This is an attempt to breakdown the larger SMIL problem into dependent pieces of work and also define a test which can say if this piece is done. What piece would you like to work on the most?
If you can see any gaps or better delineations, we can modify:
1. Preservation - Allow all SMIL elements and attributes to exist in loaded documents and save them back without modifications even if the object's they're tied to are modified. The test is a set of SMIL enabled animations and this might already pass.
2. Propertise - Actually load SMIL elebutes into structures and make them accessible via code. The test would be a code test, call out all the attributes and iterate over times.
3. Internal Time Line - We'd introduce the time dimension to our objects and we'd be able to set a time on them and read back nearest values (no interpolation). The test would be another code test to select objects, set a time and read back attributes. (getAttributeAtTime?)
4. Static Time - We'd allow the user to set a time via the ui or the command line and render that as a static non-animated drawing. The test would be a visual one, load inkscape with an existing smil svg file and pick a bunch of times and export each as a png frame.
5. Interpolation - This can happen along side any other pieces after 3, but it's a well understood problem and testing can be done with either a code test requesting attributes non-key-frame times or a visual test like 4 for non-key-frame times.
6. Renderer - We attempt to create a smooth animation from the attributes. This involves automatically progressing over time and changing the rendered output. The test would be to load a smil svg and see the animation animate on canvas once. (Play button/keycombo?)
7. Timeline UI - This is the most UI design aspect. We add in a way for the user to move through the timeline. We don't have to make the UI look like any existing products. The test would allow a user to drag through a timeline, press pause, start, etc and see the result.
8. Attribute modification - This is where it changes from being a viewer to an editor. Moving and changing attributes while at a specific time can create the required key frames in a linear time progression. Deleting those key times would be useful too. Test would be to create a simple animation from scratch.
9. Export - This is where we save out frames from the render one by one into a desired format. The test is simply the ability to get an mpeg or dir of pngs from an smil svg file.
These steps are all really basic and each one should be non-intrusive and capable of being released without effecting the experience for static artists.
The next question is; if we were to organise a project around these tasks; how far would we initially go and how much would each step cost?
Best Regards, Martin Owens