Yep, I'm the one who quoted bug statistics. OS/2 version 3 was 3
million lines of code and was built and tested to IBM's highest software
process standards at that time (I don't know what they are now) ... one
undiscovered bug per 1,000 lines of code. That was considered very high
quality code but at 3 million lines of code that still leaves 3,000
undiscovered bugs. Actually, the 3,000 bugs wasn't the estimated total
but rather the total that would likely be discovered over the next 3
years following the release date. It's all statistics but pretty
accurate and well tested methodologies for the past 30 years.
IBM taught Microsoft how to do this stuff. The programmers hated it but
they nevertheless learned how to do it. Whether they follow now what we
taught them I have no idea but in those days Microsoft's code quality on
DOS and OS/2 was about the same as IBM's. However, they only did about
40% of the code and IBM did the rest so they had a smaller piece which
makes it easier.
Anyhow, it wasn't me that was refusing to install fixes under auto
update. I think that was Khen and I think not choosing auto update is
dangerous. Many years ago I would have given totally different advice.
The rules used to be don't update and take a fix for any bug that
isn't actually causing you a problem. The reason is that small bug
fixes and incremental changes may not be well designed and well tested.
In particular, they many not be rigorously system tested to check for
possible adverse interactions with other parts of the system. Any bug
fixes or feature adds are quite likely to introduce an unknown bug in
place of a few known ones that get fixed.
But the rules are different today. I think the vast majority of bug
fixes I've gotten have to do with security problems. Fix the XYZ buffer
overflow problem or sneaky nerds will break in and take over your life.
The problem is that this *is* a problem that could likely affect you
and the consequences can be dire. Therefore, the risk of adding another
unknown bug is, as Jack Aubrey was fond of saying, "just the lesser of
two weevils".
Chuck Norcutt
Moose wrote:
> Chuck Norcutt wrote:
>> And even two years late Vista still isn't completely ready yet for prime
>> time.
>>
> As long as we are being anecdotal....
>
> We know all apps, including OSs, are full of bugs, it's simply
> inevitable. In fact, aren't you the one who quoted statistics on that?
> So why would one not want to install the fixes to those that turn out to
> be causing trouble? I know you didn't suggest avoiding them, but
> someone(s) else in this thread did.
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|