Here's some anecdotal evidence.

I spent a day recently trying to install Rails on my Air. I fiddled with it once back when Locomotive worked. I was expecting a nice clean drag-n-drop like that or MAMP. Instead I wasted half the day with broken Locomotive, then spent the rest of the day installing the dev tools (which space I can ill spare on my Air) and getting it running. I lost all desire to actually fuck wih it during that time. Now I want to nuke it because something's eating up all my free space and it started a little after I installed ROR. And I have no fucking idea where to start aside from just flushing the machine and restoring apps/data from backup.

So that's one piece of any data from a frustrated artist/occasional hobby coder who really never wants to touch the command line in her life. This is what it's like for a noob: scary, frustrating, and confusing.

So, uh... I'm a little confused here.

About a year ago I opened a terminal on my Mac and typed:

    $ gem install rails sqlite3-ruby

I waited a few minutes for it to finish, then made a new directory, created a new rails project, ran 'rails server', opened a browser, and it worked.

Has this changed dramatically? I'm not talking about people who want to test out 5 different versions of ruby and rails simultaneously. I'm just talking about the basic first-time rails developer who wants to get a rails env up and running quickly (which is explicitly the audience for this one-click installer).

If you want or need ruby 1.9.x, then that's a little more work. Install homebrew or MacPorts or whatever and install 1.9 through that, make sure that guy is first in the PATH, and then do the 'gem install' dance. But this is hardly something that needs 5 months of work and $40k to accomplish.

I had to use RVM and edit some config files in order to get Rails to run at all. Might not be a big deal of people who are used to that kind of thing, but it did not make a good first impression to beginner like me.

This doesn’t seem to be isolated to Rails though, Django seems to have gotten harder. I actually gave up installing Django due to a localization bug.

This isn't complicated. 3-4 months of Katz' time costs $40k. That's not a matter of opinion. So the only thing we have to discuss here is, "how much better can Katz make the OS X Rails experience in 3-4 months". As a Mac Rubyist, my guess is, "lots".

Ugh... Your tale sounds all too familiar. I had a very similar situation on my previous MBP. Tried to install some stuff (MAMP was a cohort in the chaos here too...) and ended up spending three days on the command line pretty much at that stage running commands that I copied and lasted from blogs and forums where people had posted after having similar issues but not really knowing from shit what those commands actually did.

Ended up backing away slowly and never looking at RoR again.

Honestly: I am an experienced programmer and all I could make of the install process (plus the attempted installation of some gems) was an utter dogs breakfast.

So as anecdotal evidence goes - add mine to the list please.

I'm not familiar with Locomotive but I would say that most of the time getting my environment working had to do with installing other technologies like resque, mysql, mongo, heroku, gems (devise,haml,aws-ses,rspec,cucumber,pdfkit,js-routes), etc.

I'm guessing the majority of people don't spend the majority of their time installing rails. It's all the other requirements that are needed to build out the application.

Further to my post above: yes - as you have pointed out, if memory serves it was the installation of certain gems that embuggered the whole thing.

On the subject of figuring out where your free space is going, check out http://whatsizemac.com/ - it's great for tracking that kind of thing down.

Where's the proof that Rails has achieved the platonic ideal of OS X support for a dev platform? What, not arguing that at all? Then why is it hard to understand how 5 months of Katz's time could be well spent there?

It's hard to understand because none of the tools the typical developer works with are ideal. But trying to make something perfect that already works well? It's a waste of developer time.

It doesn't work well, and many people spend a day or more setting it up correctly on OS X (there was one commenter in the thread announcing his project that said that he had spent 3 days getting it working right). That's the point.

I can see many ways he could spend 5 months making this great and configurable setup utility, and since so many people will use it if it's good, it's worth one person spending a long time on it. It will probably save 40k of value in terms of time for the world within the first day of its release, and after that, it will keep on giving.

Yeah, cause one thing is for sure, if atomical doesn't need it then no one does, eh?