Please correct me if I missed it, but I'm puzzled as to why we haven't seen more arguments based on "lessons learnt" from when Nicholas Clark tried releasing perl 5.8.x on regular three monthly release cycles back in 2004. I would have thought this real-world experience would carry more weight than theoretical arguments.

From perl-5.8.2 perldelta:

Starting with 5.8.3 we intend to make more frequent maintenance releases, with a smaller number of changes in each. The intent is to propagate bug fixes out to stable releases more rapidly and make upgrading stable releases less of an upheaval. This should give end users more flexibility in their choice of upgrade timing, and allow them easier assessment of the impact of upgrades. The current plan is for code freezes as follows:

• 5.8.3 23:59:59 GMT, Wednesday December 31st 2003
• 5.8.4 23:59:59 GMT, Wednesday March 31st 2004
• 5.8.5 23:59:59 GMT, Wednesday June 30th 2004

with the release following soon after, when testing is complete.

From the description of Nicholas Clark's OSCON 2004 talk Perl 5.8.5 Was Boring (And Why You Should Be Excited By This):

There's a lot happening in Perl 5.8.x -- Jarkko Hietaniemi has passed the reigns over to a new pumpking, and with this comes some changes to Perl maintenance. But despite adopting "release early, release often," priority one is actually to make releases as uneventful as possible.

Last OSCON the current version of Perl 5.8 was 5.8.0; at abstract submission time the current version is 5.8.3, but by this OSCON 5.8.5 will be current. By adopting a time rather than feature-based release cycle Clark can be confident about when the release will be, but right now totally ignorant as to what will be in it (well, apart from no new bugs, and fewer of the old ones).

Come to this talk to find out how the new clockwork release cycle helps you to plan. See how far Perl maintenance has come in the past 12 months, learn where is it going, and discover why it affects you differently depending on whether you are an end user or an author of modules on CPAN.

I'd be interested in reading a post mortem "lessons learnt" from Nicholas Clark on the 2004 three monthly perl 5.8.x release cycles but I couldn't find one. The nearest I found was People of Perl: Nicholas Clark where he stated:

There aren't enough active volunteers to go round. Releases take a surprisingly long time to get right -- there's a lot of last minute faffing that seems impossible to automate, delegate or avoid.