- From: Dominique Hazael-Massieux <dom@w3.org>
- Date: Tue, 01 Jul 2008 15:06:27 +0200
- To: Abel Rionda <abel.rionda@fundacionctic.org>
- Cc: public-mobileok-checker <public-mobileok-checker@w3.org>, Francois Daoust <fd@w3.org>
Le mardi 01 juillet 2008 à 14:53 +0200, Abel Rionda a écrit : > [...] > We think that it is sufficient for Page Size Limit and External > Resources. It might be sufficient, but you would also need to have the ancestor attribute added to the relevant <image>s in case of <object data="foo.png"><img src="foo.gif" /></object> Also, using an attribute this way is going to be problematic if a given URI is referenced several times in a given document; i.e. how would you encode the following doc: <object data="foo.png"><img src="foo.gif" /></object> <object data="foo.svg"><object data="foo.png"><object data="foo.jpg"/></object></object> So, it might actually be quite tricky to get this right... Is it too complicated to use the idea we discussed on the call last week, having something like: <objectInfo loadtype="rendered"/> vs <objectInfo loadtype="tasted"/> based on the object processing rules? > Anyway, if we had the following case: > > <object type="image/jpeg" data="image.jpeg"> > <img src="image.gif"/> > </object> > > Should the gif image be counted for the PageSizeLimit Test? Currently > the checker is taking it into account. No, it shouldn't take it into account (nor in ExternalResources). > BTW, MobileOk document is not > clear about this. Actually, the object element processing rules (linked from PageSizeLimit) specifies that you should stop going down the content of an object element if its type is image/jpeg or image/gif. (it could be clarified, but I think it does say the right thing) Dom
Received on Tuesday, 1 July 2008 13:07:48 UTC