-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 8/1/12 3:42 AM, Sergey Dobrov wrote: > Hello. > > On 08/01/2012 07:43 AM, Mathieu Pasquet wrote: >> >> I am also not sure about the <strong/> and <blockquote/> >> elements: they are shown as a recommended element to support >> (7.8), but the business rules (8.7) states that they should not >> be used, but rather <span/> or <p/> with appropriate style >> attributes. Is it only for backward compatibility, then? > > I'm using the blockquote element very intensive. I don't think that > we should give a preference to style attributes because IM differs > from web and we need messages to be more semantic than to be more > styled. I.e. we need to support the same style for all messages as > strong as it possible to prevent transformation of a chat window to > an Xmas tree, so we need to outline semantic groups rather than > style them to provide a possibility to theme them by recipient > preference. Also, it's easier to parse such message by machines to > provide some advance search possibilities and etc. So I don't > imagine the XHTML-IM without such things as blockquote, cite, code, > strong, em, etc.
As you can imagine, we had some debates about this when we initially worked on XHTML-IM back in 2003-2004. Some people were in favor of a semantic approach, others in favor of a stylistic approach (these debates mirrored somewhat the similar debates in the W3C and the web community in general). I tend to agree now that the semantic approach makes more sense for use in IM, and you have explained the reasoning quite well. >> There is the matter of the <img/> tag that accepts a data:base64 >> as a src, leading to very big stanzas. I think that maybe the XEP >> could state that whenever possible, the use of base64 data should >> be avoided, at least in MUCs, where the message is replicated as >> many times as there are users, leading to high bandwith usage >> (although if I remember correctly, most servers set the max >> stanza size to 10 KiB). Usually something between 10k and 64k. But yes, there are restrictions, and I tend to agree that we should strongly prefer pointers to external images over inline data: URLs. > agree. possibly, we need to prefer XEP-231 for that? Probably. I'll look at the XEP in more detail and propose some text on this list. Peter - -- Peter Saint-Andre https://stpeter.im/ -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.18 (Darwin) Comment: Using GnuPG with Mozilla - http://www.enigmail.net/ iEYEARECAAYFAlBkaB8ACgkQNL8k5A2w/vzmCQCdGvwlx+vf7w/ORwLjGJTCayRy 5j8AnA6XYZshvzdZ1JL2WcZoRKqtrEsq =fhID -----END PGP SIGNATURE-----