On Mon, Oct 17, 2005 at 06:04:41PM +0100, David Cantrell wrote: > On Mon, Oct 17, 2005 at 03:56:40PM +0100, Matthew Garrett wrote: > > And have to rewrite huge quantities of software. UTF-8 has its > > disadvantages, but there's very little hateful about it - it achieves > > what it sets out to do perfectly, and does so without many tools having > > to be rewritten. > > Well, apart from all those tools that have to process utf-8 text, of > course. Like M*As, string manipulation libraries, terminals, ...
MTAs shouldn't be interpreting any utf-8. MUAs should only care about getting the content-type header correct, and optionally converting to the "simplest" character set. Terminals and string manipulation libraries need rewriting, yes. The number of them that exist is quite small compared to the number of applications that store strings. UCS-4 breaks little assumptions like "A byte with a null in is the end of a string". That's fairly hateful. -- Matthew Garrett | mj...@srcf.ucam.org
