Thursday, 28 March 2013

How did you come to be using Railo? (survey results)

I've got 80-odd results for the survey I started the other day. This was one of the best and fastest responses I've had on the various surveys I've run on this blog. I wonder if this is because it was just a very short survey (one question), it was something people felt like responding to, or just that more people read this blog these days? Dunno. Anyway: no-one cares about that.

So: the results. Firstly I'll repeat something I touch on with each of these surveys I do, and that Mark Drew alluded to in a comment against the original article relating to the survey. With the mechanisms available to me, these surveys are not reflective of a very broad population base and this can have a hand in pre-determining the results.

I wrote most of this on my phone whilst eating lunch, and I've found an awful lot of auto-"correct" weirdness whilst reviewing what I've typed. I've fixed what I found, but no doubt have missed some stuff. Apologies for sounding even less literate than I usually do.

For this survey, I am reaching-out to the following "audiences":
  • people who follow this blog;
  • or my mutterings on Twitter;
  • or one of the following community forums:
    • the Railo Google group;
    • the Adobe ColdFusion forum;
    • the OpenBD Google group (with fairly unpleasant results);
    • the CF-TALK mailing list;
  • or finally the followers of a few people who re-tweeted my initial Twitter message.

The common thread here is that they're all social-media organs. The problem with this is that it's reasonably well known that a lot of CFML users simply don't participate in the community. So the survey had no chance of reaching them, and is only really polling community-active people. So what they have to say is perhaps no representative of the entire CFML user base. But this can't be helped.

Mark also had a concern that given the places I was polling, it would sway the results as everyone who would see the survey will have intrinsically arrived at Railo from some other platform. I think that logic is faulty, as I'm asking directly on the main Railo community forum (the Google group), as well as the #railo Twitter tag (which, admittedly, sees bugger-all traffic). So people paying attention to those won't intrinsically have had earlier non-Railo exposure to CFML. However I do think most will have had previous CFML exposure because I believe that describes almost all Railo users.

So that's it for the caveats. Make of these figures what you will:

Via what route did you come to be a Railo developer?

(out of 86 total)
From already being an Adobe ColdFusion developer either now or in the past 77 89.5
From already being a BlueDragon developer either now or in the past (but not having used Adobe ColdFusion) 0 0.0
From never having used CFML at all (ie: you have never previously used ColdFusion or BlueDragon) 3 3.5
Other (please specify) 6 7.0

Note: five out of six of the "other" responses should have been simply "From already being an Adobe ColdFusion Developer...", but the person was giving a more thorough career path. The other one was invalid, as it was "never used Railo". As this survey was specifically for people use use Railo, that's not a valid response.

So the final figures are:
From ColdFusion: 96% (82/85)
Direct to Railo: 4% (3/85)

If I'm honest, that's actually more people adopting CFML via Railo than I expected. But resoundingly, the bulk of people using Railo have arrived from ColdFusion.

Mark offered up some of his own figures for this. When downloading Railo there's an optional form which - amongst other questions - asks whether the person had used CFML before. Of the people answering the question, 30-40% say they're new to CFML, ie: they're not coming from ColdFusion or OpenBD, but Railo is their first port-of-call. This is great news for Railo, I think: that's a really high percentage.

But whether looking at my figures or Mark's, the majority of people are migrating from ColdFusion.

The reason I commenced this line of enquiry is that there was a discussion as to what the best focus fur Railo's official documentation efforts should be, and the Railo position  seems to be to treat Railo as if it exists in a vacuum, which will lead to it re-treading already-documented ground of working through all tags and functions etc and documenting them. I get why Railo wants their own documentation (and I have lobbied for this in the past) and not rely on people having to refer to Adobe's docs for ColdFusion. However from the community side of the scene, all that stuff is done, so having it done twice by two different vendors is actually not of any benefit to anyone: I just need to know how <cfabort> works; if it works the same in Railo as it does in ColdFusion, I don't need two sets of docs telling me the same thing in two different ways.

However where I am perpetually looking for documentation and not finding any is when I come up against unexpected behavioural differences between the two platforms in already-existing code; or secondly simply finding out what things Railo offers that is different from ColdFusion's CFML. The stuff that differentiates Railo from ColdFusion.

And my suspicion is that my requirements here are representative of most Railo users.

Given Railo's budget for docs is not a bottomless pit, I think they should focus their initial efforts on filling gaps that currently don't have coverage, rather than repeating what's already been done by someone else. Surely one of the benefits of copying someone else's language is that one gets to leverage their docs too? This is not one of those places where I'd be coy about the whole "copying someone else" thing. But that's just me.

What do you think?