On 2008-02-14, Conrad Parker wrote:> I tend to disagree with your sentiment. The specification of any > format or protocol has mandatory and recommended sections (not > "features"); MUST and SHOULD respectively for IETF and W3C stuff.Then why not make the common endianness MUST and the rest of it SHOULD? That was my sentiment, after all... -- Sampo Syreeni, aka decoy - mailto:decoy@iki.fi, tel:+358-50-5756111 student/math+cs/helsinki university, http://www.iki.fi/~decoy/front openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2
Jean-Marc Valin
2008-Feb-13 18:48 UTC
[ogg-dev] OggPCM: support for little-endianness only?
Sampo Syreeni wrote:> On 2008-02-14, Conrad Parker wrote: > >> I tend to disagree with your sentiment. The specification of any >> format or protocol has mandatory and recommended sections (not >> "features"); MUST and SHOULD respectively for IETF and W3C stuff. > > Then why not make the common endianness MUST and the rest of it SHOULD? > That was my sentiment, after all...Because "SHOULD encode using $less_common_endian" translates into "MUST be able to *decode* $less_common_endian". Or if you meant "SHOULD be able de decode $less_common_endian", then the spec is just plain broken. Jean-Marc
On 2008-02-14, Jean-Marc Valin wrote:>> Then why not make the common endianness MUST and the rest of it SHOULD? >> That was my sentiment, after all... > > [...] Or if you meant "SHOULD be able de decode $less_common_endian", > then the spec is just plain broken.This is what I meant, precisely. Why is it broken then? There is a clear expectation that certain things can always be decoded, as per specification, and there is also the possibility of doing something in excess of the base specification. How is that broken, exactly? -- Sampo Syreeni, aka decoy - mailto:decoy@iki.fi, tel:+358-50-5756111 student/math+cs/helsinki university, http://www.iki.fi/~decoy/front openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2