Sunday 25 September 2016

CFML: "It's alive!". Really now?

G'day:
Yesterday I was "helping" someone on the CFML Slack channel with some challenges they had with CFC serialisation. The conversation turned to ColdFusion's custom serialisers (see an earlier article: "ColdFusion 11: custom serialisers. More questions than answers"), and how pants I think the implementation & entire undertaking of that part of CF11 was. I'll just quote myself (full transcript of entire discussion here: Gist / adamcameron / slack.txt:

I did try to position the whole undertaking as "here be dragons"

On the face of it, it's a cool idea

But - as per most less-than-obvious things the Adobe CF Team tries to do - they completely cocked it up, having no real understanding of how ppl use CFML

[...]

What concerns me about Adobe's implementation is it seems to be very "not OO" which worries me when it's actually implemented by ppl who are supposedly Java devs.

[...]

I floated it "at the appropriate time" (nod to not speaking about the pre-release programme, but... well... f*** them), but the relevant parties just either didn't care or didn't get it or refused to get it.

It's the perpetual symptom that Adobe never consult the community before they blindly go and implement ill-conceived and poorly realised shite.

They ought to feel free to only consult with their major-client "IT stakeholders" for features like the API Manager and to a degree `<cfclient>` (hohoho), but when it comes to nuts 'n' bolts language implementation... they need to ask the ppl on the ground.

But based on that recent survey Adobe did... there won't be any more focus on the CFML language from them. They're only interested in buzzword-completion/box-ticking IT-Manager-pleasing platform features

In response to that someone observed they admired my sticking power with CFML (thanks for saying that), but my reaction was that Adobe finally "broke" me with CF2016 being such a waste of time for the CFML dev, as they did hardly anything in the language, instead of rolling something that's really nothing to do with ColdFusion: the API Manager into CF2016 and making it the marquee "feature".

At the same time Lucee 5 lost objective direction in that they indulged Micha's whim to re-write the thing to be OSGi capable (or whatever the correct term is), and mess around with the .lucee lang thing which ended up being both stillborn and evidence that LAS is out of their depth when they're not just copying what Adobe does. Harsh, sorry LAS, but that's my opinion.

I have given up on both the vendors when it comes to CFML, and personally I've moved from CFML to PHP now, but I'm still there in the CFML community helping people where I can. So - as far as the community goes - I'm suffering from "just when I thought I was out... they pull me back in", as it were. Apologies to Al Pacino there. But theatrics aside, that's my choice, so no worries.

Patrick came back with a reasonable riposte to my comments about being let down by the vendors:

Yeah, those sorts of public "divorces" (Railo-->Lucee) are never pretty, but it was really necessary. Kudos to the core of the Railo peeps who held it together. LAS is quite strong now, and I believe we're the future of CFML as a platform. I'm especially excited to see what the community can do with the extensible framework in place now. What say we all co-create a really vibrant ecosystem of products? With Adam as our cheerleader (I don't see it as cynical, mate!), I have no doubt we'll get it done.

Good on him, but I disagreed with one thing he said, and this lead to a bit of a diatribe from me:

You can't just say "What say we all co-create a really vibrant ecosystem of products?" and have it happen. The CFML community have demonstrated that they're - for the most part - takers rather than givers. Even with ppl asking for help from the community I see a sense of (misplaced) entitlement from them.

That said there's a minority of good worker bees, and a bunch of people who are good people but they're still just in CFML 9-5 (which is fine), but will not grow or strengthen the community.

Most of the stalwarts of the community are companies who have in the past invested a lot of effort into CFML (not the community necessarily, but for themselves), and it'd be a large $$$ cost for them to move. That's going to be why some of them continue to persist with CFML, and also encourage Lucee[:] as LAS is more likely to work with those small to mid-level companies to provide what they want out of CFML. Same as with Adobe and larger corporates.

The likes of Daemon and MSO.Net and Ortus and Pixl8 aren't going away when it comes to CFML due to how they've stacked their egg baskets, and I reckon they're more likely to stick with CFML than [it is that] LAS will [continue to] exist. Although one of them might consider "buying" Micha to perpetuate CFML for them if LAS founders. At which point Lucee will become kinda like OpenBD is for Alan Williamson's outfit

Sean was reading this and "emoting" with thumbs-ups in agreement and the like, then we got to chatting about OSS & CFML and that sort of thing. I'll spare you that.

But I stand by what I say above. I've been in the CFML community since 2001, and been active in the community all that time, and have been watching the ebb and flow of the CFML tide, be it ColdFusion or BlueDragon or Railo or Lucee. So I think my opinion is possibly an informed one. And I always offer it for the sake of informing, rather than posturing. People are quite welcome to disagree with me and to put a case forward as to why I'm mistaken. I like finding out I'm wrong, and I like debating (or arguing ;-) a topic with people. Provided they actually make a case.

