There’s a rumor flying around that Twitter is going to abandon Rails, so obviously this calls for us to stop and ruminate without making even the barest efforts to confirm or deny said rumor. Facts may be added, like a garnish, to later blog posts, like Parmesan cheese to your spaghetti.
For example, Evan Williams says that Twitter has no plans to abandon RoR. Hm. I guess that ought to just stop this particular line of musing right out of the gate.
Never mind. Let’s talk about it anyways.
Twitter is famous as much for its downtime, bugs, and technical difficulties, as it is for exploding into the web’s social superstar. The “Twitter will abandon Rails” line of thought would place the blame for these difficulties on Rails. I’m not sure that’s a fair placement of blame, but hey — this is the internet. We can speculate about anything, just like we can speculate about things that company’s founders say are not going to happen.
The most important question that I can think of in relation to this non-news story is so what?
For the sake of argument, let’s just say that Twitter was going to leave Rails. Would this somehow be damaging to Rails reputation? To some, maybe, especially it were implied that the reason for the switch was to eliminate the technical problems.
To many if not most others, I don’t think it would matter at all.
Rails is clearly working just fine for applications such as 37Signals various offerings, or apps like Shopify or Cullect. As long as it continues to do so, and to make developers lives easier and more enjoyable in the process, I don’t see Rails disappearing off the map any time soon.
What seems more likely is that Rails development may split into different types: Something like, Vanilla Ruby on Rails versus Enterprise Ruby on Rails. A sort of Rails and Rails2EE, if you will. In some ways, this is already happening, with products like JRuby running rails in WAR files in J2EE application servers.
Is this something that the average web developer wants to do? No. Last time I checked, most hosts don’t offer J2EE servers as a feature; probably because most average web developers neither want or need them.
(I realize that the sort of servers and hosting companies which scale much larger offer root access, so you could install J2EE servers to your heart’s content. But whatever; even thought they do, I don’t see many developers raving about how excited they are to begin doing that on their new server, in the same way as I see them getting excited about Rails, or mongrel, or mod_rails, or what-have-you.)
In my humble estimation, even if Twitter were to abandon Rails, this would not matter. In any case, it seems that Twitter will not be abandoning Rails, which makes the whole discussion moot, perhaps, but still interesting.
Your 2 cents?
Since the guys from 37Signals were the ones to develop RoR, I wouldn’t be surprised if the version they use is slightly (if not quite) different from the “public” version, so I don’t think they’re a good comparison.
That’s a good point, but…
My understanding is that 37Signals is currently running most apps off Rails Edge, which anyone else could do also; see David’s tweet about this.
I think it would defeat the whole purpose of some of the benefits of using Rails if they were using a “different” version. Now, their servers and perhaps many plugins may well be finely tuned and customized, but their version of Rails itself is certainly the same as you or I could use.
Obviously I can’t “know” that, but that is what David said, and I can’t imagine any possible benefit to changing only their version, and not Rails Edge (the development version) itself.
100% agreed on all fronts. A lot of people who use Twitter are the programmers or the server admins. In other words, they’re the people who _would_ care about Twitter abandoning Ruby on Rails. The average Twitter user doesn’t care if the site runs on Joe’s Programming Language of the Week as long as it works. People aren’t going to leave Twitter just because they switch platforms. If McDonalds starts using one brand of fryers over another, do people care? No. As long as the fries still taste the same, it doesn’t matter.
We certainly don’t run a special version of Rails at 37signals. That wouldn’t make any sense. The whole point of open-sourcing Rails was to participate in the shared commons. Running our own fork isn’t very compatible or sensible for that.
@dhh, exactly. It would seem to me to defeat the whole purpose if you were using a “custom” version. Especially all the handy features that come from rails being distributed as a gem, and being able to automatically update it, etc, etc. Running a unique fork would invalidate a lot of benefits.
Thanks for stopping by!