DHH has long said that he improves Rails based on what he and 37 Signals needs, and not what the community needs. His attitude seems to be that it's his framework and he can do what he wants to.
So, no. I don't really see the Rails hosting issue being fixed any time soon.
His attitude seems to be that it's his framework and he can do what he wants to.
No. His attitude is that it is your framework, so get to work!
People seem to think that Rails would run just fine on their $5.99 web hosts if the core team would deign to assert its magical force of will. Well, as my old boss used to say: "them" is "us". If you know how to make Rails run, profitably and reliably, on a cheap-ass shared host, go start the business and enjoy your profits. If you don't know, stop assuming that it's because DHH and Ezra Z. are holding out on you. They don't know, either.
Rails is all of four years old, and (unlike PHP) was not designed with the explicit goal of supporting shared hosting. I like Rails, and am glad you like it too, but stop expecting miracles.
I don't think anyone is expecting miracles, or for DHH to make it his personal goal to adapt Rails for the shared hosting environment. The original blog was simply pointing out that if Rails wants to see the kind of widespread adoption PHP has, that's what the community is going to have to do. It's correct.
People keep misinterpreting or misunderstanding the argument.
I think there's two separate issues raised by Dreamhost's post.
1. Should Rails have better shared hosting support?
2. Is it nice for a business that wants to make more money by having a new feature implemented to demand that an open source project make their desire a priority, without any offer of quid pro quo?
I don't see much room for debate in either. 1. Of course 2. Of course not.
2. is of course as well. Rails stands to benefit from having hundreds of shared hosting services promoting their product. The original post wasn't "Rails should support Dreamhost" it was "Rails should support shared hosting."
Rails stands to benefit from all sorts of things, but who's going to put in the effort to make them happen? Open source is all about scratching your own itch. If shared hosting support is itchy for Dreamhost they should be scratching it, not blogging about how they have an itch and someone else should scratch it for them, for free, and now.
They definitely will say that they don't care how popular is, but that is complete bullshit.
37signals is popular because of Rails, the more popular Rails is, the better for 37signals products.
As a Rails programmer, the more popular Rails is, the more worth are your Rails skills. Simple as that.
For a similar reason, the infamous OS wars are not as pointless as they seem. If I invest in Linux or OS X, I also have an interest to protect that investment, which means the OS in question should keep a minimum level of popularity.
"As a Rails programmer, the more popular Rails is, the more worth are your Rails skills."
This is still subject to the laws of supply and demand. Your Rails skillz will be valuable at a time like now - where there are a healthy number of startups using the framework for development, and a sufficient shortage of experienced Rails devs to hire. However, if Rails became popular enough that everyone decided to learn it and the number of available jobs didn't shift in proportion, you'll find your skillz less of a commodity.
So I've looked at heroku a bit, and I have to say it's pretty promising. I assume eventually they'll allow you to run your own domains off of it, maybe offer you email, etc. I actually somehow didn't meet the guys at the first Y C dinner, but I think they're going to make a lot of money.
The problem is that it's only one company. Though I'm sure they'll have clones if/when they are successful, the barrier to entry there is a lot higher than simply supporting a language like PHP on a shared hosting system. (I suppose I'm only guessing here, as I've no experience with either building a Heroku like site or operating a shared hosting one, but the sheer numbers certainly imply that.) So there are always going to be 100x more hosting options for a newbie PHP developer, even if the Heroku model catches on.
I chose PHP because I found scores of web hosts for it, and I knew that if I ever wanted to switch, I had options. Also, I think brand new programmers (think high school kids) get in largely by Googling for web hosts, signing up an account, then coding in whatever is available. If 95% of the web hosts they look at are touting PHP, that's going to play a large factor in their choice.
Not to mention that Facebook, Yahoo, etc. are in PHP, as you can easily see by looking at a URL, and that no really large sites are in Rails, and even if they were you wouldn't know it at a glance.
But then, DHH has never really said that he was passionate about the users of his framework. For an example, see the image below.
http://pocoo.org/~mitsuhiko/rails_motivation.jpg
DHH has long said that he improves Rails based on what he and 37 Signals needs, and not what the community needs. His attitude seems to be that it's his framework and he can do what he wants to.
So, no. I don't really see the Rails hosting issue being fixed any time soon.