Thank you, Rails

http://jacobian.org/writing/thank-you-rails/

by jacobian • 17 years ago

297 77 17 years ago

jon_dahl• 17 years ago

This is the tone that every language/framework comparison should take. Thank you, jacobian!

Amen. I recently dove into Python/Django (for a startup) and the whole experience really has been a blast, in no small part due to Django (I have a strong penchant for good tools).

Now that I've got a bit of experience under my belt I'm looking for more opportunities to work with this community and I love what I see. I feel like working alongside these guys is making me into a much better developer.

iamelgringo2• 17 years ago

The django core devs are one of the best thing the framework has. I'd have a beer with any of them at the drop of a hat.

subbu2• 17 years ago

You made me google for Django core team just to make sure I remember all of their names http://docs.djangoproject.com/en/dev/internals/committers/#t...

andrewvc• 17 years ago

This article is spot on.

It's a tribute to Rails' genius that so many people view it as a threat. You still see people who seem genuinely afraid to let go of their massive, overly complex frameworks, or optimizing every little bit of their PHP app. People afraid of relinquishing the slightest bit of control for convention.

What I actually see, when I hear developers decrying this stuff, is geeks afraid that they're going to lose their Genius cards now that there's an easier way.

The fact is, Rails doesn't reduce the amount of hard thinking required to write a reasonably complicated web app, it reduces the amount of boring boiler plate code. For heavy lifting Rails still gives you the tools you need, there's a reason Rails Metal and find_by_sql exist. There's a reason most of the core parts of Rails are interchangeable (esp. with the upcoming Rails 3). Solving any worthwhile problem generally requires real work. Rails doesn't change this.

The brilliance that Rails brings to the table is balance. Many of the design decisions of Rails just hit the sweet spot for a large category of apps, and remove so many of the pain points in development.

dasil0032• 17 years ago

I sort of agree, but you're coming off a little over-excited there, and I say that as a 5-year Rails developer.

The thing about Rails is that it represents the maturation of web development. It's not anything amazing per se, just a clever roll up of the best practices that emerged from the decade of Perl, Java, PHP development that preceded it. Ruby is the secret sauce that makes things just feel so slick.

I don't mean to minimize the initial innovation, it's just that Rails is no longer very unique. A lot of frameworks have largely caught up even if written in less flexible languages, and long term I see opportunities for new frameworks to leapfrog Rails.

The biggest liability for Rails is the Ruby runtime. Yes progress is being made here, but it's still a pretty sad state of affairs with regards to memory leaks and performance. I bang my head up against this every few months and it really hurts.

Another issue is the need to handle more complex applications. Frameworks like Seaside allow you to set up complex stateful apps in a more robust way than will probably ever be technically possible in Rails (again due to limitations with Continuations in the Ruby runtime).

A related issue is the question of what dark places we will end up in over a couple decades of large code bases running under Ruby with no warnings. I think it's pretty clear by now that Java is an unnecessary straitjacket, but do-whatever-the-fuck-you-want-with-no-warnings-whatsoever Ruby environment is not necessarily the most conducive to long term stability. I'll take it over Java any day, but I still have my doubts.

Finally, there is the issue of increased ajax-ification of apps. Rails was a leader here early on insomuch as it provided proof of concept that Ajax could be done easily without a cross-browser-DHTML-guru certification, but 5 years later the Prototype + Rails helpers paradigm is quite dated. Ajax has huge potential for simplifying back-end code and making web apps inherently more scalable and performant. This is one area where Rails could re-invent itself, but a mature framework like Rails is going to have a disadvantage against a fresh approach starting from first principles.

In short, I think the competition has openings.

patio112• 17 years ago

I love Rails to death and do Big Freaking Enterprise Web App Development in Java all day.

A lot of what people identify as pathological is an evolutionary advantage in BFEWAD. Straight-jacketing the programmer is often a good thing because on a team of 100 programmers in three continents (you might think I exaggerate), including (by my last count) 12 programmers who have less than 3 months of professional experience, we want to make sure the damage done by the least skilled programmer on his worst day is strictly contained.

I don't know if Rails scales to that sort of development, to be honest. The least skilled developer, touching one part of one class at the periphery of the project, is essentially one line of code away from blowing up the entire project.

I should know: since I'm the sole Rails programmer at my business, I am by definition the least skilled programmer. And since I've managed to corrupt the results of my analytics code by making a one-line ill-considered addition to my printing code (which are kept in wildly divergent classes in different packages and namespaced from each other), just because it was not obvious I was altering the global state of the interpreter. That sort of bug has never happened to me in Java.

