The path of least resistance is not SVG. It's a very promising standard, and programs like Inkscape have done wonders with it (and so has KDE), but in browser-land there are simpler solutions that are more widely supported.
So are you trying to say that it's easier to get IE to implement canvas? Or are you trying to say that it's not worth implementing the rest of SVG in the browsers that support it?
Yes, SVG needs to be in tools that target graphics designers, like Illustrator and Inkscape. And I guess, by your logic, those apps need to support canvas too.
Canvas is a simple API for drawing to a space on a web page. That definitely fills some needs. SVG fills some needs too. There is overlap. There's also overlap between plain text and html, nobody's saying we should get rid of one in favour of the other.
As for the size of the spec, that happens as a specification matures. Have you looked at the size of html5? SVG provides a lot of features, there are some common subsets that can be relied on today across numerous platforms. It would be nice if every implementation had all the features, but we can say the same thing about CSS or any number of other standards. Like an alpha channel in PNGs...
The viewer uses Canvas, which is pretty cool, but... if you're doing scaling and panning through a document and are okay with using new technologies then I wonder why they didn't build parts of it with SVG (since that's a way to do zooming & panning pretty naturally).
I don't have much experience with canvas yet, anyone have input?
I wouldn't call PDF an open format. It's well-documented, yes, but is controlled by Adobe.
Sorry if it sounds pedantic but they have made changes to the format over the years that harm preservation of data, like mechanisms for DRM and certain uses of a document through their reader.
I've used CakePHP (1.2 beta) and Drupal (4,5 and 6) but Drupal more recently. I like both very much. If I were building a site with the features you describe I'd go with Drupal. Either way you've got a lot of learning ahead of you but at least with Drupal, if you get it right, you'll have the benefit of piles of modules that are already tested and other people who are familiar with them. CakePHP can get you just as far but in my experience you're doing a deeper pile of custom code.
There are at least a couple of popular shopping carts for Drupal. I'm building a simple Ubercart site.
The other popular contrib modules you're almost sure to run in to are Views and CCK. You'll want to get some screencasts or podcasts to get familiar with those (Lullabot does some good ones - they're also the authors of this book).
I've been doing a lot with Views lately and a little with CCK. There's a pile of stuff you can do with very little custom code in the right places. It just takes a lot of effort to learn what works well together.
I've never seen any hints of intentional obscurity. It's just a matter of having thousands of users and contributers. Some are going to talk a lot about what they do, some expect you to step through the code if you want to figure it out. There's a lot of code and documentation to wade through and nobody to hold your hand.
It definitely is getting to be a big complicated project though. When you give it the second glance, be sure to talk about what you do - help reduce the obscurity.
The core Drupal download is good for a brochure site but it powers some massive sites. It runs my current site, integrated with Gallery 2. I used to like Wordpress but my site outgrew it.
I've been using Drupal off and on since 4.7. In version 6 I've really gone into it and am building a couple large sites with it. I'm using a pile of contrib modules and they integrate very well together. The Views module and the theming system have minimized the amount of custom PHP I have to write.
The learning curve is very steep and it has historically been hard to find the "right" pieces on drupal.org but that is improving steadily. I think the difficulty with finding the right (and current) info is a reason that people come away with the impression you've got.
I have a theory that it's impossible to prove anything, but I can't prove it.