Andreas' UI and design blog

Wednesday, March 08, 2006


I just found this on boingboing and I couldn't stop laughing for several minutes... only at the media lab... :)

Just imagine a blender controlled by voice. Here is a video. And here is some text describing it.

Monday, January 30, 2006

Fascinating blog about work on the new MS Office UI

I just listened to a talk about work on the new UI for Office (called Office 12). During this talk the speaker gave a pointer to Jensen Harris' blog on the new Office UI and it is quite a fascinating read and also the comments to the blog entries are definitely worth checking out.

Saturday, January 28, 2006

Approximate watch

I just heard of the Talus watch on boingboing ( It's an interesting concept for a watch because typically watches are trying to be exact. Well, this one is not. A couple of years ago I heard a talk about time pieces and the one thing that I still remember of that talk is how people use analog and digital watches differently: A digital watch gives you *exact* time, what time is it now. The analog is not quite as exact typically, but not because it couldn't be. Rather you don't look so closely! And here is maybe the main difference in use as well.

The speaker gave a fascinating example, a simple experiment which I have tried a couple of times (successfully). If you see somebody with an analog watch look at the watch, walk up to him/her and ask him for the time. S/he will most likely have to look at the watch again. If the person wears a digital, the most likely can give you the time. Now why is that? It's because analog watch users actually don't look at the *time* but very frequently they look for another piece of information derived from time, for example: how many minutes do I have till X, how long am I waiting here already, etc. The difference to looking for the time is that they are looking for an interval, not for a time point. So if they have just looked at the watch to and found out they still have -- say -- 5 minutes till the bus leaves, they may not know the exact time. They could probably figure it out, but looking at the watch again is just easier.

So where does this discussion lead and what does it have to do with the Talus watch? It seems that watch is sort of an odd case in between the analog and digital watch. It doesn't give you the correct time, but an approximation. For instance it tells you "a bit after 5" instead of "5:03", or "nearly 1 thirty" for "1:27". But it shows it digitally, so it's not an analog watch either. And the watch has a special mode to show the exact time too.

Now I wish I'd run into somebody who has one so I can do my little experiment with them and figure out what would happen.

Oh, here is a page about that fascinating time piece:

Wednesday, December 28, 2005

Remote control without batteries

A friend send me a pointer to this page: The 50 Greatest Gadgets of the Past 50 Years.

On place 21 (right on top of the page this link should take you to) is the "Zenith Space Command" form 1956, the first remote control. Of course the first (wireless) remote control is a totally cool gadget. But there is something that I found especially amazing about this piece: it doesn't need batteries!!

The way this is done (according to that page) is through sound. The space command has 4 buttons and each makes a small hammer strike a piece of metal which causes some ultra sound which the TV then picks up. Simple, elegant. You probably don't even have to point the remote at the TV for it to work. And it doesn't need friggin' batteries!

I'm an environmentally conscious person and I think we are way too battery happy these days. In fact you seemingly cannot buy almost anything (except maybe a hammer) that doesn't require some kind of battery. And I ask: is this really necessary? Even modern batteries are a hazard in the waste, and a battery is really a lot of valuable resources simply discarded after getting very little energy for a lot of money. Can't we finally find a better solution? Couldn't we build systems that are less reliable on batteries? (in another, forthcoming post I'll talk about batteries again)

As the space command shows it clearly is possible to build - for instance - a wireless remote without batteries. Yes it had only 4 functions and we probably don't want to have a brick with 50 different hammers in it to make a modern remote. But electronics has made so much progress since then, I cannot believe there isn't an alternative to batteries for many applications.

Maybe I'm totally day-dreaming now, but maybe the energy you get from pressing a piezo-electric button might be enough to power an IR LED, or there is some other effect that doesn't require a stong power source such as a battery? I'm not an electronics expert so somebody else will have to figure this out. But let's start thinking about alternative solutions and maybe we can come up with something and reduce the number of batteries we have to buy and eventually throw away (because I assume most batteries never see a recycling plant anyway, at least not in the US).

