> The biggest argument behind date-based versions is that they are simple and 
> easy, but are we giving up too much potential meaning for that convenience?  
> Ubuntu is both an amalgamation of many different products and in some ways a 
> "consumer" of software, so I think their versioning priorities are 
> understandably different.

No, I don't think we would give up too much potential meaning.  I think any 
special meaning that "encodes" software features in version numbers is mostly 
developer and system administrator centric, wheres a date-based scheme would be 
more easily consumed by the general users, and thus is more user or customer 
centric.  In my opinion, any information that would be "encoded" in a version 
number is just as accessible and useful when found in documentation.  So, maybe 
we give up some meaning, but not too much and it's not that important.

Also, if we use a self-explanatory stable versioning scheme that does not 
require a special decision every time a release rolls around, we are likely to 
avoid spending much time discussing this issue in the future, like talking 
about whether this or that architecture change warrants a major or minor 
version change, etc.  In my book, that's not very high return on time invested, 
so if we can avoid doing that twice a year, that's a win.

Alexey Lazar
PALS
Information System Developer and Integrator
507-389-2907
http://www.mnpals.org/

On May 17, 2012, at 13:40 , Dan Wells wrote:

> Hello all,
> 
> To answer the most immediate questions, I certainly agree that the next 
> version should be called 2.3 and come out in September.
> 
> Beyond that, I think I prefer traditional version numbering over Ubuntu 
> style.  I think Rogan is dead-on with his "anxiety" comment, but it can also 
> work the other way.  That is, there is something about a major version number 
> change which makes it easier get excited about and rally around.  It has a 
> pronounced "big-deal" feel which is almost completely lacking from Ubuntu 
> releases.  Then again, I suppose the actual goal might be to avoid "big deal" 
> releases, but where's the fun in that?
> 
> Another common use of major versions is implied external compatibility, and I 
> could see us possibly wanting that someday.  For instance, we might say that 
> if you write a VuFind connector now using public APIs, we will do what it 
> takes to not break it for 2.x, but when 3.x rolls around all bets are off.
> 
> The biggest argument behind date-based versions is that they are simple and 
> easy, but are we giving up too much potential meaning for that convenience?  
> Ubuntu is both an amalgamation of many different products and in some ways a 
> "consumer" of software, so I think their versioning priorities are 
> understandably different.
> 
> Dan
> 
> -- 
> *********************************************************************************
> Daniel Wells, Library Programmer Analyst [email protected]
> Hekman Library at Calvin College
> 616.526.7133
> 
> 

Reply via email to