A lot of techniques which are considered kosher in the Rails community don't scale to large distributed teams really well. Monkey patching is the obvious example -- if you have two developers who have different understandings of what Array#pack is supposed to do, you're going to run into HELLACIOUS, difficult to debug bugs. If you do not quickly clue into the fact that Array#pack was redefined in a file you've never seen before that was committed last Tuesday by your most junior developer in the Seoul office, enjoy spending the next several days going over code you swear looks right looking for the error.

tptacek7• 17 years ago

The Rails Javascript helpers are a huge blind alley, and something that should have been jettisoned a while ago --- as is RJS.

I got addicted to Sinatra because it pretty much begged me to try Haml + jQ --- which is an amazing combination --- and we recently ported all our Erb's to Haml in our Rails app as well.

I think it's a major design mistake to hide or abstract Javascript now that jQuery has pretty much won the JS framework war.

blasdel2• 17 years ago

The Rails novelty I've been most thankful for is their initial popularization of the use of non-monolothic web servers. Every webapp running on your machine doesn't have to be served up end-to-end by a single Apache config, much less with your goddamn programming language runtimes linked in to the same binary -- but nobody seemed to know that at the time.

Yes, there was plenty of emotional bullshit in the community as solutions were hyped and abandoned. I think Passenger was initially a massive logical regression when it was mod_rails, but now that it's mod_rack it no longer agitates me so.

Thank you, Rails, for helping the world realize how awesome HTTP is as seamless, simplifying middleware.

dasil0032• 17 years ago

I dunno man, for most of the apps I run (2k-10k lines), the memory requirement is usually the biggest pain point.

blasdel2• 17 years ago

No doubt, MRI has the worst GC of any programming language implementation in active use today, and the design of ActiveRecord/ActiveSupport/etc. push it over the cliff.

Even if it mostly failed at pulling it off, Rails convinced even some PHBs that Apache is not the alpha and omega of HTTP.

kevinholesh• 17 years ago

This is exactly the kind of attitude I like to see. Rather than argue why a certain framework or language is better, I'd rather reflect on how much things have changed. I remember 5 years ago programming in object oriented PHP.

I'm much happier now with Rails and Django.

Jim_Neath• 17 years ago

I hear that.

I spent nearly 4 years doing OO PHP. I enjoy coding a lot more now I'm rocking ruby/rails.

If anything this article has made me more inclined to check out python/django.

dannyr• 17 years ago

jacobian is on fire with his blog posts lately. However, what he said that "Windows folk look down on us Mac users" is in my experience false. It is definitely the other way around.

amoeba• 17 years ago

It goes both ways and I'm not sure one side is guiltier than the other.

rimantas• 17 years ago

My impression is that Mac users tend to look down on Windows—the OS itself not its users, while liking Apple products makes you douchebag, sheeple and brainless victim of Apple marketing. Maybe that's because the most of Apple users do have experience with MS products which is not always the case on the opposite side.

mreid• 17 years ago

My guess is that it was a bit of self-deprecating humour.

eam• 17 years ago

I liked the civil approach!

sailormoon• 17 years ago

Excellent article!

10ren• 17 years ago

The "love or money" struck a chord. It's similar to "when your hobby becomes your job, you no longer have a hobby." It certainly adds new concerns. There's the finding that rewards can be demotivating: http://www.gnu.org/philosophy/motivation.html (HN: http://news.ycombinator.com/item?id=783939)

There's pg's idea, of working hard to make lots of money, so you are free to work on what you really enjoy.

And 37Signals' idea, of a lifestyle business that you enjoy working at in the first place (assuming you resolve the above added concerns and demotivation).

mattjaynes• 17 years ago

My first thought when reading this is "The Django guys even think about Rails?"

I don't know much about the Django community, but in the Rails community I seldom hear any discussion of Django. The few times I've heard talk of Django in the Rails community, it's along the lines of "Oh yeah, it's great for x/y/z use cases." It's hard for me to imagine DHH or other Rails notables writing anything remotely similar to this post.

Is this just me being sheltered? Or is there more of a pre-occupation with Rails in the Django community than vice-versa? If so, why? Just curious to hear what other's experiences have been.

davidw• 17 years ago

I think that some of the Python guys have a bit of Ruby envy: Python has steadily been gaining momentum over the years as the "scripting language", and along came Ruby and in the blink of an eye it had, at least in terms of publicity, surpassed Python, at least where web stuff was concerned, thanks to Rails.

carbon84• 17 years ago

Yup. Ian Bicking talked specifically about this at pycon last year.

iamelgringo2• 17 years ago

There's generally an awareness and a decent discussion in the django community about other web frameworks in general. Even though I use Django full time, I usually keep an awareness of happenings in other Python frameworks like TurboGears, Pylons. And I usually tend to have an awareness of big events that are happening in Rails.

bad_user2• 17 years ago

I don't know much about the Rails community, but they should definitely look at Django.