Saturday, December 17, 2005

Web 2.0 explained

As the Web 2.0 buzzword gets used more and more one wonders... what IS is exactly? Who better to write an article about Web 2.0 than Tim O'Reilly. I happened to stumble over this article more or less by accident and I'm glad I did. It's a lengthy read but it's absolutely worth spending the time.

As I'm just finishing up writing a book chapter with the title "From Social Navigation to the Social Web" I couldn't have found this article at a better time. (The chapter will be in a book on the Adaptive Web published by Springer some time in 2006)

Thursday, December 08, 2005

Why can't I light my personal flame thrower?

I recently purchased a personal flame thrower - aka Professional Cooking Torch. Or, as most of us would call it: a creme brulee doohickey. Essentially a little butane torch. These things - especially the larger ones -- have pretty hot flame so you want to make sure there is some safety mechanism on it. Especially as there is also a lock, so that the torch continues burning even after you take your thumb off the trigger... Make it child proof, so to speak. Of course UI people know that the plain english translation of child proof is: "nobody above 6 years of age can figure out how to operate that darn thing" This cooking torch is no exception. Although I have it for over a week and have used it several times I still need at least 5 tries to light it AND engage the lock.

The basic layout of the device is pretty simple. There are two slider switches, one on the left and one on the right. There is also a trigger button at the back and below it a little safety latch. The slider on the right side controls the size of the flame -- simple enough. This means the thing really has only 3 controls to worry about. Yet, despite this simplicity it's obviously hard enough to consistently turn this thing on and lock the flame.

Part of the problem seems to be the lavish labeling of the controls (read: non-existing). The right level, the one the controls the flame size is soooort of labeled with + and - (barely visible). The other controls have no labels whatever. Essentially it should be a simple two button operation: you need to press and lock the safety latch and then press the ignite button (the big red button at the back).
Well not so fast. I press the safety down it stays down. Then press the ignite button. OK. Flame comes on. Release. Flame goes out. Something makes click. So far so good. Next iteration: I press the safety again, it locks, I press the ignite. Nothing happens. Hmmm. Press safety, it doesn't lock, press ignite. Nothing happens. Once more. Press safety, it locks, ignite, it does ignite. Press. Safety, lock, ignite, it does. Again, Safety, doesn't lock, ignite, nothing happens. *sigh*

Read manual. That doesn't help much (of course not. These manuals never help, but it tells me how to make Chocolate Creme Brule, if only I could reliably turn this thing on). So I play some more with the lock lever (on the left) which makes the whole procedure more complicated, because now the torch could be in a "mode" where it simply cannot be turned on. Or at least I thought so because the thing never lit whenever the lever was in the back position. Eventually I figured out there is actually no mode and there is no connection between the lock and the inability to light.

This should be simple, but it isn't. After way too much experimenting I think the safety latch doesn't always properly release when you extinguish the thing. Waiiiit... isn't the safety latch the ONE piece on a little flame thrower in your kitchen that should be the most reliable piece? Aside maybe from the gas tank itself? Of course I'd rather have it not come on when I want to than the other way round. But why does a simple cooking implement have to make me feel stupid? And then it turns out it's probably not even the design itself that's so bad, but maybe bad workmanship? Again... do I want to have a potentially unreliable butane torch sitting around in my kitchen?

But as I'm a geek I will not admit defeat. I will practice till have mastered "Le Torch". After all I don't want to feel stupid when I use it and friends are around. Or maybe I'll ask a friend to do the torching and gleefully observe how s/he cannot get it to light... this puts the whole torch into a new light. It's not a creme brule torch at all... it's an entertainment device!

PS: As I'm bitching already... they could really be more explicit about how to gas up the torch as well. Essentially you connect a gas cartridge to a valve at the bottom and press down. Gas flows from the one canister to the other. Fine. But how long should you fill it? How do you recognize when it's enough? The manual only says "don't overfill" and that a backspray of liquid gas from filler valve indicates that the torch is full. Helloooo? Back spray of liquid gas? How about some information how to avoid back spray!? I don't think I want any backspray of liquid gas in my kitchen. A simple "this takes about 15 seconds" or something along those lines would help tremendously here.

