This is so awesome that we get to reimplement the same graphics effects we did on first PCs from the mid 80s, but this time - fully loading CPUs on machines that would have been cosidered supercomputers back then!
It's called progress. And because of it, we programmers will never be out of work.
Although you seem to mostly be talking about new platforms, I'd add another reason for this perpetual reimplementation: the amazing lack of knowledge transfer from one generation of programmers to the next. Nearly everybody seems to start from scratch, and of course we all think our ideas and programs are original and far superior to what came earlier.
I've often wondered about this. Part of it may be that the real practice of programming is totally divorced from the schools that (supposedly) teach it. Another part is obviously that the quick succession of languages and platforms reduces earlier work to the equivalent of ancient indecipherable tablets (if it's even available). On that last point, one of the unexpected pleasures of working in Common Lisp is being able to easily read 20-year old code. It turns out they weren't as stupid back then as we might think!
Great point about the "undecipherable tables" that any program eventually becomes as languages, platforms and OSes change.
Beyond PR and vested interests, the full backward compatilibity of java (even when it hurts) and of Windows (up to XP) account for their survival a lot more than anyone would ever suspect at first thought.
This bit is quite funny:
Function to work out approximate width of a word based on Arial font letter widths.
Probably possible to do by putting the word in an offscreen div with style set so it's only as wide as content, then checking the size of the div...
function fontWidth(word){
// arial font
var S = "a68b68c68d68e68f31g68h68i31j31k68l31m108\
n68o68p68q68r39s68t39u68v50w90x68y68z68A\
90B90C90D90E90F78G98H90I28J59K90L68M108N\
90O98P90Q98R90S90T68U90V90W130X68Y90Z68'\
30.42!29?68-41/41=72";
var l = word.length;
var s = 0;
var x = 0;
var c = 0;
for(i=0;i<l;i++){
c = word.charAt(i);
x = S.indexOf(c);
if(c>="0" && c<="9")s+=69;
else
if(x>=0)s+=parseInt(S.substring(x+1,x+4));
else s+=68;
}
return s/10;
}
It's funny, now that Javascript interpreters are getting fast and decent effects can be written in Javascript you start seeing lots pages like this.
I say it's funny because it reminds me of the explosion of similar animations that were created in Flash a few years around the time that Flash 5 and 6 came out. Animations that were mostly cool and revolutionary-seeming because of the technology rather than the actual content.
I'm not knocking this type of creation--I was really fun time to be working in Flash with so many 'breakthroughs.' Now it's cool to see a similar surge in creativity happening in a language that is even more ubiquitous than ActionScript.
My thoughts exactly, I was also involved in Flash experiments from the days of Flash 4 even, when it was horrendous to "code" in.
I can't help but find it slightly irritating when people remake these in Javascript these days to have praise heaped upon them for interesting new experiments. ;)
What browser are you using? If you try the latest Safari or Firefox nightly, you'll be amazed how fast that thing is. Javascript VMs are progressing at an amazing rate.
It may be premature to say that it's neat just for the sake of neatness. It may not have a direct UI usage, but we could probably learn something about mouse events and drawing by examining the code, for example. Also, from a UI perspective, you migh be able to make mousing more efficient by dragging UI elements around in a cloud around the mouse. (You could fade them out so that they were just barely visible, for example, until you right-clicked or shift-right-clicked to pin them down and use them.) It looked, at first glance, like the effect was running on a Canvas, so I don't know how easy it would be to do outside of one.
Also, call it mkn's first law: There's always something you're overlooking. (Where 'you' is the hypothetical you, not you in particular.) :o)
We've all heard of single serving sites: http://news.ycombinator.com/item?id=399941. Here the parent commenter is pioneering the latest incarnation of the trend: the single serving HN account.
I guess my point was that it would be a translucent cloud of items that followed the mouse. This way, you'd be able to see/read what was under the cloud but could switch your attention to the translucent cloud to see what was there.
Context menus are invisible until you right-click. Also, there's been some evidence that we remember the locations of things better if they're arranged radially (like the demo approximates) rather than orthogonally (like in a context menu). A combination of a translucent constant reminders of tasks available plus a radial arrangement of those tasks floating around the mouse might make it easy to remember what's available after a right-click. It might be distracting as hell, for all I know.
One thing's for sure, it would be more than just a context menu as we know them now.
The effect has to be adjusted slightly to be used in a UI context. If it is changed so that the item nearest the mouse cursor snaps to the cursor, then suddenly you'd be able to precisely and easily select an item from a group of tightly-packed items. So the UI advantage might be that you can see a lot of items, but can precisely select any of them.
Also, for UI purposes, it might be beneficial to remove the "springy" effect on the items. When the mouse isn't moving, the items shouldn't be animating. In other words, the animation should be driven by mouse movement, with no elasticity.
It's called progress. And because of it, we programmers will never be out of work.