Django currently is more modular. The ease with which you can integrate third-party components is unparalleled in my experience. I also like Django's ORM a whole lot more.

Rails is awesome, other frameworks are too ... but keeping blinders on is not good on the long term.

jeremymcanally• 17 years ago

Rails 3 is taking a turn towards that direction. More modular, open to more ORM's (we have more than ActiveRecord these days, you know! :P), etc. And I'm happy for it. I really like that aspect of Django (and somewhat of Merb, though it was painful sometimes with all the deps), so I'm glad to see it coming to Rails.

carbon84• 17 years ago

Rails borrows plenty from django and python at large. Merb's slices and Rails' engines were directly inspired by Django's apps. Rails is getting automatic escaping of strings in templates after Django showed it was a better default. Rack was directly inspired by WSGI. Rip is based on pip. Rails developers aren't "keeping blinders on." Exactly the opposite.

megamark16• 17 years ago

I wonder if that is due, in part, to the fact that there are so many web frameworks for Python (Django, TurboGears, Pylons, etc) that maybe we tend to discuss web frameworks in general a little more then the Rails guys. I'm sure there are some out there but I've never heard of any other web frameworks for Ruby.

bct• 17 years ago

Ruby has a lot of web frameworks too. All your experience means is that you don't pay much attention to Ruby (which is fine, I'm only vaguely aware of the 3 you listed for Python)

jamesbritt3• 17 years ago

"I'm sure there are some out there but I've never heard of any other web frameworks for Ruby."

Which is real shame, because there are many really good options for Ruby Web development.

What people are saying here abut Rails I felt about Nitro, whihc came out at the same time but had a more well-considered implementation. It was thread-safe from the start, much faster than Rails, did less mucking around (if any) with core Ruby classes, made it easy to swap in different ORMs, offered a trnsformation pipeline that was simialr to Rack middleware; in fact, it seemed much like what Rails3 is going to be, only five years ago.

But it got no traction in Rubyland. In fact, at one MountainWest RubyConf, Chad Fowler flat out told people to stop working on Nitro, because there was Rails.

There's a weird myopia among many Rubyists, a lack of curiosity about what else is happening.

nopal• 17 years ago

This attitude is one reason that I really like the Django community. I've found, almost without fail, that most Djangonauts are friendly and patient with the framework's users, new and old alike.

carbon84• 17 years ago

Yeah, friendly with django users, which is nice if that's all you are. The problem, however, is what he mentions in the post: "I’ve noticed the tone of the arguments in the Django community getting nastier — especially when it comes to Rails." Hopefully the ideas in this post catch on and I'm glad a prominent Django developer is calling it out.

I mentioned Ian Bicking's great talk at this year's PyCon elsewhere in these comments, but another thing he touched on was framework fanboyism, and how it's mainly the result of people being new to the framework and trying to convince themselves that it's the answer to everything. Whether it's Django or Rails or any other technology, excitement is good, but not when it becomes tribal and adversarial toward perceived competitors.

recurser• 17 years ago

Ditto, every time I get stuck and hit the django IRC channel I'm amazed how friendly and helpful the devs are. I don't think I've ever left there without learning something.

marknutter4• 17 years ago

I love this post. I've only recently gotten into development (< 3 years) and my biggest complaint about the development world is all the fighting between languages, frameworks, etc. I credit Rails with being the reason I was able to get into development in the first place - it made web development extremely accessible to people like me with little to no app development experience.

So all that said, I really appreciate the author's candor. The negative attacks people make against other frameworks and languages really discourage people who are just getting started with this stuff because it casts doubt in their minds about whether or not they bet on the right horse. I may actually play around with Django now to see what it's all about.

clutchski2• 17 years ago

In addition to the cultural shift it brought to the web community, Rails' is due a debt of gratitude for its technical innovations as well. Django, Merb, Pylons and others are all iterations on the Rails approach to web development. Django, it seems, is like a teenage boy nearing the end of his adolescence and realizes that just maybe his parents aren't as stupid as he once thought.

jdunck• 17 years ago

FWIW, Django was in private use before Rails was made public, so while Django and Rails are kindred, it's not reasonable to call Django an iteration of Rails.

clutchski2• 17 years ago

good to know. thanks.

Thank you, Rails

Related Stories