Though it’s Friday back home, it’s Saturday here in Sydney, and it’s been a a long week of meetings. Or meeting, I should say… the first SVG WG F2F of 2007, and my reunion of sorts. I first joined the SVG WG last January, and my first meeting was in Sydney. The winter (read: summer) Sydney meeting, hosted by Canon, has become a tradition. I sure can’t complain. As usual for a face-to-face, we’ve been sequestered in a hotel room all day, though with a beautiful view of Manly Beach right out the sliding doors. At night, we go out for dinner, then back to work, and to sleep. Read on for a day-to-day account of my trip thus far…
I rushed to the airport, but I wasn’t the only one who was running late. My flight was delayed, which meant that I would miss my connection in LA for Sydney. Surprisingly, United Airlines did the right thing, booking me on a non-stop cross-country flight on Delta to get me there on time. It was tight, and I arrived at the departure gate after the boarding had already begun… only to find that I needed an Australian visa to board! Last year, this was taken care of by the travel agency, so I didn’t suspect I’d need one. Luckily, they filed one electronically. I’d only bought my ticket 2 days before (I wasn’t sure I was going to be able to make the meeting until the last minute), so I guess everything worked out as well as could be hoped. When I got to my seat (not as nice as Qantas by half), I was pleased to discover I was one of a few with a row of 3 seats to myself. I stretched out and was promptly unconscious.
Monday disappeared into the aether of international time zones and sleep.
I arrived well-rested in Sydney around 8:30 in the morning, and took the train to Circular Quay (which Australians quaintly pronounce “key”), then took a ferry across the harbor to Manly. I sat out on the deck, affording me a gorgeous view of the water, the Sydney Harbour Bridge, and the famous Opera House. I really enjoyed the crossing, with the fine weather and smell of brine on the breeze. I arrived just after the first day’s meeting had jumbled together. I popped into the conference suite’s shower to freshen up, and we spent the day resolving the SVG Tiny 1.2 test suite, among other issues. I’m staying with my friend Andrew Shellshear of Canon, so we split to his house in Marrickville for the night right after dinner.
I was surprisingly tired for having slept through the flight; I’m always surprised how much travelling takes it out of you. Unfortunately, I had to get up at 3 AM for a 2.5-hour meeting back at the office, and couldn’t get back to sleep before we had to leave again for the day.
We began the process of creating an errata for SVG 1.1, yet another time trip. This task stretches back years before I joined the WG, back to 2003 or earlier. In short, an errata is a list of corrections and clarifications for problems with a spec, mostly raised by implementors who found the parts of the spec to be incomplete, too hard to implement, or in conflict with other specs; ambiguity in a spec leads to bad interoperability between browsers, and is an author’s nightmare, especially with designers who want pixel-level control over the appearance of their graphics. Sadly, the errata has been so long in coming that it has an almost mythical status at this point. The task of compiling these issues has fallen on the shoulders of several people over the last 5 years, and has defeated them all… until now! The aforementioned Andrew rose to the task and we spent most of the day in formal group review of each issue. Some of them were so old they were no longer relevant, and others required a good bit of spec archaeology to suss out, especially with the high turnover the WG has had over the years. Each time someone leaves the group, a lot of wisdom passes with them… Jon Ferraiolo (erstwhile of Adobe, and who left the group several months ago) was especially clued in to the intricacies of SVG’s history, and each person knew a certain piece of the pie. Recreating the decision-making process is not pretty. But some issues become clear through hindsight, so we were able to get quite a lot done. I was pretty tired from jetlag and my meeting in the wee hours, but a short mid-morning nap brought me back to.
