There's a difference between stagnant and growing in the right direction. I think the author believes that RoR is moving away from what many thought was one of its core pillars: The ability to go from nothing to an up and running web app quickly.
Lots of things aren't stagnant -- few things move in the right direction. I'm not saying that RoR isn't moving in the right direction, but I think you've created a strawman of sorts.
IMO, as long as Rails is being driven by real application development, I'm happy. Compared to "enterprise" frameworks which are built in isolation from the real world, Rails movement seems more grounded in pragmatism.
How in the world do the changes lead to a situation where you can't get an app up and running quickly? I don't see that at all.
There are other points where the article is plain wrong to. "It used to be that you just used Prototype, now it goes to vast efforts to be agnostic." No, it switched the default to JQuery, thank god.
I came from PHP and did a lot of work with CakePHP and some with Zend Framework. I found CakePHP fun to work with but often limiting. I found ZF unpleasant to work with and often just as limiting because the effort involved in doing certain simple tasks was often too much.
With Rails I was able to crank out some basic functionality quite quickly and the only hang up I've come across is learning how to create plugins. Then again, the plugins I need are often out there in the world and I could probably get away with never having to write one.
I still think Rails is incredibly easy to learn the basics of but it also is well developed and thought out leaving plenty of room for more advanced behavior without running into the same walls I feel I ran into with PHP development.
Since I started with Rails 3, I don't have any perspective on how difficult it is to keep up with the changes but I must say I'm excited about 3.1 and I don't mind having to dig into whatever information I can scrounge up to familiarize myself with it.
I skipped the Rails 3 upgrade just simply because I couldn't justify the time sink. rspec made it ridiculously worse with the difference in Rails support between 1.x and 2.x. I'm guessing with Rails 3.1 coming out I'll have to upgrade just because most of the plugins I use have moved on and don't backport bugfixes.
It'd be nice to evolve an app without having to constantly wasting time on version compatibility. It's an absolute nightmare if you have any non-trivial-sized app.
Very true. Frameworks should strive for a solid API that makes it easy for people to write add-ons, and avoid adding anything that is not essential.
Of course, one hacker's essential is another's bloat, but developers should give things a long hard look before adding stuff.
It's the reason I like Ramaze so much. Has what's needed, gets out of your way, it's easy to evolve or extend if you know Ruby, and it respects your opinions.
This is where you lost me. I spent about 2 hours, aggregate, spent solving upgrade-related problems across a bunch of different apps.
Where is the time sink?
There are entire sites like http://railsplugins.org/ to help you figure out if your plugins are Rails 3 compatible. Unfortunately, there simply aren't enough users to crowdsource the compatibility of all the plugins out there.
But, if you think backward compatibility is vastly overrated, there's really no seeing eye-to-eye on this.
Also there are a lot of deprecated things that must have been removed in 3.1 ― TBH I haven't kept in touch with 3.1 development but I'm sure a lot more has happened besides the default switch and Asset Pipeline. :)
Basically in the end, Rails is a curated framework. It either fits your requirements or it doesn't. There is no need for it to work for "all possible" cases.
Also like I said you don't have to use the defaults. :)
"Opinionated" was the original description. "Curated" is just some buzzword that means "people look after it" rather than "automatically generated", which is, uh, pretty much going to be par for the course for a maintained framework.
"Opinionated" communicates the idea so much better - it means not only that yes, some choices were made, but that some of them are the result of strong opinions, and do not necessarily try and accommodate all points of view.
So basically you agree with point that Rails' rapid changes are making it less approachable, but see it as a good point?
The norm in the community is to actively refactor.
That's it. When the Rails folks redesign the routes system, not only does my app get cleaner, but I learn stuff about how to write a good DSL. When Blake Mizerany wrote Sinatra, not only did we all get a super simple new tool for prototyping web apps in, like, zero seconds.... but we all learned something about how far you can take the "hello world" simplification trend.
We're all trying to get better at programming. We're all trying to advance the state of the art in software design. And we're all learning from each other, and it's awesome.
In many ways it's just a microcosm of what's happening in the Open Source world at large. Node.js is teaching Ruby devs important stuff, just as the Node authors seem like they probably took a cue or two from Sinatra. It's why I left the Microsoft world years ago.
Loose federations of heterogenous interests with high incentives to integrate are so much more interesting than centralized organizations.
When the rails folk redesign the route system it's not a refactoring cause a refactoring keeps functionalities unchanged, it does not add/remove/change corner cases. Which has happened in all major rails releases and some point ones.
"Refactoring" as a term was defined in that book. It has very much been diluted into "I changed some code because I think this is better," which is a shame because it has a very clear definition.
I should also note that having a test suite allows you to refactor with more confidence, but it's neither a requirement nor changes the definition at all. So a "TDD-assisted refactoring" isn't different from a "refactoring."
If you knew C really well, and someone said "Hey, you can be more efficient building a website in C#/ASP.NET", would you have that hard a time believing them?
I wouldn't. But if they said that I could write systems code better with Objective-C I'd say prove it.
Unfortunately whenever I see RoR people talk about ROR vs anything else, including ASP.NET MVC, they talk in broad strokes with statements like, "Obviously dynamic languages are far more productive than static ones." Or "People who program in Ruby are much better developers -- which is obvious to prove, because they program in Ruby."
I'm a former ASP.NET MVP, a codebetter.com blogger, I wrote Foundations of Programming, I own a handful of .NET OSS libraries. Yada, yada, yada..
All of that is pretty meaningless, I'm not saying to show off (most of that stuff isn't show off worthy). However, by all accounts, I'm an expert C#/.NET programmer who's been in the thick of of any progressive .NET movement. I know C# and ASP.NET (MVP or WebForms) much, much better than I know Ruby and Rails. Yet, somehow, I'm still more efficient in RoR (and thats only getting more true). I can get as low level about this discussion as you want, it won't change my opinion.
The two big wins for me: I write a ton less code and code is easier to test (this is largely since IoC is a language feature).
I know C# and ASP.NET (MVP or WebForms) much, much better than I know Ruby and Rails.
The point? If you want stagnant frameworks there's plenty to pick. A lot of them aren't easier to learn, but none of them require you to keep learning. This is great because it tends to attract all the programmers you want to avoid into obvious communities.