Date:         Fri, 2 Oct 1998 12:02:58 GMT
Reply-To:     Howard Schreier <HSchreier@USA.NET>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Howard Schreier <HSchreier@USA.NET>
Organization: Institute for Global Communications
Subject:      Version Numbers (was: DROP A COLUMN ...)

On Wed, 30 Sep 1998 08:59:23 -0400, Douglas Dame
<dougdame@HPE.UFL.EDU> wrote:
 
[lines deleted]
 
>Along this same line of thought, there was a thread not too long ago about
>someone wanting to be able to show end-users some data norms on variables
>contained in a warehouse type SAS app, such as the mins, maxes, means,
>zeroes, missings, etc. This is also content that seemingly could be most
>usefully accommodated within the structure of the dataset header. (Although
>I think I would opt to have that processing be turned off as the default, we
>certainly wouldn't want the overhead of computing all that stuff routinely
>for every throw-away work dataset created during a job.)
>
>Maybe something we can hope (and/or lobby for) in SAS version 7.02 ???
 
It would be more like Version 9. Really. It seems that the Institute
is going from one extreme to the other in its numbering of versions.
 
Version 6 to Version 7 took what? About 12 years, I think, with lots
of decimal increments and TS-level qualifiers in between.
 
Now, as I understand it, V. 7 is out the door or nearly so. V. 8 will
follow a few months later, but is limited to taking care of loose ends
and problems from V. 7. So new development initiatives now are aimed
at V. 9.
 
I personally would like to see numbering more like what other software
firms do; integer increments for significant design or functionality
changes, fractional or other clearly subordinate increments for
maintenance releases, added platform or device support, etc. The more
extreme practices are just uninformative to me.