On our third day, we went through the various SVG modules we intend to publish, including Filters, Vector Effects, Fonts, Pinned Clip Video, Authoring Tool Guidelines, and my own Accessibility Note. We are moving to a modular approach to specs for a number of reasons: they are easier and faster to spec out; easier to review and comment on; better for non-SVG technologies and specs (such as HTML) to reference and to use; and easier for developers to implement and for clients to choose which features they want in an implementation. The chief danger is that the different modules have to be designed carefully to integrate well with each other, but this is a risk in monolithic specs as well. We plan to avoid splintering (differences in feature availability between UAs) by publishing profiles, which are collections of modules along with other non-SVG requirements on the UA (such as DOM support, scripting support, etc.).
We also had a good discussion about public accountability, including a publicly-available issue tracking system; we really need a way for both us and our commentors to know the current state of any given issue, and to make sure that it was logged and been discussed and resolved (there have been issues that got lost in the shuffle of email threads, and that needs to stop). We also resolved to create a related public FAQ/wiki, so that people can see why certain decisions were made in the spec; this wiki and tracker will also help us, what with the turnover of membership. In practice, this may have to wait until we hire our new W3C staff contact, which we hope will be within a few weeks.
Canon took us to a nice Indian restaurant for a hosted dinner, where were joined by several past members of the SVG WG: Phil Armstrong (late of Corel), Craig Northway (late of Canon), and Dean “Dino” Jackson (our recent former staff contact, who had grown a surprising head of hair… I’ve only ever seen it shaven in all the years I’ve known him). Cameron MacCormack (Batik developer extraordinaire) is also a vegetarian, so we enjoyed the meal… Indian food has a great millienium-old vegetarian tradition.
In Boston, at the WebAPI meeting, I’d taken on the task of editing the ElementTraversal specification (among other things). This is one of several scripting interfaces that were originally part of the SVG Tiny 1.2 spec, that are particularly useful for mobile implementations (SVG viewers for phones), with whom the SVG WG has been working closely over the last few years. But they are not directly related to graphics, and would be useful for desktop HTML browsers (and authors) as well, so at the encouragement of some other Working Groups, we agreed that this functionality would be handled under the aegis of the WebAPI WG. But the editor of the spec left the W3C (being a standardista burns people out at an alarming rate), and ElementTraversal was orphaned. This is bad for SVG and the mobile implementors, who have been counting on this functionality being officially specified. So after dinner, prompted by the gentle nagging of Nandini Ramani of Sun, and the assistance of Chaals (who joined us this week), I pulled a late-nighter and wrestled the simple spec into something close to publishable form, then crashed in the empty bed of the suite. (More on ElementTraversal in a later post).
We started our last day of meetings with liaisons, moved on to publication roadmaps, followed up with the Print spec, and closed the week out by working on the errata yet again. Tim “tor” Rowley and Jonathan “jwatt” Watt of the Mozilla SVG team, while not part of the SVG WG, were especially helpful in enumerating some errata items of pressing concern, so we gave those particular attention. The intent is to publish something in a couple of weeks… a subset of the whole list of things that need to go on the errata, and to add to it incrementally.
At the end of the day, Andrew Shellshear handed over the Canon membership baton to Anthony “Fat Tony” Grasso… literally. Andrew had made a mighty SVG baton out of an axe handle, replete with XML instruction elements and the new SVG logo. Andrew is moving on to other projects in Canon (hopefully still with SVG); I’ll miss him as a kindred spirit, and as a creative demiurge who really helped out the group.
We had dinner outdoors on an esplanade near the beach. Chris Lilley regaled us with some history of the Web, hypertext and HTML, the IETF and the W3C, and the work of Sir Tim Berners-Lee, Dan Connelly, and Dave Raggett… he’s even spoken to Ted Nelson a few times. I’d love to read a book that weaves the whole story of hypertext together, from teh Intartubes back to the creation of the WWW and the competitive technologies, through the work of Nelson and Vannevar Bush, back to the monks who annotated manuscripts. After dinner, we went our separate ways, and the F2F was over.
Andrew let me drive his car back to the house, as a lark. I’ve never driven in a mirror-image-driving country before (frankly, crossing the road on foot is challenging enough), but by the end of the 45 minutes, I was much more comfortable with it.