Archive for July, 2003
“My presentation - It’s crap but at least it’s not a PPT presentation”
HTML spec defines a file format only, SVG spec defines format, reader, viewer. SMIL gives you wonderful markup, but because the content allowed can be any, there’s loads which is actually proprietary because of the content. SVG says what the client can do, so authors and vendors know what they’ve got to do.
What do you need to make an SVG client? XML/XML Namespaces/Unicode/SMIL/PNG/JPEG [CSS 2/HTTP optional] - which is a lot! It gives us real Internationalisation, including visual fidelity of fonts with embedding them.
XLink allows us to put links on other XML NS aswell as SVG.
What we need: XML Events, XForms, XFrames, SMIL audio/video, SOAP, XPath (If it’s a W3 tech, stick it in it seems
Things that need fixing - ID (a general XML problem, which attribute is an ID?) XForms would allow SVG to phone home - styling is left to SVG/CSS - css ain’t really up to it.
Demo’d Deng which can do SVG/CSS/XForms etc. Deng is in Flash!
Plugins SUCK!, plugins are a poor concept, they expose mechanism in content, don’t integrate with browser UI
The SVG core client, SVG gives XML DOM etc. does compositing, rendering etc. CSS box model for text, RCC to create SVG presentation, to give us a rich client for web-services and the Semantic Web.
People already using RCC which seperates pDOM from cDOM, despite it being released 3 days ago.
SVG Tiny competitionTEST Promote and showcase SVG design, 30k SVGT MMS greeting card, giving Fame, fortune (well okay not fortune), but great google karma - w3 is number 14 in the world on google rank. Nokia will give a prise, a 3650… (nice phone)
“Should developers band together with some agenda?” - People could market, perform a consortia to develop a browsing tool, but mostly we need organisation.
SVG has tremendous untapped potential, it’s incredibly powerful and web compatible, tools are in the infancy, and we’ve only begun to appreciate what’s possible.
Graphic Artist, Software Developer, Web Developer <- we need convergence of roles to create web-applications.
We need new authoring tools, export filters are insufficient, tiny fraction of language, the novel features are teh missing ones, and it’s not round-trippable. We need native SVG editors, that focus on design objectives not generating SVG primitives
Native SVG editor, needs IDE and graphic design components.
One SVG many CSS - seperate style from content - re-purpose graphical content, different mediums/priorities etc.
One CSS many SVG - Leave style to stylesheet, change design etc.
Demos of the techniques involved
Lots of discussion on using CSS in SVG, very poorly used at the moment, need to get it out there, it’s important techniques, especially as now we have viewers which export external references.
XSLT to transform XML to SVG, demo of a chart… SVG could be a web drawing API, to render other graphical formats, e.g. MathML/GML/world3D etc.
SVG as graphical UI for web-apps - skinnable components, seperating app-logic from presentation logic.
- Novel features
- Markers and text, e.g. markers on a path that are text, e.g. for cutting cloth demo.
- Text and filters
- Animate filter effects, e.g. shimmering mirrors or light sources moving.
- Animate CSS properties
- Animate Gradients
- Gradients and filters
- Matrix maths with filter effects.
The challenge for software vendors: create a UI aligned with objectives not primitives, and integrate workflow between design and programming elements.Hmm, earlier we were on seperating content from design, why do we need tools that integrate elements we’re seperate them.
The challenge for programmers: take a Graphic design course, combine features to achieve goal, and develop the craft of SVG.Hmm, we should aim on getting workflows that can combine the 2 people working in their own disciplines, and get good products out, programmers will never be good designers.
Philip Mansfield from SchemaSoft is demoing a stunningly good “swell” magnifier - which allows contextual magnification - the normal magnifying glass hides context, demos on UK tubemap, you can see which line is connected/adjacent too when you have a magnifying glass.
Phil’s solution is an SVG/Vector magnifying glass solution, where there’s a transition phase between the zoomed and the original. Tough to describe, but excellent to view.
Interior region is a transformation, nice and simple, the transition region is an feDisplacementMap filter, and pixels from an input SVG image used to spatially displace another image. Phil explains the algorithm, hopefully the slides will be available later, as I’m not live noting them…
Demo is unfortunately not publically available, and we’d need to talk to about licensing… which is a shame, but not unexpected and it’s great!
This is a pretty poor set of notes on the SVG Panel
Initials are www-svg folks or svg-developer guys.
Jon Ferraiolo’s presentation on the stuff formally known as RAX, very well attended not surprisingly.
- SVG Open 2002, had lots of data/presentation seperation issues.
- Lots of discussion between SVGOpen 2002 and 2003, to work out RAX/RCC.
- UI in SVG is hard, want to seperate data & pres. USE element disappoints, mix XForms, mix static/dynamic layout, stuff is positioned precisely, but people want dynamic layout.
- RCC solves or helps the limitations, it’s a foundation for creating widget library, cleanly seperates data vs presentation (different DOM’s), can have layout algorithms, and it’s extensible!
- Examples: simple flowchart example, (From the draft spec), impressive UI demo, checkboxes, select etc. created by Peter Sorotokin. - shows that it’s only 6kb (gzipped) Another Peter demo, which is a tree list. (This is all excellent tech. - libraries are good, but is there a revenue model, will anyone actually pay someone to produce SVG?)
- The entire presentation is RCC’d
- Explanation of syntax.
- What it means? - General component framework, joe can use sally’s components seperate etc.
- It’s the first pass on the syntax, still issues etc. Open issues Authoring tools impact - there’s alternate proposals to do similar Live Templates/dSVG - Please help us review the proposals.
