- From: Ivan Herman via GitHub <sysbot+gh@w3.org>
- Date: Mon, 28 Mar 2016 09:54:02 +0000
- To: public-annotation@w3.org
> It's consistent with the embedded content pattern, rather than the values in selectors pattern. The use of text here is because we used to use Content in RDF to embed the SVG resource's representation. > Well, that is past. Not relevant any more for this version... > The same with the CSS Style. I think this is correct, rather than value ... we're embedding the content of the representation into the graph, rather than providing a stand alone value that is not a representation. > > I do not find this argument convincing. For the selector, the *role* of the textual svg content is the same as, say, the xpath content for the XPath Selector: it is the value that governs the selection for an XPath selection. The fact that we are talking about some content that, *in other contexts* is interpretable as a stand alone image, is irrelevant. Note that the "Content" class ('pattern', if you refer to) appears twice. One is the SVG Selector, the other is Style (I admit I did not look at the latter for my comment). But for Style we use `value` and not `text`, so we _are_ inconsistent. B.t.w., in the Style element there is also a `format` argument, that is irrelevant or, more exactly, superfluous, because we are talking about a CssStylesheet object! I have the impression that the Content class itself should be therefore removed from the model altogether, and simplify both usages. It is an overcomplication without any benefit that I would see. > If there isn't an id, then it's just a blank node. So long as the representation is available from the URI in id, I don't see a problem. > I am not sure what you refer to. > That said I do see the point about type-jacking. We solved that already for Semantic Tags where the resources were very likely not to be ours. On the other hand, here it seems that finding an SVG document in the wild that happened to express the right area for the selection is so improbable as to not be worth worrying about? > It is again a matter of consistency. The Open Web Assumption has already directed our deliberation in the past, and we decided to be consistent with it (see our discussion on roles). We should apply the same here. Bottom line: I stand by my original proposal... -- GitHub Notification of comment by iherman Please view or discuss this issue at https://github.com/w3c/web-annotation/issues/191#issuecomment-202325540 using your GitHub account
Received on Monday, 28 March 2016 09:54:05 UTC