Slashdot Mirror
Sun's JRuby Team Jumps Ship To Engine Yard
itwbennett writes "'To be honest, we had no evidence that Oracle wouldn't support JRuby, but we also didn't have any evidence that they would,' said Charles Nutter, explaining why Sun's entire 3-member JRuby team will be leaving the company to work for application hosting company Engine Yard. Nutter called getting hired by Sun about two-and-a-half years ago and being given the chance to work full time on JRuby a 'dream come true.' And said that the decision to leave Sun came down to making sure 'JRuby will get to the next level.'"
I think it was more something like "Engine Yard wants to hire all three of us. Oracle won't give us a yes or a no on whether we'll even be here next week, and the magic eight ball says 'outlook not so good'. Let's take the offer we have in hand where someone in management will at least know we exist."
I worked for a Fortune 100 that took two years to getting around to closing our office--that's how long it took them to notice we were there and bother to send out layoff notices. Between continued employment at a place like that and a place where someone actually wants to employ me, I'll take the latter.
Anyone who loves or hates any language, platform, or manufacturer, doesn't know what they're talking about.
Python is OK because it's mature and you can build good infrastructure such as Zope with it, but talking about organisations using RoR being out of business in favour of PHP and Perl?! ROTFL.
Oh naive curr. This is not your fathers Internet where running code and rough consensus meant anything. I'm sure with the proper publicity photos, you tube videos and social media consultancy this project can be hi profile, media centric and the darling of those who tweet.
Whats usability or performance go to with anything today? Hell, this is one of the mild examples.
I'm surprised ICANN hasn't already contacted them for their new language registry yet. Better get certified quick before the price goes up.
Need Mercedes parts ?
Quite right. In a few months' time they could all be out of a job with no prospects when Oracle has an inevitable cull and eventually realises they don't want them. However, they have the opportunity to jump ship to an expanding company now who will directly rely on and need the technology they are developing. They would be stupid not to jump ship to that in the current climate.
It's the ruby VM that's slow. There's nothing about the language itself that makes it slower than any other dynamic languages.
> If there's one thing even slower or more pointless than jvm and ruby,
> it's ruby on rails. I guess someone didn't get the memo,
> but RoR's 15 minutes are up.
To the contrary, people are cranking out new Rails apps at a furious rate, and lots of Java and C# apps are getting ported over to Rails. It's good times.
The Army reading list
...is that now EngineYard has full-time folks working on Rubinious, JRuby, and Ruby 1.8.6. It's Ruby implementation central over there.
The Army reading list
I'll one up that: Based on my experience in several software engineering companies (video games), my policy is now to find a new job ASAP when your company gets bought out, even if they DO say they want to keep you! (That being standard operating procedure -- not hearing that is the equivalent of an immediate pink slip.)
Engineers can only really advance by switching companies. Definitely best to jump ship when some people are talking about you and you have some leverage, no doubt about it.
We know where leadership by an anti-intellectual "strongman" who scapegoats minorities and likes boisterous rallies goes
... the ramblings of a guy named Nutter.
Since Malda likes em young, I'd guess he is a chronological 11-13 year-old.
Do you even lift?
These aren't the 'roids you're looking for.
But they would be even stupider not to ask for a counter-offer from Oracle before formally jumping ship.
Except they're in the nebulous legal gray zone where the Sun Board has approved the deal, but DOJ & the SEC haven't given their blessing yet. The companies continue to make decisions and operate as separate entities, and people can actually get in legal trouble for attempting to do otherwise.
So... They couldn't go to Oracle and ask, and Oracle couldn't have given them an answer if they did.
Benchmarks are actual observations -- what we in the real world like to refer to as "science".
If you've got a benchmark of a larger application that you'd like to share that shows just how much the JVM sucks, I'm definitely curious. Ruby 1.9 works well for me now, but I'd been considering JRuby as a performance boost. If you can show me it's actually a failure, that would save some time.
Don't thank God, thank a doctor!
In the aftermath of a buyout, three employees trying to jack their soon-to-be-former employer for a raise is a recipe for negative attention.
The thing about the aftermath of a buyout is that the purchaser (legitimately) takes months or years to understand what they've bought, and decide what to keep and what to prune. Unless Oracle bought Sun to get Jruby, then calling attention to yourself by seeking a counteroffer is a good way to move to the top of the "keep or cut?" list, and in a way that makes cutting all the more likely.
And generally speaking, "large, deep-pockets organization(s)" are no more stable than a startup, from the grunt's perspective. At any moment, you're one spreadsheet away from being laid off to improve the quarterly statements.
Anyone who loves or hates any language, platform, or manufacturer, doesn't know what they're talking about.
And we're seeing a lot of speed improvements coming down the pike with Ruby 1.9 and Rails itself should prove to be significantly leaner and more modular with Rails 3.
Uhhh, yeah. And pretty good, thanks. I guess it must have had something to do with my recognizing my own value to the organization and approaching management honestly and openly, instead of skulking around like a venal little creep. But by the same token, if they didn't want to pay what I was worth (to a viable competitor) then there was little reason for me to continue working there.
Breakfast served all day!
Not sure about PHP, but when you make such comparisons between RoR and Perl solutions you should at least be aware of recent developments such as the Catalyst framework.
Not to mention things such as perl5i which tries to aggregate most of what is known as modern Perl.
Perl is an evolving language and Perl code from 8 years is very different from modern Perl code.
I used to work with Charlie Nutter and Tom Enebo years ago, when we worked on the same Web team. And I was thrilled to hear when they moved to Sun, really was the best deal you could imagine. Note that JRuby wasn't actually bought by Sun, but remained a separate project, only the developers were paid by Sun to work on JRuby. So I wish them the best as they move to their new digs.
Good luck, guys!
So we take one of the slowest and most bloated virtual machines for static languages, the JVM,
Boy, do you have absolutely no idea what you're talking about! JVM one of the slowest VMs? You need to get your head out of the '90s. It's one of the fastest now. And JRuby is one of the fastest Ruby implementation. (Definitely faster than Ruby 1.8, which is dreadfully slow.)
You're correct that Ruby's dynamic typing doesn't go well with JVM's static typing, which is why Scala is much faster than JRuby. Still, the JVM is so impressive, and Ruby nice enough to work with, that JRuby is still a pretty good idea.
JRuby is a faliure for ME and for my project. This doesn't make it a failure for him and his project.
I looked it over, as a part of looking Java over. It's not a good fit. Doesn't mean it's not the right answer to some other problem. (OTOH, I'm not certain why JRuby is better than Groovy or Jython. My guess it that it would have lots to do with re-using existing code.)
If you're going to be using ANY of those, all time critical portions will need to be written in Java. No surprise there. It's like Ruby and C, only the interfacing is easier, and the result isn't quite as fast.
Personally I looked them all over and chose D (Digital Mars D). It's weakness is that it doesn't have many libraries. But it can interface to C libraries and, slightly, with C++ libraries. (OK, I'm using the beta version, D2.x.) It's not quite as fast as C, but it has built-in garbage collection, and it has an object model sort of similar to Python's. For my project that's the best fit I've found. Yours is likely to be different. Maybe you should chose Ruby, or Python, or C, or C++, or LISP, or Ada, or ...well, if you get much beyond that, I won't understand what features of the language make it desirable. I don't know Haskell, Clean, Mozart, OCaML, etc. Assembler...well, if that's the best language you're working closer to the metal than I would choose to work. But sometimes it's the right answer.
Don't diss a language just because it isn't right for what you're doing right now.
I think we've pushed this "anyone can grow up to be president" thing too far.