Ask HN: Kill it or use Flash? (don't say flash please) | Hacker News Reader

Ask HN: Kill it or use Flash? (don't say flash please)

4 pointsdaviding10 comments

I've had this weekend project for a while now that's more or less reached it's natural end. Here it is:

http://www.ZoomRead.com

One frequently asked thing I've found is that people don't want to use Silverlight with it. I've implemented a pure script version (you see it if you don't have SL) but the performance is nowhere near as good. This just now leaves flash...

I'm not keen to use ActionScript etc but not sure what else to do. So HN, is it time to kill my weekender just as a fun learning project or is it worth knuckling down and going flashy/flexy on this baby?

View on HN

So far it looks like a really cool site!

A couple of things: First off, using Silverlight or Flash will drive away the free-software-only users (though comparatively, that's a pretty small number), so if you do go with Flash, keep the JavaScript version.

Second, is there anyway you could go with Python or Perl? Those languages can usually be implemented better, they're compatible with almost all browsers (no extra plugin needed), and they're free. Java is also a good language to try, though some will disagree with me on that. :)

EDIT: cough cough Had that backwards! Sorry! :)

Other than that, great job!

Thanks!

In terms of the browser front-end it uses a mix of jscript and Silverlight. The pure jscript alternative is still quite nice, but hard to make any faster (Chrome works best). You can see it here even if you have Silverlight (Windows/Mac, with Moonlight on Linux - man, that's slow though...)

http://www.zoomread.com/ZoomReadjs.html?map=2010-10-03_04-40

The 'back-end' is all Python, as in the server generates tiles as it goes and dumps it out for the front-end to find (the browser bit).

I think with any new site it's hard unless you use completely open software with no browser dependencies, as any barrier can become a stumbling block.

Linux/OSX/iOS is (I think, happy to be wrong here) the primary desktop of a lot of leading tech people - and if they can't use my app then it's hard to reach 'mainstream' people. I think ZoomRead will fall victim to that unfortunately...

PS Had a blast doing it though! :)

Where is the bottleneck in the JS version? The major browsers are coming close to releasing GPU-accelerated canvas. You might just be a bit early...

For example: the JS version is choppy, but usable at 2560x1600 in Chrome on a PhenomIIx4. However, if I switch to the Chrome Canary build (http://tools.google.com/dlpage/chromesxs) with --enable-accelerated-2d-canvas it is significantly less choppy. At 1920x1200 it is downright snappy.

The bottleneck is mainly in rendering. Plus I actually took a lot of 'blur n zoom' effects out already (which is where the silverlight version really excels I think). I'll put them back in and retry (it's a kind of blend like anim)

I'll check out the 2d canvas - thanks for that. Ironically the IE 9 works very well too, although that group of users always seem to have Silverlight already :)

With traffic non-existent to the site it might just be too early, or perhaps some best suited as a native iOS app instead.

>In terms of the browser front-end it uses a mix of jscript and Silverlight.

I see now. I think I had things a bit backwards. :) I think JS/Flash, Silverlight, or Java may be your only choices here for now.

>Linux/OSX/iOS is (I think, happy to be wrong here) the primary desktop is a lot of leading tech people

I think your site could be very popular amongst news junkies, who primarily use Windows/Mac.

I ran it with silverlight. A couple of points:

I was using my thinkpad and could not zoom in/out (a problem with some sites) using the scroll (nib+middle button).

Could this not be done with HTML5? Canvas? I was thinking while this is kind of inefficient on my laptop it would be pretty cool on an iPad or large format touch device.

I think the technology you end up using should be something you will want to use on future projects. For example, if you endup choosing flash, will you want to keep doing flash stuff?. This fun project will serve as something to bring more attention. So what are you working on next? use it as a trampolin to build on your next skills.

My next project is more about pure HTML5, so I'm unlikely to want to keep flash around. Good advice, in that ZoomRead was good to finish up as I had enjoyed working on it - flash, not so much for me I think.

I fail at clickable title text, so:

http://www.ZoomRead.com

Your site is great, and the js version works fine for me.

Reply on news.ycombinator.com