Geoff Bowers (of afore-mentioned Daemon fame: he's the boss; and I think he's El Presidente of LAS) joined the conversation and suggested we should not be having such conversations, as they hurt the CFML community. And Geoff's position is that "The CFML Community" is a meaningful entity.

I disagree.

I disagree with a few points or inferences here:

  • Any conversation by experienced members of a given community will possibly have merit, even if they suggest that the best course of action for the community members is to prepare themselves for the time when "the community" doesn't really mean anything any more.
  • Even if the points being discussed have no merit, I disagree with the notion that the discussion should not be had because one member of the community doesn't like the sound of it, or it doesn't match their agenda.
  • I don't actually think "The CFML Community" is a thing in and of itself which automatically deserves preservation. This might seem odd, but I'm interested in the welfare of the members of the community, not some completely intangible label "The CFML Community". That's meaningless. I will help the individual people in the CFML community whenever I can (or, realistically, when I feel like it, but that's quite often ;-), and I think helpful advice to people who rely on CFML for their crust these days is: plan an exit strategy. It's convenient for Geoff to deny this (I'll get to that), but CFML as a language has had its day. It had a unique selling point a decade ago: it made making dynamic websites quite easy compared to where Perl or PHP or JSPs were at the time. It does not have that USP any more. CFML does a bunch of things really well, but not that much better than any number of more popular (and easy!) languages do these days. The thing is the other languages bring their own participatory communities with them, and with those communities come rich OSS efforts which ensure whatever one wants to do that might be off-piste for a language: there'll be a third-party module for that.


But I digress: no, I don't think "The CFML Community" is worth protecting. The community members are, and my best (-quality and -intentioned) advice is to start planning an exit strategy. You'd have to be an arsehole to place the preservation of "The CFML Community" ahead of the welfare of its members.

CFML has precisely one remaining relevant purpose: to keep CFML devs in work if they can't do anything else. CFML is not going places. It's staying where it is. But the ice floe where it's positioned itself is melting away, growing smaller and smaller, and nothing's gonna reverse this.

Geoff disagrees. That's cool, but other than "I disagree" and a few half-hearted ad hominems ("none taken", btw), he didn't actually refute a single thing I said. I'd be dead f***in' keen to hear what he has to say though, to demonstrate how CFML is still alive in any meaningful way though. And how it's a good prospect for uptake, and how CFMLers shouldn't be planning an exit strategy.

My cynicism suggests to me that Geoff is mostly concerned about "The CFML Community" because it's commercially prudent for him to do so: he's built a company and a livelihood on CFML. So it's beneficial to him for there to be "The CFML Community" out there, as they're potential revenue opportunities. That's completely fine and I see where he's coming from. But it seems disingenuous to me for him to be suggesting myself or others don't have the community's best interests at heart when we offer advice that is not convenient for his business model.

Another issue I'll touch on in this increasingly rambling diatribe is Geoff's assertion that my observations have an actual deleterious impact on the community. This is seemingly based on the notion that any negative comment is intrinsically deleterious, and for "names" such as myself to me making these negative comments is more deleterious still. I guess cos ppl might be inclined to listen to my advice. Heaven f***in' forbid.

He's probably right. But let's get things in perspective here. Things that are deleterious to "The CFML Community":

  1. ColdFusion being a paid-for product in 2016;
  2. Adobe's lack of interest or lack of ability to keep the language contemporary;
  3. every other bloody language being more contemporary and compelling;
  4. and free.
  5. The apathy of most CFML dev resulting in an under-represented community to engender momentum; 
  6. Adobe's lack of ability to market ColdFusion;
  7. Railo v Lucee schism;
  8. Lucee 5 taking an inordinate amount of time to see light of day, for what seems like developer-self-indulgence;
  9. LAS dithering about the place and seeming very incoherent when it comes to what they're doing;
  10. [daylight]
  11. [more frickin' daylight]
  12. Me saying stuff Geoff doesn't like.
One could quibble about some of those items there, but it's the scale of things that is significant. Almost all of CFML users don't know who the f*** I am. A decent subset of those that do can think for themselves and will take what I say with a grain of salt, or already be (dis-)agreeing with me. And the others could do worse than consider what I have to say, and weigh it up for themselves.

But no CFMLer who cannot form their own opinions will have any idea who I am, and will never see any of my written opinions.

I hasten to add that the list above is "things that are deleterious to the CFML community". I see Lucee as a net force for good in the community. But if we're assessing what is unhelpful in the community, then they're a more significant player in all this than I am. And even Lucee's influence one way or the other is completely inconsequential compared to Adobe's.

But go on then, Geoff and others who think CFML is alive and well and is something that should be suggested for new adopters. Tell me why it is. Tell us... "The CFML Community"... why that's the case. Don't just call me an uppity dick or a big meany or whatever the f*** personal attack is easier for you to make. Do something.

Why is CFML still relevant?
Why should new people adopt it?
Why should people stick with it instead of perhaps protecting their future by examining other options?
Why should people not at least suggest that might be something to do?

Go on then...

--
Adam