I have begun to be able to articulate something I have been wrestling with for a long time around Ruby on Rails, and I think it potentially applies in all kinds of places.
Rails is a fantastic tool for developers, almost everyone accepts that. It's true. I just realized though, that's exactly the issue. When we were working on p0pulst for example, RoR saved us some time and was pretty to look at and easy to code. A lot of Rails structural stuff, migrations for example, were really not all that helpful. When more than one developer is involved, you spend a lot more time screwing with migrations than you would have spent just changing the
damn table in the first place. I know the counter argument begins with, "But you should" and I think that's part of the problem. Same goes for generators, blah blah, the specifics don't really matter that much. For sure though, Rails was more fun to work with overall than say, PHP which is usually kind of a pain, looks ugly, etc. It was so easy to work with in fact, that I felt a little lulled by it. I have to wonder out loud, and I'm thinking about Perl vs Python here say, does the presence of a right way to do things stifle or encourage innovation? Is easier and more "maintainable" always (or even often) going to translate to "better for the user?"
Here's the thing: a great tool should really go unnoticed. Unnoticed by whom, you ask? Ahhh, but you already know the answer to that, my little butter bean, a great tool should go unnoticed by the PEOPLE who use the application. It turns out that applications are in fact mostly for the people who use them, not the people who write and maintain them. It's a one-to-many thing. So a cool tool, a weird data structure, a deployment strategy named after a place where swallows frequent, etc, just doesn't mean shit, man. Apps succeed and fail based on how much people like them, how appropriate they are for the culture and the time, how many people get into them and stay into them. Myspace was written in Cold frickin Fusion! Imagine anyone writing a real web app in Cold Fusion! And of course Myspace was a nightmare by any developers estimation. But real people out in the world literally didn't notice or care at all except that Myspace was always kind of slow and that was enough of a bummer that they headed to Facebook even though they couldn't keep their totally bitchin' purple skull page background. I think Rails is still so god-awfully slow and silly to deploy for the same reason, but on the other side of the planet. None of us thought in 2004 that Rails would still be a production liability 2008.
The conventional wisdom is something like, "well, no one wants to hand code add/update/delete ever again," but context to context, I wonder where and what the joy really is and where innovation and new ideas really sprout. The reasoning is supposed to be that a tool which inspires more joy for developers will lead to a better application. Where is the Awesome Dial set on todays web applications compared to 1999? Are we still talking about the joy of problem solving? The joy of that perfect 4-table query? The joy of an optimized data tier? The joy of an app that hasn't restarted in 600 days? All I mean to point out is that there are certainly flavors of joy. Even Cold Fusion must inspire or have inspired joy in SOMEONE, right? Is REST-ful also joyful? Is it pro-user? Is Rails? Was writing ASP in VBScript?
I read today where Engine Yard took another 15 million in funding, and I can't help but wonder about these kinds of things. Is Engine Yard maybe a kind of HMO for Rails apps? Here's hoping that they work on MERB with some of that money. We had to support Smirk with PHP and I have to say it was fantastic to write some SQL.