- Should be in Basic, no DOM or script in Tiny, there is Java though on handhelds so maybe a Tiny DOM, or have it in with Java since J2ME exists.
- Learn more from the draft, try them out etc. watch the developers list.
Declaritive iteration idea? can external non-viewer prog do the RCC’ing instead, e.g. from a C?
Script can do the iteration, maybe some XSLT future to do.
Perhaps more events to provide interfaces outside of ASV.
Live Templates guy asks - Weakness of RCC is the content is a direct substitution would XSLT be good, can we have a method to fire off an XSLT?
Jon Agrees fully!
Dean asks to view the source? looks neat
Dean asks where are the tools to generate this stuff?
David Dodds is illustrating some ways of convert SVG documents into a text description of the document, simply by analysing the SVG document, only low level stuff can be done like this, but no real smarts, so not so useful.Various ways to look at an image, Left->Right, Top-Down, Inside-out.
- Shows an SVG document with RDF description of a barchart, RDF schema used is daxschema it contains Below, isNear, AtRight, curve, circle and all sorts of stuff - similar to W3’s linearisation schema (DanielD and Ivan Hermann)
- The RDF generates crap comments, true, but not very English. Converse a linguisitic grammar, (Doug Schepers was just showing me a really interesting Linguistic grammar editor/viewer)
- System can recognise contents and see if it’s a barchart, and answer other questions like “How many rectangles?, is there animation?” or do things like “Put a large green line under the bottom text.
Benjamin Camplin, gave a very interesting talk on his SVG maps which had a haptic (touch paper) output and could talk and interact with the user, unfortunately I had no power for my laptop, so no proper notes, they should be online though, the map of canada he demoed with audio sounds and talking was cool though.
Well attended, about 50 people, Dean apologises it wasn’t better prepared, but he’d spent the night getting the SVG 1.2 spec out.
- SVG folks are fanatical
- In AUS Gents for toilets in Swedish - gents in 50 languages, In canada, just a picture.
- Accessibility is about allowing the most number of people to view your SVG
- who’s it for, everyone, Visual impairment, Audio impairment, learning difficulties, motor difficulties, others e.g. Google.
- W3 takes Accessibility seriously.
- Graphics is the last frontier of universal accessibility.
- What is multimedia at the moment - images, ausio, animation, most stuff on the web is useless, inaccessible - generally need visual ability, and motor control.
- W3 does PNG, HTML, SVG, no audio, no video, SMIL and SVG for animation.
- WAI and SVG - WCAG, ATAG, UAAG review it all.
- SVG is more accessible as it’s vector, it’s textual, and SVG stuff.
- infinite zoomability, componentize SVG elements into groups which can be described.
- SVG is text, view-source etc.
- title, desc, metadata in SVG, what’s missing, structure, references, alternative equivalents.
- SVG text is really text.
- Seperating content/presentation - tough in SVG, stylable with CSS/XSL, user style sheets.
- SVG DOM makes SVG more Accessible!! Thanks for saying this Dean - should help me make the anti-Shadow Tree arguments on RCC
- SVG G element allows you to group elements, define presentation order, use element, etc.
- Try to use semantically rich language for as long as possible, just use SVG for the rendering - use RAX/RCC stick it in the metadata element etc.
- Is Scripting evil? - Device independent events - focusin, focusout, activate.
- Demo of clock, loads of circles to make animated script, DESCRIPTION element in group contains the updating time the same as the clock morphs.
- Shows the W3 logo, the text is text, and inside the SVG doc, is lots of descriptions, telling you all about the image including fonts etc.
- Demos SVG2HTML XSLT (see W3 site
- Use TITLE/DEFS.
- use xlink:title
- structure content
- text as text
- don’t copy elements, use USE.
- watch out for colour blindness.
- Provide control over animation.
- Use your common sense!
- Think about writing a program to get to it.
- The future - Semantic meaning, seperate DOM/rendering, navigation/selecting, usability
- What should we do? - SVG spec will always allow people to make inaccessible. WCAG, UAAG requirements from WAI
- Ronan Oger asks which countries have laws:
- In AUS websites must comply with WCAG 1.0 (Dean doesn’t say A or AAA) - In reality few do, especially with graphics.
- Ronan asks how do we encourage our clients to allow us to author accessibility.
- Cost can be minimal if you do it from the start.
Robert Woodbury’s presentation, Purposes of geometry in the Semantic Web. I missed the very start
- Sub-parts not defined in the SVG tree, where’s recursion?
- SVG provides tools for interacting with semantic information.
- SVG planets demo, orbiting planets showing orbit spans, (lifted from some Oz guys)
- Specify patterns in SVG geometry, if SVG is going to be interesting to designers, needs to be representing patterns, and infer things.
- rules to combine lines into other lines - ie convert a red/blue overlapped line into a single line - creates a minimal number of lines.
- Design Process - Brief, Layout, Massing, Detailing
- DAML and SVG representation of each stage, Describing room layouts
- Sem. functional, Spatial room layout, Spatial wall representation, Sem. wall representation.
- DAML creates functional layout.
- SVG representation converts that to a room layout.
- It’s not the rooms that matter, it’s the walls.
- Transfrom SVG Rects, into SVG lines.
- RDF rep. of walls, has_function, has_room_rightSide, etc.
- Extending from info-viz to design/production - it needs an inference layer.
- I ask “Why an SVG view at all?” Geometric data is info, 1st thing you do with geometric data is visualise let the svg view have it. I’m still not convinced that SVG needs an inferencing capability, but Rob did suggest that, that was my geek bias.