Here’s the top 20 fixes:
http://www-10.lotus.com/ldd/R5FixList.nsf/WhatsNew/2a05020a88eafd09852573770069fefd?OpenDocument
and the announcement letter:
Subject: 7.0.3 officially released - why so many regression
Hi,
Sounds nice with the new release. But am wondering; on the top 20 list 8 regressions are listed. This is almost half of the fixes.
From where do those regressions come - is this a sign of laziness or mistakes in the development process - that fixes are not being moved properly to the next release? Is just a shame that so many old problems - once fixed - are introduced when moving to a new release…
Kjeld
Subject: RE: 7.0.3 officially released - why so many regression
There are many reasons besides “laziness”. I would guess that the issue right now is developers trying to take the most important bug fixes from the Notes 8.0 effort and retrofit them back into the ND7 codestream. That is well worth doing, but often leads to regression issues in release such as this one.
As an aside, it would really help if when asking a question like this you did not leap to the worst possible conclusion. It is a perfectly fair question, but it comes off as rude and makes an answer less likely when you use words such as “laziness”, as you seem to be prejudging the people who are working very hard on this. I am not suggesting we should all always be polite and forgiving of IBM’s lackings, but just that there is no need for antagonizing people when they have in fact caught and fixed the problems.
Subject: RE: 7.0.3 officially released - why so many regression
Hi Ben,
Please forgive my “lasiness”. I dont want to put the developers in the position of being lazy. I know they arent and I appreciate the work they do. When I mention “lazy” it maybe to the process of moving code and fixed to new releases. - If it happens because people forget to document or if it is the app which cant transfer the changes.
It wasnt a conclusion - just a question why it happens.
But again - I am sure the guys do a hard job to give us the best code.
Kjeld
Subject: It’s perfectly normal that the Top 20 contains regressions, because…
… The Top 20 are the top 1.4% of fixes in terms of priority. All things being equal, these regressions should be given priority over new bugs of similar scope and impact. Ergo, the Top 20 should have a high percentage of regressions!
Out of the 1407 bugs on the fix list, there are clearly going to be some regressions. A more interesting question would be what percentage of all 1407 fixes are regressions, but I’m too lazy to take a go at answering that question. (Maybe Ben can shed some light on it, if/when he does his usual bit of magic analyzing and formatting the fix list.)
As to why there are regressions, it is because IBM programmers are normal human beings who are working on a massive code base that has gone through numerous releases over a span of 20 years; and when code evolves over time and many fixes are put in, the probability that the next fix will break a previous fix will always go up. This is as sure a thing as the law of Entropy.
-rich