yet when someone fails to handle it, they don't move on until it's fixed.
I don't think I understand that statement? Do you mean: When programmers encounter a problem with handling utf-16, they neither try to correct it, nor fail loudly?
From my perspective, the first problem is that when you read a file, and there is simply no way to know what it contains. Unless you know upfront what is in the file, there is no sensible mechanism for deciding how you should decode (or encode?) the contents of that file. Guess wrong and you produce nonsense. Know wrong and you produce nonsense. Download text from the web and if the webmaster has--through laziness, incompetence or maliciousness--miscoded the mime-type and you're stuffed.
The second problem is, as I said, variable length encoding. With a fixed length encoding, if you want to fetch the 3 millionth character of your data, you add 3e6 to your base and fetch it. With variable length, you have to inspect every single one of the intervening 6 million to 12 million bytes.
Imagine if the postman had to walk the length of Yonge Street inspecting each house name in order to deliver a letter :) (Or IP traffic if...)
In reply to Re^6: Does String::LCSS work?
by BrowserUk
in thread Does String::LCSS work?
by BrowserUk
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |