version numbers

Tony Linde ael at star.le.ac.uk
Wed Mar 10 01:16:15 PST 2004


Agreed. Well stated, Ray.

Tony.

> -----Original Message-----
> From: owner-interop at eso.org [mailto:owner-interop at eso.org] On 
> Behalf Of Ray Plante
> Sent: 09 March 2004 18:24
> To: interop at ivoa.net
> Subject: version numbers
> 
> Hi VOers,
> 
> Now and again I see someone asking about the logic behind a 
> version number, wondering how vZ.Z.ZZ came after vX.X.  I'm 
> sure I've confused my fair share of people with my own 
> choices.  I would like to recommend that we practice the 
> pattern typical of most revision control systems (that I am 
> aware of), as in the following example of a sequence of version
> numbers:
> 
>      0.1         // an initial pre-production draft version
>      0.2
>      0.2.1       // a minor revision on 0.2
>      0.2.2
>      0.4         // perhaps 0.3 never saw the light of day
>      0.8
>      0.9
>      0.10        // we're not ready for 1.0 yet
>      0.11
>      0.11.1      // further refinement
>      0.11.2   
>      1.0         // yeah!  ready for prime-time!
>      1.1         // revision on 1.0
>      2.0         // major revision over 1.x; applications may break.
> 
> This is based on the following principles:
> 
>   o  Each field between a period (.) is an integer.  The 
> version is not 
>      to be interpreted as a floating point number.  That is 0.91 is 82
>      revisions after 0.9.
> 
>   o  When the first field is 0, the standard/app is a draft release 
>      with significant development or review still expected.  
> 
>   o  The third field is a branch in development.  In our 
> case, we should 
>      use this to mark minor changes (perhaps backward 
> compatible) beyond 
>      the last two-field version. 
> 
>   o  An increment in the first field represents a major 
> change; users will 
>      likely have to change their applications to be compliant 
> with the new 
>      version.  
>   
>   o  Which field is incremented--that is, whether this is a 
> minor or major 
>      change--is up to the authors.  
> 
> cheers,
> Ray
> 
> 



More information about the interop mailing list