With the recent pulling of the 5.4.3 release due to an issue and looking at some of the other past monthly releases that also broke functionality/3rd party modules. I think it might be time to reassess the monthly release schedule.

Whilst I like the intention of monthly releases. They should undergo sufficient testing to ensure nothing breaks and not just the core provided services. And if there is a breaking change enough time and warning for the third party developers should be given.  This doesn't appear to be happening currently.  As a hoster I am seeing both an increased workload to update on a monthly cycle as well as increased support calls when things go wrong.

I would also be interested to hear from 3rd party module developers on their thoughts. I would think that they are currently spending a lot more time on checking (or trying to check) their modules for the monthly releases at the cost of developing new functionality.

As a suggestion would it be better to alternate between beta and production/stable releases on alternating months. Exception being if there is a security issue of sufficient import.

So what does everyone think?

Antony

I agree with Joe on this one.  While there is always the chance that a release will have some major flaws in it, this is something that can be fixed by better quality assurance rather than slower release cycles.  The slower the release cycle, the more changes and the bigger the delta between releases.  As long as problems are found early and communicated well, I don't see the drawback of the occasionally DOA release based on regular releases.    If anything, the 5.0 release showed that until a release is put up live, most people don't test them.  If you've got a mission-critical app, then it pays to be a few releases behind and choose which one you want to move to carefully.

The difficult part of the equation is coming up with a solid QA system, and I suspect the eventual solution is a mix of widespread testing by third party app developers (something I know DNN corp is developing) and possibly automated QA/Smoke Test type situations.

The bottom line is that releasing new versions of a product with an install base in the 00,000's is always going to be difficult and there are going to be some problems. To me the solution is tying better automated and compatibility testing into the release process rather than slowing it down, which just puts more pressure on individual releases to be perfect because you have to wait longer to get them.

I agree - keep the Release Early, Release Often attitude...

I personally do not upgrade to the latest version unless I absolutely have to - it needs to be tried and tested first for me to put it on a live site, 5 out of 5 of my last upgrades over the last 2 years have failed.

It's a pity as this is not the case in other software markets (iphone apps, Apple OS, Windows updates, game updates) but DNN releases do not have a great history to be completely honest.

---

Entrepreneur

PokerDIY Tournament Manager - PokerDIY Tournament Manager<
PokerDIY Game Finder - Mobile Apps powered by DNN
PokerDIY - Connecting Poker Players