- From: Doug Schepers <doug@schepers.cc>
- Date: Wed, 2 Mar 2005 02:27:43 -0500
- To: "'Dean Jackson'" <dean@w3.org>, "'Chris Lilley'" <chris@w3.org>, <www-svg@w3.org>
Hi, Dean and Chris- Thanks for the quick and reassuring replies. Great to hear we've still got arbitrary flow. I guess, though, due to unforeseen complexity, that the flowing mechanism detailed in the Spec will only be a guideline, and not normative? Not ideal for interop, but if that seems like an unattainable goal, then so be it. Regards- Doug doug . schepers @ vectoreal.com www.vectoreal.com ...for scalable solutions. Dean Jackson wrote: | | | Hey Doug, | | On 1 Mar 2005, at 12:23, Doug Schepers wrote: | | > | > Dear SVG-WG- | > | > The wording here seems to indicate that text-flow to | arbitrary shapes | > may have been dropped. I hope that this is not the case. I | feel that | > that is a well-designed and useful feature, | | It hasn't been dropped. It certainly will not be in SVG Tiny | 1.2, but it will most probably be in the next draft of the | full specification. | | > and fail to see how it intersects with CSS's text layout | properties, | > which can do no such thing. | | Right. | | > | > It has been suggested on this list that most authors do not want | > pixel-level control over appearance and that specifying a mandatory | > algorithm in the Spec restricts UA implementors from using custom | > solutions to achieve the desired end (thus reducing a possible | > competitive advantage). Most SVG authors (and Web authors | in general) | > I know are more interested in control and consistency of | performance | > and appearance between implementations than in proprietary | performance | > gains. I think it is crucial that authors know what to | expect across | > various UA and Oses; otherwise authoring is a nightmare (as in the | > Browser Wars). | | The problem we face is that it is extremely difficult to | define an algorithm that does automatic line-breaking in a | well-internationalized manner. We had long discussions with | the W3C I18N Working Group who suggested a pretty simple | algorithm (fairly similar to what was already in SVG 1.2) but | with the warning that it was completely unsuitable for a lot | of scripts. They, and others on this list, said we needed to | make it possible for the user agent to implement its own | algorithm (or use the platform). Leaving this out was an | oversight on our behalf. | | It would be nice to have a specification of good line | breaking, but I've been told that it is an extremely hard | thing to do, and is only mastered by a handful of people at | large companies that do lots of text work (eg. Microsoft with | Word, Adobe with InDesign, etc). | | > Please do keep textflow in arbitrary shapes, and do mandate a | > particular algorithm to do so. If a UA implementor wishes to add an | > additional algorithm, allow them to do so using an explicit | attribute | > value like text-breaking='custom', or somesuch. | | That's the current plan. | | > If I'm being paranoid and jumping to wrong conclusions, | um... Never mind, | > then. As you were. | | Just because you are paranoid doesn't mean we're not out to get you. | | Dean Chris Lilley wrote: | | On Monday, February 28, 2005, 8:23:25 PM, Doug wrote: | | DS> Please do keep textflow in arbitrary shapes, and do mandate a | DS> particular algorithm to do so. If a UA implementor wishes | to add an | DS> additional algorithm, allow them to do so using an explicit | DS> attribute value like text-breaking='custom', or somesuch. | | DS> If I'm being paranoid and jumping to wrong conclusions, | um... Never | DS> mind, then. As you were. | | Tiny 1.2 has flowing text in rectangles, as it did before. | | Full 1.2 has flowing text in arbitrary shapes, as it did before. | | User agents are allowed to pick the best line break | opportunities for any language, for example a Thai SVG | implementation is allowed to use a Thai dictionary to do line | breaking between words (Thai has no spaces between words, and | only breaks between words). They were forbidden to do that | before, in an attempt to get consistency between implementations. | However it didn't actually give that consistency. So we | changed it. In particular, an implementation that already has | a CSS line layout engine can re-use that and be conformant. | | Algorithms will be provided, to help implementors. But better | ones can also be used, if available. | | -- | Chris Lilley mailto:chris@w3.org | Chair, W3C SVG Working Group | W3C Graphics Activity Lead
Received on Wednesday, 2 March 2005 07:27:49 UTC