Saturday 31 January 2015

Lucee: what to tell your boss about this whole RailoLucee thing

G'day:
I'm sure the Lucee bods will come out with their own whitepaper on this shortly, but here are my thoughts / analsyes.


Railo

Railo has basically discontinued. I am 95% confident there will never be another release of Railo. Any bug fixes will only ever be done by the residual community. However most of the community itself never really participated in committing.

Railo is no longer a viable option to continue with, in a commerical context. So this pretty much makes the decision for you.

Yes I know "nothing has happened to Railo, it's not as if it's gone away", but this is a bit specious. Railo is, indeed, still there as much as it ever was. But that's it. That's the way it will remain from now on.

Note: I am keeping an eye out for something from Railo in response to the recent Lucee news, but I've seen none. If I see any, I'll share it. If any of what I say above is not accurate, I'll update it, and let my readership know.

Community

I am fairly confident the entire Railo community upped-sticks and moved to the Lucee community at the end of last week. If you persist with Railo, you're gonna be in a community with roughly as much input as the OpenBD one has. That is to say it's mostly tumbleweed, crickets and the odd owl hooting.

I am still subscribed to the Railo Google Group, but once the Lucee questions die down, I will unsub. I will still watch for Railo questions on Stack Overflow, though.

Code migration

Lucee'a version is 4.5, which obviously is intended to have some relation to the last Railo version: 4.2. It's built on Railo 4.2, with a coupla tweaks. I guess one should consider it more the equivalent of Railo 4.2.3? So the migration from Railo to Lucee should be undertaken with the same care as upgrading from Railo 4.2.2 to Railo 4.2.3.

Seriously: it's the same code base, it's the same team working on it. The name's changed, but it's mostly the same beyond that.

I think the migration would be very low-risk.

From my perspective - I only ran Railo Express, and I have nothing in production - it was simply a matter of replacing railo.jar with lucee.jar, and I was running Lucee. I have not checked thoroughly, but I have seen no issues at all yet. Nor would I expect any, given Lucee's lineage.

Risk

This is a legit consideration. The "death of Railo" (let's face it, OK?) took everyone by surprise, and it does reveal internal instability in that project that's only seeing light of day now. However Lucee's raison d'etre is to deal with these issues. I also think they have learned from their lessons, and the whole approach of having the Lucee software application administered by Lucee Association Switzerland is actually a good move, and a brilliant message that the Lucee Team realised things weren't working with Railo, and Lucee's approach is the solution to that.

I also know that there's great backing for the association and the Lucee project itself from the likes of Pixl8 an Ortus, et al, and they definitely have the right spirit in their approach.

Reality

It's business as usual. Micha and Igal are writing the code. The same community people are helping out in the same way, and doing it all cheerfully. Mugs like me are still around helping and hinering in equal measures. Status quo.

Bottom line

Yeah, this was definitely a bump in the road, let's accept that. But I think it's one that was just necessary to ensure our open source CFML is stewarded in a way that has the interests of the product and community first and foremost in its mission statement.

And why should your boss care what I say?

Ah... they shouldn't. But in whhat I say above, I don't think I'm misrepresenting reality, so they should read what I say, listen to you, and draw their own conclusions. The conclusion should be they're good to go with Lucee.

--
Adam