I am just about to start a new package for dotfolio. The package will be used to create concept maps for showcasing learning artifacts stored in the e-portfolio system. I have chosen concept maps as a mechanism for showcasing, as it provides context for the learning artifacts. This differs to OSPI, as they use a matrix for showcasing.
I want the concept maps application to be interactive, and have looked at using Laszlo. However, after listening to Lars from Google talk about Google Maps at a conference last week, I have been inspired to look at other approaches to developing a Rich Internet Application.
I would like to use Laszlo, as it compiles down to a flash object, which can then be played in any browser with a flash plugin. This kind of solves my concerns for cross-browser support. However it means mixing Laszlo code in OpenACS, meaning there is a dependency on using the Laszlo Presentation Server to compile the Laszlo source code before deploying the application. This adds a layer of complexity due to a dependency on third party software.
The alternative is to use JavaScript and SVG. There are some interesting things that can already be done in SVG and JavaScript as shown here. The advantage using this approach is that there is no dependency on a third party system. The disadvantage is that SVG isn't widely supported yet. However, Lars from Google mentioned that SVG will be supported in a future release of Firefox. So hopefully his insight into the future of web browsers rings true.
If I chose the JavaScript/SVG approach, then users will have to download the Adobe SVG viewer plugin. This would be preferred over using one of the nightly mozilla builds that include SVG. Creating a dependency on using the Adobe SVG viewer plugin would mean even IE users can view the application. Or I could implement a Factory pattern to create VML for IE browsers, and SVG for Mozilla browsers. For the timebeing I would prefer to focus on SVG though.
What do people think? I'm interested in your opinions.
