- From: Henri Sivonen <hsivonen@hsivonen.fi>
- Date: Mon, 19 Aug 2013 13:38:51 +0300
- To: Heydon Pickering <heydon@heydonworks.com>
- Cc: HTMLWG WG <public-html@w3.org>
On Mon, Aug 19, 2013 at 12:13 PM, Heydon Pickering <heydon@heydonworks.com> wrote: >> Perhaps you could explain, in this case, why so many discussions have >> taken >> place on the subject for such a long time > >>Because the thinking of the semantics for the sake of semantics (as in >>more semantics the better, since semantics are good) as opposed to >>thinking of semantics as a way of getting someone else's receiving >>software to exhibit some commonly useful behavior is a common trap for >>people to fall into. > > Fortunately, I do not personally struggle with the concept of semantics re > their consumption by software. > > Blockquote's lack of a clear metadata solution is an obvious anomaly > (discussed at great length here: http://html5doctor.com/blockquote-q-cite/ > and > http://www.projectevolution.com/activity/how-do-you-solve-a-problem-like-blockquote/ > and > http://oli.jp/2011/blockquote/) and is deserving of a clearer, simpler > solution that makes > easier work of writing and consuming blockquotes for all (human and nonhuman > alike). Based on quick skimming, all those three articles are exemplars of falling into the trap that I was referring to above. They start from the observation that attribution for a quotation is an identifiable piece of text and then jump to the assumption that there should be explicit markup for identifying attributions for quotations. None of them approached the problem from consumption use case side. None of them appear say stuff like: "If the attributions for quotations were explicitly marked up, it would enable us to develop a user-facing browser feature X that would be so useful that the usefulness would justify the cost of the implementation, the standardization and the evangelization of getting a web offers to use the markup. Furthermore, feature X would be useful to have even if the bulk of existing quotations don't use standardized attribution markup." Semantics that don't enable useful user-facing features on the consumption side are waste of everyone's time. Even if the attribution for quotation could be, in principle, explicitly identified, if there's no strong consumption side use case for having it explicitly identified, it doesn't need markup (especially not standardized markup). -- Henri Sivonen hsivonen@hsivonen.fi http://hsivonen.iki.fi/
Received on Monday, 19 August 2013 10:39:25 UTC