Friday, December 02, 2005


A recent article on the ipodobserver shows that again that even very elegant designs can have their issues. The article reports about concerns about RSI (repetitive stress injuries) from using the iPod click wheel. Well, my first thought was just to laugh, but of course this concern does make a lot of sense if you really use the spin wheel a lot. Although I have a very long list of - say - artists on my own iPod I would probably very rarely spin the wheel so much that it becomes an issue, there are probably people who use the thing very differently than I do.

The lesson to take away from this is a reminder that you just have to observe users, not think what you believe users are doing. And I know I am occasionally guilty of that (aren't we all when there are close deadlines...)

Wednesday, November 23, 2005

Spell something with Flickr

Here is a reeeeeally cool (and useless?) Flickr mashup (hmmm - does this actually qualify as a mashup?) I found this when I surfed to Julie's Jumble which I found because Julie left a comment on by blog (thanks Julie :) )

It is Words from FlickR Just enter some expression and you get it spelled out using images from Flickr. Here are a few examples:

Now, how cool is that?

Cool: Speculative ligatures for internet expressions

Here is a neat little page for people who care about typography. Imagine abbreviations like lol, wtf etc would have evolved before there were computers? How would they change over the years, centuries... what would they look like?

Sunday, November 20, 2005

Driving into a parking garage

Oh my, just created the blog and I have something to bitch about already? Geez...!

Yesterday I drove into a parking garage as I have done hundreds (thousands?) of times before. I stop at the gate, open my window to get the parking ticket and hear the voide "push the button for the ticket" or something along those lines. And suddenly a thought struck me: why in the world do I have to push a button to get this ticket? Obviously the machine knows I am here, otherwise I wouldn't have gotten that voice prompt, right? And the button was even blinking, if I remember correctly.

Now this IS interesting. Why indeed do I have to push this button? Couldn't the machine simply push out the ticket and give me the voice prompt to please take the ticket? I am pretty sure I have once been at a garage where things indeed worked like that: you drive to the gate, the ticket gets stuck out at you and you get a voice prompt. So why not here?

Let's step back for a moment and assess the more complete situation. That particular garage entrance was right off a street. You turn off the street and there you are at the gate and get the voice prompt. This is different from some garages where you might have gone through some driving to get to the gate. I'm thinking of garages where you drive down a dark alley to get to the gate or where you drive down/up a ramp. Could this make a difference?

Well, I think it could indeed. Because if you turn off the street and there you are, especially in a place where there is maybe not much space to stop for a moment, it could happen that somebody stops at that entrance for a moment to take a phone call or whatever. And as the gate is right off the street, this would trigger spitting out a ticket.

Big deal? Well, maybe! Let's assume somebody stopped for a moment just to take a cell phone call. Voice prompt happens, ticket gets pushed out, nobody takes the ticket, gate stays closed (we have to assume the gate opens only when the ticket is taken, because you want to ensure that whoever went into the garage got their ticket) car drives away again. Gate still closed, ticket still sticking out of the machine. What now?

Does the machine pull the ticket back in? Probably not. Does the machine just spit out the ticket so it drops to the floor? That could be risky because somebody arriving later might drop their ticket and then they could pick up the wrong one (very bad). Just leave the ticket in the machine? Not good either because the next poor fellow that comes to the garage now has a ticket with - say - 10 minutes of parking time on the ticket already (unless the system is smart to actually record the time the gate really openend to let the holder of that particular ticket into the garage. Or somebody would have to talk to the machine to collect the bad ticket (which is expensive with today's cost of personnel)

So maybe in this particular situation it isn't quite so stupid after all, to give the voice prompt and ask the driver to push the button. But, only in a situation like this one. I don't think it makes any sense to ask people to push that button if we can be certain that whatever car drives up to that gate indeed wants to get through the gate.