The new app was originally written with d3 but, by the time of deployment d3 was responsible only for scaling the axes, and the rest of the graphic was built largely around custom code written around canvas elements.
Ah, I see in carbocation's link now. It just seems odd to say they moved from d3/svg to canvas, because d3 is just a layer on top of canvas (or the dom), and if you look at their source (not chart.js, the minimized source on the website) it's full of calls to d3 still. Maybe more useful to OP to say that it was originally written in d3, then optimized.
I agree with being at the point where I'd happily pay for it (I'd actually pay for a mobile and faster desktop version that sync, as it's a bit sluggish on my netbook). What Joel said about it being a list of lists (a very simple idea), is what I like so much about it. Currently I'm using it for standard simple use cases (keeping track of stuff I need to do), non-standard use cases (tracking my weightlifting progress, between each set), and collaboration (my landlord and I keep a list of things to do). Oddly enough, I don't use it for software at all (what I investigated it for in the first place).
I introduced my (non-technical) girlfriend to it, and after a short explanation of how I used it, she took to it immediately, and uses it daily to keep track of her to-do list.
Yeah, it's kind of a mystery to me why my fiance doesn't like it. I suspect it has more to do with fearing the unknown when it comes to new software over actual usability concerns.
The way I see it, she really likes wedding planning, so I'm not going to rock the boat too much by pushing Trello, haha.
I think some of the examples he gives at the end of the article are good in winning confidence, such as "we hate spam as much as you do" and "you can unsubscribe at any time". These things are hard to communicate through interface alone.
As for the Paypal thing, I hear what you're saying, but I think the microcopy should stay. It does no harm, fits just fine into the interface, and if it sways even one person who blanched when they saw the Paypal logo (next to the credit cards or otherwise), it was worth it.
Like vimperator? I use it, and I love it. In fact, I'm thinking about moving to a windows manager such as awesome or Xmonad for a completely mouseless experience.
I fail how to see how using HTML and Javascript keep something from being an application. To be an application it has to run on the desktop, and not the browser?
It's about level of interactivity, responsiveness, and robustness. Shiny web front-ends to a relational database just don't compete on the same level as desktop workhorses. Mint, Gmail, and Freshbooks are nice web "applications", but their innovation is in solving simple problems with incredible interfaces. The innovation in Textmate, and Photoshop is that they enable a high amount of manipulation and automation that simply can't be delivered over the web at this time.
To be an application, you certainly don't need to meet the levels of interactivity, responsiveness, or robustness as Textmate or Photoshop. Maybe to you that's what an application is, but under a more standard definition all a piece of software needs to do to be considered an application is to interact with a user to do a specific task. It has nothing to do with value judgements. A crappy cgi guestbook is technically a web application. It doesn't matter that it's slow and only does one thing.
I don't know whether geekcorps is still active, however, you should be more lucky with the Lion's Club International, which organizes similar projects : http://www.lionsclubs.org/
