Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: (ASCEND) 5.0Ap34



On 23 November 1997 21:14, Jason Eggleston [SMTP:jason@jet.net] wrote:
> Kevin Smith wrote:
> >
> > as we can with a lab environment). Only full releases, like 6.0 will
> > get a full regression test suite.
>
> Why?

I think that the need to ask this question says much in itself. A "full" 
regression test (that I must say normally is pointless, as problems still get 
through) is something that takes some considerable time to complete (weeks to 
months) and the number of tests and reports should reach (for a unit the 
complexity of the MAX) some hundreds or thousands of different tests in 
different environments.

The user community (us) need changes, fixes and new features much quicker 
than this, and so a "weekly" release cycle (or is it bi-weekly now Kevin ?) 
means that only the most rudimentary "functional" testing can take place. The 
idea of the patch branch is to change as little as possible, so that the 
regression tests done on the original release will be "mostly" valid, and   
only small, focused changes to the source base will happen. The incremental 
branch gets a quick "Does this boot up and connect calls ?" type testing, but 
it is down to the user to do the final satisfaction testing in their own 
environment. If you cannot afford either the time or the money to have your 
own adequate test facilities then get out of the kitchen, 'cause it gets hot 
in here.

Those users (typically corporates) who need a very defined set of functions 
should not be interested in incremental releases, just patches to fix known 
problems. In most cases this group of users can wait the 6 months or so 
between full releases anyhow.

Peter

++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>