- From: Dr. Olaf Hoffmann <Dr.O.Hoffmann@gmx.de>
- Date: Mon, 14 Apr 2014 20:07:17 +0200
- To: www-svg@w3.org
Stephen Chenney: > The SVG language already contains sufficient functionality to express > the star element use cases presented thus far. > ... > > Philip's argument is that the cost of the > star element is unnecessary. He says nothing at all about the necessity of > other features. > This is your/his personal point of view - or this is a (maybe small) fragment of the world you/he can see from your/his point of view ;o) Unfortunately nobody has the overview from outside, therefore it is useful to collect results from different points of views to get a slightly more complete impression ;o) If we reduce this idea of 'everything is already possible' to the essentials, we can replace SVG completely by PNG and MNG. In general with this one can simulate graphics with scripts on the server and there is no need for vector graphics at all. But in practice there is still a problem, often one does not want to have files of arbitrary size to simulate a specific functionality and one does not want to write (or pay someone to write) programs, that realise the simulation. And therefore SVG is of some use and new features with the capability to reduce file sizes by orders of magnitude are of some use as well. The other simple alternative is just to provide the formula for example for such a 'star' like element and the interested audience can imaging what it is by just interpreting the formula - by the way, this works for most graphis and decoration issues, therefore with this idea we can save all graphics formats and CSS, because we have already the feature 'text' - this provides already 'sufficient functionality to express' everything ;o) And it is pretty effective as well - you just have to write 'apple' - a few bytes for a complex idea or abstraction, typically orders of magnitude more compact than any image of a specific apple - what a compression efficiency ;o) Just say 'star' and everybody gets what you mean ;o) (not really, interpretation especially for 'star' depends strongly on knowledge, culture and personal point of view, therefore hard to write a proposal that covers a wider range of 'star' abstractions of different people, even if we just care about naive graphical 2D-object-abstractions - but this difficulty makes the problem pretty interesting). And another relation concerning this - introducing a new element for such objects is providing vocuabulary, words to express this idea in SVG. As you can see, if we follow this argument or reduction to the essentials consequently and in an efficient way, the internet formats could be reduced again to something very simple, you just need imagination and simple text a story, a fairy tale to get it all together again within the brains of the audience. But unfortunately many people are not satisfied anymore with pretty words, they want alternatives, less effective alternatives to express something. And once again, if this desire is something worth to care about, it is worth to care about the techniques to provide this at least somehow effective in a finite and acceptable time for the audience. Some people want stars or more general objects with rotation symmetry or at least somehow related to polar coordinates, therefore it is worth to think about the question, how to provide this functionality in an effective way, for example no need to study mathematics, natural sciences or techniques to realise some simple shapes, currently in SVG there is a need for this knowledge - I would not have started with SVG at all without having studied before, in my case physics with several lectures in mathematics - but is SVG only intended for academics with the capability to approximate everything by brute force mathematics? Maybe, hopefully not. Obviously to require some basic knowledge depending on the complexity of the problem cannot be avoided for all kinds of objects, but for some commonly known problems SVG can provide features to solve such problems in an elegant and effective way. For example the path element does not only have the M command - would be sufficient as well to approximate every path, but with cubic Bezier curves it is much more efficient. You can use such path data as well to approximate every gradient effect or filter effect - therefore no need for them in a minimalistic scenario, but it is convenient for many people wanting to use gradients or filters effects to have such features. Or animation - should be enough to approximate everything with set elements, but it is much more convenient and effective to have values, keyTimes, keySplines, additive and accumulate to do the job. Therefore the question about the star or polar element is not: 'Do we need this?' It is more: 'Does the proposal simplify common tasks in a significant way?' And because there is more than one proposal: 'Which one covers more shapes of the intended type and is therefore more useful for a wider range of authors to solve problems?' In the past it was not sufficient for the SVG group just to have an element for regular polygons (and maybe directly derived shapes as the boundary of a non convex one). Then it was not enough to have this proposal for a polar element, covering much more shapes like abstractions of blossoms or asymmetric star like structures. But it was ok to put in the requirements for SVG 2 to add a path like element with polar coordinates. This is good for lots of static shapes (if this would at least appear in the SVG 2 draft). But if an author really wants to animate a star or a similar shape in an effective way, something like the star element or the polar element is needed as well. Even worse, a star like structure is often more an object with only the corners/vertices related to polar coordinates and the curves between are often more related to cartesian thinking, therefore a mixture somehow. Will it be often used? Well, symmetric and asymmetric star like objects are used for centuries or millennia often to represent some ideas and organisations. Many technical products have some rotation symmetry or at least parts of the product have. Therefore sure - it is often used. Looking at the past and taking into account the limited capabilies of the human brain, my guess is, that the usage will continue for the future. But this is neither true for the past nor necessarily for the future in SVG, because there is no simply way yet to note such objects. Olaf
Received on Monday, 14 April 2014 18:07:46 UTC