OLPC Human Interface Guidelines

[mdlbear]

If you haven't been hiding under a rock for the last two years you've probably heard of the One Laptop per Child project. But, like me, you probably haven't heard much about the software apart from the fact that it runs on Linux. That's too bad, because it's brilliant.

Go read this blog post as an introduction, then go read the OLPC Human Interface Guidelines. They're simply brilliant. They actually use all four edges and the corners. They bring the mesh network into the interface. They get rid of the stupid "Caps Lock" key and add "View Source". Most of the software is written in Python, one of the two best teaching languages in existance at the moment (the other is Smalltalk). They have a journal, for infinite undo and version tracking.

Go read it, and ask yourself what Microsoft and Apple have been doing for the last 15 years.

Comments

[randwolf.livejournal.com]

Is there educational research behind this? Any testing? I'm starting to get the feeling of a vast educational experiment, and I don't like it at all; hope it's wrong.

[mdlbear]

I don't know the details. The project was started in the academic world -- I think at MIT -- so I wouldn't be surprised if there were some papers available on it.

The UI is available cross-platform, so I'd guess that they've been getting several months worth of user feedback by now.

[quentin-long.livejournal.com]

"[W]hat Microsoft and Apple have been doing for the last 15 years"? That's rather an unfair comparison. They are constrained by a highly relevant factor which the OLPC project can completely ignore: Namely, the need to actually make money. As well, OLPC is utterly and totally unencumbered by backward compatibility -- all those legacy issues (for hardware, software, etc) which Apple and Microsoft must address in some way.
All of which said and acknowledged, the OLPC machine looks real interesting. I wish OLPC the best of luck, and I hope it all just works...

[mdlbear]

They could have introduced a new UI at any time as an alternative -- the changes between Windows 3.1 and Windows 95 were pretty big, and could have gone farther. Apple had opportunities, too. The tight integration between the toolkit libraries and the OS in both Windows and Mac would have made it trivial to offer radically different UIs without changing the apps.

[quentin-long.livejournal.com]

Sure, they could have made arbitrarily large changes. But how much of a radical change can you get away with before people decide it's not worth bothering with, ergo don't spend money on whatever-it-is? Basically, I'm saying the whole "legacy issue" includes what people are used to. Apple definitely changed the interface when they introduced Macintosh, and subsequent events suggest that Apple's changes weren't on the "too much" side of the line; how much farther could they go before the crossed that line? The "gotta make money" thing is highly relevant here...
OLPC, contrariwise, is pretty much starting with a blank slate, since their intended "customers" have absolutely no previous experience whatsoever with computers. No legacy issues at all, whether it be hardware, software, user expectations, or whatever else.
Feel free to criticize Apple & Microsoft (the latter company in particular)for being insufficiently innovative. But when you ask "what have they been doing?", you imply that they haven't been innovative at all, and I think it would be more accurate to say that Apple, at least, has been as innovative as they thought they could get away with—or maybe a little more than that, given the market constraints they must deal with.

[mdlbear]

Apple's lack of innovation is due more to the company's UI religion -- in particular the top menubar -- than anything else. And there were lots of things they could have slipped in during the OS-X transition, like good integration with X, that would have been unnoticable to casual users but would have made a big difference in what third-party software was available. They could even have gotten rid of the stupid top menubar at that point -- they might have lost a handful of diehards, but gained a horde of Windows and Unix users who already know and prefer per-window menus.

They haven't been as innovative as they thought they could get away with in the marketplace, only as innovative as their religion will let them be.

[quentin-long.livejournal.com]

Um... you [i]are[/i] aware that Apple's UI "religion" is based on bunches and bunches of actual testing to see how real, non-geek human beings use computers, aren't you? The "stupid top menubar" makes a great deal of sense, IMAO, even if the kind of people who think that debugging code is [i]fun[/i] may disagree. I think the kindest interpretation I can put on your remarks about "religion" here is that you dislike the fact that Apple didn't design the Mac to suit [i]your[/i] particular tastes in computing machines. If that's true, fine -- you have your preferences, and that's okay. You just might want to consider that people who don't share your preferences might have valid, non-religious reasons for doing so.
As to "only as innovative as their religion will let them be", what do you think of OpenDoc?

[mdlbear]

It's also true that most -- if not all -- of that testing was done decades ago in an environment when very few people had so much as seen a graphical user interface. Things have changed. Screen sizes, for one -- all that testing was done on 640x480 screens; on large terms the distance term tends to dominate, especially if accelleration is set low or off for some reason.

There are other disadvantages to the top menu bar that may or may not outweigh the advantages (one is the inability to make good use of "focus-follows-mouse" and "raise-on-focus"); I don't think the research has been done.

I use the term "religion" advisedly -- one of my coworkers was a high-level manager at Apple not long ago; the term is even used internally among the skeptical. He was, IIRC, in charge of OpenDoc. It was an interesting experiment, but ultimately a failed one. Open document formats like XML that allow a proper separation between code and content win every time over "object oriented" formats that amount to making a particular -- usually closed -- application the "class" that has total control over that format.

[mdlbear]

It's also worth noting that Microsoft's OLE (and whatever it's evolved into) is similar to OpenDoc in spirit, and has all of the same problems.