Internationalization review of Web Workers spec

As requested by the Internationalization Core WG [1], I have reviewed the Web Workers specification draft of 2012-03-13 [2] from an internationalization point of view.

I found no internationalization issues in this specification.

Here are more detailed notes from the review:

- Sections 4.6, Processing model, and 5.1, Importing scripts and libraries, specify that the source code of web workers is decoded from UTF-8, with no allowance for other encodings. This follows our published best practices for choosing character encodings in specifications [3].

- The same sections also specify UTF-8 as the character encoding for non-ASCII characters in the query component of URLs, with no allowance for other encodings. This follows the same best practices.

- Throughout the specification, URLs are resolved to absolute URLs as specified by the HTML5 specification [4], and managed as absolute URLs. The URL resolution algorithm in HTML5 converts to an ASCII form, using IDNA ToASCII conversion for the host component and percent-encoding based on UTF-8 for the path and query components (the latter can use other encodings in HTML5, but Web Workers uses UTF-8, as noted above). The algorithm intentionally deviates from the relevant RFCs in a few points. I assume these choices have been carefully evaluated in the context of HTML5; any issues would have to be discussed in that context. It is good that the Web Workers specification reuses the HTML5 model.

- I found and reported two editorial issues that are not related to internationalization [5, 6].



Received on Monday, 16 April 2012 02:23:27 UTC