() Patrice Dumas <[email protected]> () Sun, 24 Feb 2013 19:20:47 +0100
On Sat, Feb 16, 2013 at 04:16:02PM +0100, Thien-Thi Nguyen wrote: > > - mkixin records blob lengths individually, not cumulatively > > Before, for two blobs of lengths A and B, BLOBS-INDEX would > record lengths A and A+B. Regarding blobs indexing, I think that it would be better if there was a separate length and position. That way it could be possible to intersped blobs with text, for example text that describes the blob and would be used as a comment. For example truc.png jsfdjkfsdkjfsdkjfsdjkfdsjfsd machin.jpg mqdsmlsdqljsqlmdjqmljdsmjlqsdmljsdmldsqlmqjdsdqslm So the index would be 2 informations, the length of the blob and an information that allows to find the blob, either the absolute position in the blobs, or the total length (blob+following text between blobs). Interspersing blobs with text seems like a step backward to me. The current (lazy) design is nicer to low-resource renderers. Re redundancy (position info), that suggests that there should be some kind of integrity-checking tool. Hmmm. FYI, IXIN source is now in a public repo, and there is also a dedicated mailing list (reachable from the project page): http://git.savannah.gnu.org/cgit/ixin.git?h=p http://savannah.nongnu.org/projects/ixin I don't mind continuing its discussion here, of course. -- Thien-Thi Nguyen GPG key: 4C807502 (if you're human and you know it) read my lisp: (responsep (questions 'technical) (not (via 'mailing-list))) => nil
signature.asc
Description: PGP signature
