- From: Elliotte Harold <elharo@metalab.unc.edu>
- Date: Mon, 11 Oct 2004 19:36:21 -0400
- To: public-xml-binary-comments@w3.org
After skimming the first public working draft of XML Binary Characterization Properties I noticed three properties that seem to be missing but which would be essential to a successful format: 1. Architecture neutral: the format should not unduly disadvantage any particular computing platform. For instance, it should not be significantly slower on a platform that uses big-endian vs, little-endian ints, 8-bit, 32-bit, or 64-bit words, or that has a particular floating point format. 2. Programming language neutral: the format should not unduly disadvantage any particular computing language; e.g. Java or C. Types should not be defined around the native structures of any one language. 3. Human language neutral: the format should not make text in one language larger than it would be in its native encoding. For instance, it would be wrong to require UTF-8 as the sole encoding of text since this sub-optimal for many languages such as Chinese. However, it would be appropriate to allow individual documents to choose the encoding that's most appropriate for them. For instance a document in English might use UTF-8, a document in French might use ISO-8859-1, and a document in Chinese might use UTF-16. However, if only a single encoding can be chosen then it must not compress the English alphabet at the expense of other writing systems. -- Elliotte Rusty Harold elharo@metalab.unc.edu XML in a Nutshell 3rd Edition Just Published! http://www.cafeconleche.org/books/xian3/ http://www.amazon.com/exec/obidos/ISBN%3D0596007647/cafeaulaitA/ref%3Dnosim
Received on Monday, 11 October 2004 23:36:24 UTC