a more active thread on the same topic is here: http://news.ycombinator.com/item?id=479410

Off topic for Bespin, but on topic for their site, I just had one of those "if you want your customers to use it make it easy to play with" moments. I actually caught myself weighing the benefits of seeing it first hand vs putting my email address into another database. Then when I clicked register to fact check this comment it gave me (well, Chrome) the inferior-browser-F-U:

We would love to have you try Bespin

Unfortunately, we are using exciting new technology in HTML 5 that only leading browsers have implemented.

We want to push the Open Web forward, so for this tech preview you will have to use a new browser.

We have successfully tested Bespin on Firefox 3 and WebKit Nightly, so try one of them!

Oh well. :/

Ugh...

Canvas is fully supported in FF1.5+, Safari, Opera and Chrome out of the box. And you're only a short hack away from having Canvas emulation in IE5+. It's inexcusable not to build something like this cross-browser, when the underlying technology itself is cross-browser.

Well... it could be worse you know... you could get the superior-browser-F-U. Not long ago, my main bank's website told me to upgrade my FF install to IE6.

What would you suggest they do if your browser doesn't support the things they need it to support?

If a browser had been tested and found lacking in some type, I'd hope they'd state the feature that they were relying on that was missing or buggy. The verbiage made it sound like Chrome wasn't tested (or they're pretty arrogant in their dismissal), but something like this would be preferable:

"This probably isn't going to work. What do you want to do?"

[Let me try anyway] [Get me a tested browser] [oh, nevermind]

I understand the behavior as it stands for things like banking or medical websites, where real life things of value are at stake and non-experts may use it, but this is demonstrating an experimental widget on a "labs" site, where the whole point is to try out weird new stuff in browsers.

This wouldn't have been so noteworthy, except that Netscape/Mozilla/FF just spent a decades fighting "browser-ism" and even have a bugzilla category for it: (https://bugzilla.mozilla.org/show_bug.cgi?id=407187 as an example)

I've been using a javascript based editor for this:

http://www.heclbuilder.com

It seems to work ok, although it's not as full featured as their system.

As for me, though... they will pry Emacs out of my cold, dead hands.

I hope they've succeeded in creating a viable web-based editor, which I (and I'm sure many others here at HN) have thought about before. But it's possible they just have their heads in the clouds...

I've always thought of code editors as a bulletproof example of apps that has to be written in native code.

Very interested to see the code and how they did all this in Canvas in a sensible way.

Are canvas drawn apps going to be the launching point for Web 3.0?

I'm really amazed at the performance. This is Javascript, drawing characters pixel-by-pixel? Whoa.

Canvas is pretty fast for a lot of things. Check out Twiddla for an example of what you can pull off. It's just a bit limited in what it lets you do cross browser.

It's sort of like the days of IE4 vs NN3, with two technologies doing the same thing with a small common set of things you can do in both. Only now it's Canvas vs. SVG. You can accomplish a lot of stuff, and both can do a lot more, but there's just not enough overlap right now to really build anything truly cool.

It's not as pixel-by-pixel as you might think; canvas has a fillText() method (that is presumably implemented in native code) to draw text.