| By Yakov Fain |
Article Rating: |
|
| April 9, 2009 09:30 AM EDT |
Reads: |
1,480 |
The Register has published an article about Twitter considering moving the mission critical portion of its code from Ruby on Rails to Scala.
This sounded logical to me, even though the article has been dated by April first. But then Artima has confirmed the news a couple of days later.
As usual, the bad part of me goes, "Told, ya!" as people who read my posts knows that I was never a fan of RoR. I really respect DHH and his followers for doing exceptional job in exciting so many software developers with Ruby and RoR. But when these boring men with cash (BMW$) come into play, things change. The problem is that BMW$ don't get excited that easily. Their low-IQ brains go like this, "I may invest $X into this startup, and on exit in five years (IPO or takeover), it should bring me $Y". They always had problems with math, and the only formula they've managed to learn was this:
if ($X >$Y){
// Run away
}
These filthy rich BMW$ don't give a damn what's the name of the programming language is that the startup is using for development of the software that will allow them turn $X into $Y.
Ruby makes developers happy, but if you are entering the business world, you'd better use a language that makes BMW$ happy.
Twitter deserves credits for being extremely creative - they hired professional artists who came up with this picture of a little whale laying at the bottom of the ocean shown several times a day when Twitter's servers go over capacity. They've also created a bunch of other cartoons that were supposed to act as prozac on people learning that Twitter is once again on unscheduled maintenance.
RoR is great when you need quickly create prototypes of a Web CRUD application. It may be a good fit when it comes to developing Web sites for small businesses. But when it's time to architect enterprise-grade mission-critical applications, I'd stick to tried and true Java servers. If Scala can run on JVM that's fine as long as some serious performance benchmarks can prove that it's as fast and reliable as Java EE.
Now, I'm wondering how long does it take to Twitter to realize that it's time to find a decent stateful replacement for their AJAX client, which will allow further lower the load on their servers? Shouldn't take more than a couple of years.
If my Twitter account won't be banned, it'll remain the same twitter.com/yfain
