Will_Newsome comments on (Virtual) Employment Open Thread - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (276)
So the good news is, you need none of these "stacks". Really. What you're trying to do here is run before you've learned to walk.
Web programming is as simple as writing a program that does
and redirecting the output to "index.html" when you run this from shell. This gets you a dynamic Web site generated from a program that you wrote.
If that's too simple for you, congratulations: you've mastered enough of the fundamentals to get on to the next step. Something like parametrize the Web page to greet you by your name.
You may be protesting at this point that running the program whenever you want to update the Web site wasn't your idea of "Web programming" - you want a Web site that updates on the fly.
Now we're talking about Web frameworks. But still you don't need a "stack", you need to build up an understanding of how Web servers work, and maybe study a few architectures of historical interest:
With something like PHP or (if you lean toward Ruby) Sinatra, you're reaching the point of building up arbitrarily complex Web pages in response to HTTP requests, on the fly. At this point you'd better have a solid understanding of how the dance unfolds between Web clients (typically browsers, but not just) and Web servers. What are requests and responses, headers, parameters.
Rewrite your Hello World example in Sinatra. The next exercise is to orchestrate a simple interaction where your Web app displays a form with a text box asking for your name, and a Submit button which takes you to a page greeting you by name. Just this little loop requires you to stay on top of a surprisingly large number of abstractions spanning several layers.
Note that at this point you still don't need a database, knowledge of CSS is superfluous, and client-side scripting very much not even on the agenda. Discussion of design patterns like MVC should be deferred until you have experienced for yourself the issues that these patterns are supposed to solve. (That way, you can judge for yourself to what extent the big frameworks like Rails actually solve the issues. Affective Death Spirals are just as deadly in programming as in the rest of your intellectual development.)
When you feel comfortable with this kind of program (check that by writing a more complex one like a calculator or a tic-tac-toe game), you can move on to one that handles arbitrarily large sets of data. By now if you feel like tackling MySQL, go ahead and do that, but you could also focus on your learning programming (as opposed to learning technology). Get your data from a text file, maybe YAML which is like XML, only much more friendly.
Anyway, that's the kind of learning that works for me: instead of getting in way over my head, I try to cut it down into chunks that I can master one at a time.
I'm confused, you're talking about data for a tic-tac-toe game, which should be stored in YAML? I got a tad confused because YAML sounds like HAML and I'd just been reading about HAML, then I realized I'd never used XML and didn't even realize what it was for, so I tried to see if it was related to HTML which I've used a lot of, but it looks like it's not, it's just a way to format text files so that programs/scripts can import data from them. HAML on the other hand is a prettier version of ERB which is a way/template to use Ruby to make HTML use variables. So HAML is this thing I shouldn't be using yet because it's complicated and I should be focusing on Sinatra/Ruby whereas YAML I should start using eventually because it's a way to format data for easy interpretation by ruby scripts and the like. Or did you mean YAML should be used for other things? Am I totally off?
No, something like tic-tac-toe has a bounded set of data to keep track of, no database or other long-term storage required.
Once you've mastered tic-tac-toe or the like the next kind of Web app you're likely to tackle is one with an unbounded data set: for instance, a list of user accounts, or a list of posts and comments. Data like this should survive shutting down and restarting your server.
The traditional approach is to use MySQL to store this kind of data, but I would advise against getting a premature interest in MySQL, as there's a risk that doing so will warp your thinking. YAML is an alternative for storing such data that would keep you more in the Ruby world.
All four of HTML, XML, YAML and HAML share the two initials ML - they're all "markup languages". It's easy to get confused because the differences between them are minute, in the grand scheme of things, and only relevant to people who care so deeply about what tools they work with that it magnifies the differences.
The one you can't do without is HTML, so it makes sense to use only that for a little while. (Actually in the past few years HTML has been redefined as a dialect of XML, so using it means you'll also be using XML.)
Using HAML makes a lot of sense, but there's also a bias it (and ERB, and every similar templating language) induces that you want to be aware of. When you start using these things you start thinking of your generated Web pages as "documents with Ruby variables".
This is fine for some class of features but it will wreck your design skill if you generalize the approach too eagerly. Some (indeed many) portions of a significant Web app should be thought of as Ruby objects with the ability to output a HTML representation of them. Things like row/column tables; menus; data entry forms; recursively nested structures like LW comment threads; and so on.
YAML, and pretty much every data serialization method is overkill for Tic-Tac-Toe - you can store a game just as a few integers if you represent the board as a magic square: http://www.reddit.com/r/programming/comments/9904e/interesting_alternative_to_captcha_pic/c0bw41g
Neat!