J. O'Conner

As you pass data from the browser to the application server to the database, opportunities for data loss lurk. I highlighted some of those conversion points earlier, but I neglected a browser issue. The JavaScript layer has its own lossy points of interest. One of those points is the escape function.

The escape function “encodes” a string by replacing non-ASCII letters and some other punctuation symbols with escape sequences of the form %XX, where X is a hex digit. Unicode characters from \u0080 through \u00FF are converted to the %XX form as well. Unicode characters in higher ranges take the form %uXXXX. So, as an example, the name José will take the form Jos%E9. Go ahead, give it a try below:

The problem with this is that the escape mechanism is broken if you want to use UTF-8 as your document encoding. If you were dynamically composing URL strings with parameters, those parameters will definitely not be escaped correctly. Instead of Jos%E9 that URI component should really be Jos%C3%A9.

Fortunately, JavaScript has resolved the problem, but the solution means you’ll have to use another function. The escape function is deprecated in ECMAScript v 3. Instead, you should use the function encodeURI or encodeURIComponent. These functions convert their argument to the UTF-8 encoding and then %XX encode all the non-ASCII characters. Two forms of the function exist so that you have greater control over whether characters like “?” and “&” are encoded. You’ll need to check your documentation for details. You can experiment with the encodeURIComponent function here:

What’s this mean for you? Maybe nothing if you’re hopelessly attached to ISO-8859-1. However, if you’re trying to reach a global market with your product, chances are very good that you’ve decided to use UTF-8 for your character set encoding. That’s an excellent choice, but you’ll have to manage the conversion points. In a nutshell, that simply means that you’ll need to use UTF-8 from front to back consistently.

Part of managing those conversion points is consistently providing well-formed URIs to your application server. If you use JavaScript to manipulate data or to create dynamic URIs in your application, make sure you toss aside that deprecated escape function. Take a look at encodeURI and encodeURIComponent instead.

Although JavaScript itself uses Unicode internally, you can still run into charset conversion problems. Consider the following example of charset conversion issues with a very simple HTML and JS file. Continue reading ‘JavaScript file encoding’ »

Last week a friend asked an interesting web-based localization question. He surprised me with it. I wish that I had considered it before, but I had not. In my less than complete analysis of his problem, I found a solution, but I don’t know if he’ll like it. Hell, I barely like it myself, but it’s what I came up with quickly.

Here’s the question:

I want to localize a bit of JavaScript, nothing fancy now, just localized text strings. Once the JavaScript in pulled down into my browser, I don’t want to make any more trips to the server to get text. You have any idea how I can get localized strings in my JavaScript?

So here are a couple options:

Continue reading ‘Localizing JavaScript with JSPs’ »