RE: SVG2 polyfill

Fwiw, David Leunen has been working on FakeSmile [1] that is sort of a polyfill for SMIL-ish things. Am not sure if he’s still actively doing it or not.

 

Then there is the <replicate>[2] and <random>[3] stuff that allows Processing.js type to create interpolations, replications, randomizations, tessellations, nonlinear gradients and 2.5D stuff. Both FakeSmile and <replicate> involve parsing out the modifiers of SVG elements (pattern, filter, gradient, etc) so that as objects are replicated they can be incrementally changed (even using keySplines to adjust the rate of change). See example [4]. It also allows for values of quantitative attributes (not just x and y) to be controlled by values taken from a user-defined SVG path, providing much more precise and intuitive shaping of curves than keySplines would affort.

 

The desire to have declarative randomness in SVG has been often expressed, and this addresses the ability to run skeins that have either fixed or randomized start-seeds. 

 

The core business of parsing values lists, as well as from-to, and interpolation is common between <animate> and <replicate> being based on the same conceptual model of the universe, so some of the code used in replicate could prove useful to new work on polyfills. Replicate also has the advantage that it extends concepts present in SMIL in new directions very consistent with original design intentions of SVG.  Several of these extensions to SMIL’s functionality are things I hope that SVG3 may be willing to consider since a lot of the HTML/CSS/SVG playground rules have been established by SVG2.

 

I think we may have used some of FakeSmile in <replicate> , I can’t recall, but the .js stuff in <replicate> here [5] is open source. Since I’ve recently figured out how to parse CSS rules [6] (never tried it before, but CSS is actually starting to work with SVG in some places now, so I have started to get interested), I’ve been considering extending <replicate> to include the parsing of CSS rules, so that one might take those things that CSS offers that SVG doesn’t yet (like 3D rotation, text flow, and blend modes[7]) and somehow modify during the process of extrapolation along a path.  Not quite sure yet, what gradually modifying a perspective transform along a Bezier curve would give us that modifying an affine one doesn’t, but it could prove interesting! I’ll see if any students are interested in such, and that project may have a bit of new life.

 

Am also interested in using SVG + Replicate +random as a declarative syntax that could then be cross-coded into WebGL, so that a 3D model is quickly sketched in that declarative environment and then poofed (technical term) into a full blown WebGL with wireframe and shaders.  Again a fun project, though I think it might require something like a summer-of-code project rather than just an undergrad volunteer or two.

 

Cheers

David

 

 

[1] https://leunen.me/fakesmile/

[2] http://srufaculty.sru.edu/david.dailey/svg/SVGOpen2010/replicate.htm

[3] http://cs.sru.edu/%7Eddailey/svg/RandomTalk.html 

[4] http://cs.sru.edu/~ddailey/svg/svg2.svg

[5] http://granite.sru.edu/~svg/rep.js <view-source:http://granite.sru.edu/%7Esvg/rep.js> " 
[6] http://cs.sru.edu/~ddailey/svg/3Dsimple0b.svg 

[7] https://ello.co/ddailey/post/hwqd8hmolnwx7-vp4zocfq 

 

From: Amelia Bellamy-Royds [mailto:amelia.bellamy.royds@gmail.com] 
Sent: Wednesday, August 24, 2016 5:24 PM
To: Frédéric Guimont
Cc: www-svg
Subject: Re: SVG2 polyfill

 

I don't know of any polyfills currently being worked on, although I've heard a few people talk about maybe making polyfills for particular features.  It's something I'd like to work on myself, but that's probably not going to happen without funding.

 

It is certainly something working group members would like to see. I'm not sure how much support would be available to help with the project, but do keep us informed if you start work on it.

 

And please file spec issues (https://github.com/w3c/svgwg/issues) if any sections are unclear or inconsistent about what the behavior *should* be.

 

Best,

Amelia Bellamy-Royds

 

On 24 August 2016 at 14:41, Frédéric Guimont <frederic.guimont@savoirfairelinux.com> wrote:

Hi all,

Is there anyone working on an SVG2 polyfill? If not, would there be any interest in such a project? I'm thinking of proposing an R&D project on the topic but I need more info first.

Thanks,
--
Frédéric Guimont, Consultant en logiciels libres
Savoir-faire Linux

Téléphone  : 418-525-7354 #362 <tel:418-525-7354%20%23362> 
Ring ID    : d9396b8004d26120f1e948ac7a075ab7dd165077
www.savoirfairelinux.com