Daniel wrote:
> Spot on, Chris, and I do my little bit of testing, *as part of the
> community* , by running the Beta versions as they are released (still
> using 2.21b2, as shown in my sig, below, whilst waiting for the Council
> to release 2.22b1).
And yet you, and all the other beta testers, failed to identify this
problem before 2.21 went live. Which indicates not that you, and they,
are not providing a useful service, but instad simply confirms Philip
Chee's statement that there /is/ no formal regression-testing program,
and there never will be, as a result of which each and every new
version of Seamonkey is capable of being fundamentally flawed.
Chris argues : "[regression testing] is rarely conducted by developers
("developers" meaning the people that wrote the code)". In a world
where the sole aim is to produce a new version every $n$ weeks, for
small $n$, regardless of whether or not this new version is (a) wanted
by its potential users, or (b) offers any significant improvements over
the previous version, I have little doubt that the statement is true
(it could hardly be otherwise). But in a world where quality and
stability matter, and where the developer(s) take a pride in his/her
/their work (e.g., TeX), exactly the converse obtains; can you really
imagine Knuth releasing a new version of TeX without first being
99.999999% certain that it is 100%-compatible with the previous
version except in those areas where new and significantly improved
functionality is offered ?
Philip Taylor
_______________________________________________
support-seamonkey mailing list
[email protected]
https://lists.mozilla.org/listinfo/support-seamonkey
