will it parse the complete package name from the end and
> split the name at the first "/" character it finds? Can the version
> number include a "/" character? The ABNF does not appear to disallow that.
> 
> 
> What do we want to do with version?  I would suggest that we soften the MUST requirement.
> 
> Chris.
> 
> 
> -- 
> Chris Boulton
> CTO & Co-founder
> NS-Technologies
> m: +44.7876.476681
> _______________________________________________
> MEDIACTRL mailing list
> MEDIACTRL@ietf.org
> https://www.ietf.org/mailman/listinfo/mediactrl
> Supplemental Web Site:
> http://www.standardstrack.com/ietf/mediactrl