Dave Laribee

So, the choice basically boils down to:  Roadie or Groupie? :)

Not sure (really just Not sure) if I agree... I see MS MVC as a pluggable solution where you can switch out the part which doesn't suit you... DI, Views... so if Castle produces something better (like a better DI container)... maybe it will be pluggable in MS's solution (at least it sounded like that in Scott's video)?

I like the Castle stack but I sincerely think we need to put a 1.0 stamp on that thing asap. The beta/RC thing is one more contributing factor to the smaller adoption. There will always be bugs. Heck, look at RoR, they had 2 maintenance releases in the last few days.

Roadie: Doing the hard work, getting stuff done, etc

-vs-

Groupie:  Hanging out with the rockstar programmers

It's difficult to imagine a complaint that Microsoft moves too slowly, from those of us in the trenches who only just this last year got to start working with Visual Studio 2005 and still have yet to uncover all of it's capabilities.

Dave,

I'm a former .NET developer. I've been working with Ruby for about 3 years now (I don't keep track of specific dates :-) ), and Rails professionally for about a year.

If you really want to be where the action is, skip Rails.

Merb is more than twice as fast. Much easier to follow code (my #1 reason), with a much friendlier, open-to-feedback community. The Framework itself is more "OO". Actions render whatever is returned from them. No "render :text => 'Hello World!'", instead it's just "return 'Hello World!'" (call #render if you want to render a template instead). Build re-usable Components with Merb. Wrap up your Controllers, Views, and Models for a specific piece of functionality just by dropping them into a Gem. No voodoo required. With a plugin a few of us wrote, pass arguments to actions for parameters similar to Monorail (requires ParseTree gem, *nix only).

DataMapper (my own project). Peek at http://datamapper.org/why.html for just some of the reasons why. Oh, and it supports multiple databases with different mappings out of the box.

DataObject.rb is a port of the best part of the ADO.NET interfaces (Connection/Command/Reader) to provide a consistent, fast way to access (so far) Sqlite3, Mysql and PostgreSQL: dataobjects.devjavu.com. Soon I'll switch the DataMapper to use it as it's basis. Within a week one of the most prolific (involved in DataMapper, DO.rb, Rubinius, Merb and a core developer for jQuery) guys in the Ruby OSS community, Yehuda Katz, wrote this in his spare time.

If you have questions or need help with any of this, feel free to stop by #merb or #datamapper on irc.freenode.net.

True, some of these projects are young, and are going through pretty heavy development right now, but if you want to be on the edge, this is where it's at. If you want the Ruby skills that will carry you into any project, Nitro, Merb, Rails, etc, with ease, this is where you'll develop them.

It's also difficult to imagine a that Microsoft moves too slowly, from those of us who market a competing product and can see years of hard work duplicated by Microsoft is a matter of weeks/months

Careful when searching for Merb. It also stands for "Montreal Escort Review Board".

Guys,

In many ways MS does and will continue to lag a community:

NUnit/MbUnit/xUnit.Net not to mention the NSpec type extensions will continue to move faster than MSTest

JQuery/Prototype/Scriptaculous move faster than whatever Atlas is called now

and

Rails is marching on.

@Sam,

Merb sounds cool.  Is it a little more suitable for more complex domains?  I'd be a little spooked to use ActiveRecord on some of the databases I've worked with this past couple years.

@Jeremy: You don't _need_ to use ActiveRecord to use Rails.  You can rBatis or DataObjects.rb or Sequel or ActiveResource or your home grown DBI models if you want.  Just like you can swap out the views the ASP.NET MVC framework, you can swap out the model and view implementations in Rails.

That said, MERB does look very cool.  Its been a while since I looked closely at it and I plan on remedying that.  :)

The notion of "rockstar programmer" is just ridiculous.

@Jeremy Couldn't agree with you more. The fact that castle/nhibernate/nunit will almost always move quicker than microsoft is what really drives me to use those tools over MS tools. I take a look at MS tools but unless there is realy value in using a MS tool I just find the open source equivalent

Wouldn't the Cathedral and the Bazaar be a better analog? Microsoft is only releasing their cutting edge stuff when they think it's done, which is the cathedral approach and projects like MonoRail release whenever a piece is done, the bazaar way.

Also, my rockstar programmer idols are Scott Guthrie (In a cathedral) and Miquel de Icaza (In a bazaar).

I guess I'm imagining who you mean by "rockstar" and that has a big impact on my response...

I think your often better to stay a little behind the bleeding edge. With any "advance" we get a lot of excitement in the blogosphere, the "rockstars" move in and start arguing then a couple of months later they have moved on. I don't tend to get excited about things till I see how they can be useful to my work.

Quite often I find lagging a few months behind the cutting edge, but keeping up with the discussion about the cutting edge, means you don't end up wasting time on some new thing just because one of the "rockstars" has gotten all excited about it.

So as an example I'm keeping an eye on things like FIT/BDD and so on but haven't rushed in to use them yet because there is still a lot of debate and the tools are still evolving or (seem to) have some serious issues.