- From: Larry Masinter <masinter@attlabs.att.com>
- Date: Fri, 14 Jul 2000 18:45:36 -0700
- To: "C. M. Sperberg-McQueen" <cmsmcq@acm.org>, "Martin J. Duerst" <duerst@w3.org>, <www-xml-schema-comments@w3.org>
Re coding systems for strings: I don't understand your lack of sympathy. Every 'string' in a every programming language I know potentially can contain control characters that are excluded from XML. (Actually, I can think of *one* programming language, MOO, that at one time didn't allow control characters in strings). Many database systems are set up so that most strings allow control characters in them that are excluded from XML. What a schema language *could* do would be to invent or select a public convention for including such characters. For example, RFCs 2047 and 2231 include ugly encodings that would work for this purpose. Perhaps you can do better, but it seems unreasonable to do nothing. The problem is that a program should be able to RELIABLY translate from internal strings to XML data, without having to raise exceptions about 'unencodable string character'.
Received on Friday, 14 July 2000 21:45:34 UTC