User: Password:
|
|
Subscribe / Log in / New account

Seeking Enlightenment (The H)

Seeking Enlightenment (The H)

Posted Nov 2, 2012 15:23 UTC (Fri) by ledow (guest, #11753)
Parent article: Seeking Enlightenment (The H)

1990's - "Using OpenGL will let us use the full power of the graphics card to take away the work from the processor and let you have a faster desktop.

2000's - "OpenGL desktops can have more fancy features and we can do 3D things and shading and transparency and all sorts of cool effects."

2010's - "Using OpenGL to composite your desktop when you don't want the fancy effects that slow everything to a crawl is stupid, so we now fallback on using the CPU".

Well done, desktop guys. Nothing like being visionaries. (I'm not referring to the Enlightenment guys here!)


(Log in to post comments)

Seeking Enlightenment (The H)

Posted Nov 2, 2012 22:26 UTC (Fri) by pboddie (guest, #50784) [Link]

It's a sorry tale of how people can be enticed to make "demo-ware" which looks great when demonstrating something to other people - spinning that desktop cube makes everyone feel like Tom Cruise in "Minority Report", or whatever - and which gets the audience cheering and punching the air at an Apple-style keynote, but which adds relatively little to the experience for most users.

Back in the late 1980s, it really was a usability enhancement to be able to drag a window across a desktop and have the contents actually dragged along with the movement of the mouse, instead of the window outline being dragged and the window then being replotted (which was all the Apple Mac could really do at the time). Back then, opening or resizing a window on a fast system - not a Mac - didn't need a stupid animation of the window outline: it was impressive enough to just have a window appear or resize more or less instantly, instead of having the system stall for time.

What seems to have happened since is that many of the user interface "experts" have unquestioningly perpetuated the rituals of the supposedly most innovative environments. So that means animations of distorted, semi-transparent depictions of windows, complete with wobble effects that even Pixar would be embarrassed to use, ostensibly so that even mundane operations can seem "exciting" to people not actually doing anything more than tinker with their desktop.

I'm not arguing against things looking good, but I don't see anything like the bang for the buck that the introduction of "solid" window manipulation gave us over twenty years ago.

Seeking Enlightenment (The H)

Posted Nov 2, 2012 22:47 UTC (Fri) by johnny (guest, #10110) [Link]

Speaking of Apple, they do do a good job keeping OS X free of useless graphics effects. There was a time when I thought OS X was boring, because it didn't have any fancy spinning boxes and wobbly windows, like Linux desktops. These days, when wobbly windows are "been there, done that", I'm very happy with how conservative OS X is.

Seeking Enlightenment (The H)

Posted Nov 2, 2012 23:01 UTC (Fri) by raven667 (subscriber, #5198) [Link]

I don't think the "wobbly windows" phenomenon has anything to do with UX designers or usability testing and everything to do with developers coding the feature just because they could.

As far as new accelerated graphics capabilities enabling new UI constructs, I would say that drop shadows titlebar fading on windows are useful for subtly hinting which is the active window. A big one would be Expose-style window management which is very useful. Animations on program start/minimize and desktop change can also be useful to help leverage spatial memory which is useful to some

Many of the UI additions which are the most helpful can be the most subtle though, the flashy stuff is probably more about showing off the technology than any kind of professional UX design.

Seeking Enlightenment (The H)

Posted Nov 3, 2012 16:54 UTC (Sat) by pboddie (guest, #50784) [Link]

Your point about spatial memory is a good one, and I'm willing to concede that I'm arguing against myself in that visual feedback is a useful thing to provide to users and such feedback often needs improvements in hardware capabilities. In fact, I missed out the feature that was even more important than moving and resizing when applying "solid movement" of screen content: scrolling. Having immediate feedback when scrolling a window using a scroll bar is the difference between a usable application and a frustrating one.

I don't see the ubiquitous "cover flow" and "reflections on a shiny horizontal surface" to be visual feedback improvements in the same league, somehow.

Seeking Enlightenment (The H)

Posted Nov 5, 2012 11:07 UTC (Mon) by sebas (subscriber, #51660) [Link]

So you're essentially saying that "useless animations are useless". I agree with that. :)

Seeking Enlightenment (The H)

Posted Nov 5, 2012 12:22 UTC (Mon) by sebas (subscriber, #51660) [Link]

Correction to the above:

2010: Graphics drivers still vary in quality to the point one would call it "they suck". We need fallback paths whenever the driver or GPU is incapable of handling an operation. Yet, the fallback path can still run into performance problems. In many cases, the effects work well and bring noticeable usability improvements.

I'm not sure I'm understanding the intention behind your comment, even after reading it a few times. In general, the effects do not slow down the user experience, but it's certainly possible in some cases (usually caused by bad drivers). KWin, at least, has fallback pathes, it can use XRender (when a good XRender implementation is available) or fall back to an entirely "2D" rendering, when even XRender does not perform well enough. Others chose llvmpipe as fallback, but that is not a guarantee for good performance. In general, some effects can be offloaded to the CPU, especially for "relatively cheap" effects, that can be a good trade-off to still get usability enhancements, even with bad quality graphics drivers.


Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds