I couldn’t understand some parts of this article, but it sounds interesting…

As for the ‘beastly JVM’, well JRuby is moving along quite nicely and almost has complete support for Ruby on Rails applications. Not to mention that Graeme has already shown that early testing of Grails outperforms Ruby on Rails.

I’m aware of eapps BTW. I just looked and they have some good stuff since I last saw them.

Check out http://eapps.com for both Java and Rails done right and with GENIUS support. That way you can still host the last of your Java apps whilst you learn RoR

I think initially smaller companies will lean towards RoR while the bigger ones go towards Grails if for no other reason that Grails is Java based and thus there is a wider developer base (for the time being) which is a consideration when it comes to staffing. With regards to the bigger companies, it will be the developers themselves going towards Grails more than a ‘company’ decision. I’ve certainly been one of the developers before that started using a new framework for an application for whatever reason (RAD, curiosity, whatever) and the company didn’t care as long as it ran on their chosen platform, i.e. a JavaEE server and it was documented. Smaller companies might be more inclined to have specific reasons for one over the other, especially if that company is basing their product on the software being written (again I point to 37signals).

There is enough buzz and interest in Rails however that I think the performance side of it will increase at a faster pace than Java’s did back in the day.

Re: “As for the performance I intentionally left that out because I thought I would get more backlash about it still being new, Java having a 10+ year head start, etc. etc.”

I would like to note that this is wrong. Technically, Ruby is actually only a few years younger than Java’s initial revs (and about the same in terms of publicly-released years), and certainly far older than Groovy; and that Rails is also certainly older than Grails.

I will admit though, that whereas the JVM has had popular exposure since a few years after it was publicly released, Ruby appears to only have had a popular presence since the appearance of Rails (and only recently received a VM). Furthermore, Hibernate has had more time than ActiveRecord to mature.

But companies don’t choose a development platform on the hopes that performance will improve, but that the performance is sufficient and that the development effort is reasonable. It seems to me that Rails may be a more mature framework and will be appealing to shops where performance is not a top priority, but that Grails has the performance advantage, and can work with pre-existing Java web development efforts, and is built on top of very mature pre-existing Java frameworks, and therefore is more appropriate/comfortable for development when performance is more important, but development speed cannot be sacrificed.

I’d personally like to see the RoR developers respond to Graeme’s performance comparison. If that means that they run the test using JRuby, a better CGI, or some newer version of a RubyVM, so be it. At this point, I believe that the ball is in their court to prove that they can match Grails performance.

If by “Grails is going to be for serious development and serious projects” you mean the ‘enterprise’ market I agree UNLESS JRuby really takes off and supports RoR really well. To call the current RoR applications out there not serious (such as the apps developed by 37Signals) isn’t fair though.

autocrat: Fair enough! I did see the roadmap about the middlegen integration and I honestly forgot to mention it, so you are correct there.

As for the performance I intentionally left that out because I thought I would get more backlash about it still being new, Java having a 10+ year head start, etc. etc. There was a time when Java wasn’t all that speedy and it took a few years before the JVMs came along that made it really shine. So to be fair Grails does have the advantage because it runs on those VMs and I’m sure Mongrel, etc. will catch up. Then again, perhaps we could be running Ruby on the VM and then we would have some real benchmarks!

Good debate gentlemen, thanks for the comments!

Yes, RoR might well be the next PHP. Fine..take the market. Grails is going to be for serious development and serious projects.