Miscellaneous and Useless Information

Human-computer interaction


It’s a geek’s dream week: CES and Macworld. There have been two announcements that have caught my attention.

[Photo of HP MediaSmart Server] The first is Microsoft’s Windows Home Server, which will be sold by HP and other vendors. It makes it easy to share files and stream video, music, and photos to PCs and Xbox 360s at home, access your files remotely, and automatically backup your data onto the server’s hard drives, which you can add or replace while the server is on. It’s aimed at homes who have more than one PC. The server does not have a display, keyboard, or mouse; you administer it on the PCs you already have via a web browser. And you wouldn’t have to do much administering at all — the user interface looks to be quite simple.

Some geeks are already proclaiming that there’s no need for Windows Home Server, since there are already open-source NAS that you can install on top of a Linux box. Give me a break: Windows Home Server is meant for people who don’t know what “NAS” stands for and don’t want to administer a Linux box. I want one.

[Photo of Apple iPhone] The second announcement is, of course, Apple’s new iPhone. It looks absolutely stunning, and the user interface looks way beyond any other phone or PDA out there. But now I have a few burning questions.

  • Can I write my own programs for the iPhone? It has some version of OS X.
  • How does the iPhone’s OS X compare to Mac’s OS X?
  • Can I upgrade the iPhone’s software like I can upgrade a Mac?
  • Can I get access to the accelerometer, proximity sensor, or ambient light sensor?

If I can write iPhone applications, then the iPhone would make a wonderful research platform for mobile devices. If not… well, it wouldn’t surprise me, since Apple has a history of being closed (see the iPod). I hope they prove me wrong.

I read a lot of news online, geek and otherwise, but lately I've been trying to read more about human-computer interaction online. There are certainly a lot of blogs and web sites on society and technology in general, but I'm thinking more about resources that focus on HCI and usability. Here's what I've found so far, in no particular order. I welcome more suggestions:

By the way, I'm not sure if it's a good idea to mix HCI entries with everything else, like entries about food, in one blog, but it seems like having two blogs wouldn't really solve the problem.

Also last month, I went to a San Francisco Bay Area ACM meeting on component-based user interface design, which turned out to be a talk on UI design patterns. Aha! It turns out that SAP has developed an extensive collection of design patterns for their domain of business applications, and they have an extensive web site for designers. The main points I got from the talk:

  • SAP has 3 types of patterns
    • patterns for frequently discovered user requirements
    • patterns for composing UIs from components
    • patterns for executable UI components
  • The only way to drive adoption of patterns is to build them into the tools that designers and developers use
  • SAP validates their patterns through user testing
  • Patterns do "freeze" UI innovation overall, but the patterns themselves will evolve over time
  • There is always a tension between those who use the patterns and those who deviate from the patterns to innovate on the UI

Time to plow through my backlog of stuff to post. Last month I went to the BayCHI meeting on Beyond Search: Social and Personal Ways of Finding Information, which was about social search and recommender systems. I got a few things out of that meeting.

How do you recommend to someone that they'll like obscure song B if they like popular song A? How does the connection between those two songs get made in the first place? The speakers from Netflix, Live365, and Pandora agreed that you need experts to make that connection, since the public at large doesn't know enough to make that connection on its own.

Also, if a person likes a popular song, you can't recommend a completely obscure song that appears unrelated even if it is. The user must be gently led down the long tail.

Finally, I started using Pandora for listening to music. I gave it one song, "Venus de Milo" by Miles Davis, and it set up a station that played that song and about 100 other songs that are similar. An instant instrumental jazz station, perfect for listening to at work.

I just came back from two intriguing talks from this month’s BayCHI meeting. The first talk was about Chandler, the open source PIM that seems to have been under development forever. Mimi Yin talked about Chandler’s design philosophy and how it’s different from typical e-mail/calendar programs (her slides are online). For example:

  • There is a universal inbox, called the Dashboard, that can hold anything: e-mail, calendar, documents, etc. These go into one of three categories: Now, Later, and Done. Stuff moved from Now to Later can be “tickled” so that it moves into Now at a specified time. The idea is that things go back and forth between Now and Later, picking up more information about how they get done, until they are actually Done.
  • Stuff can go anywhere. An e-mail message can go directly into the calendar or a to-do list, and it also stays in your Dashboard.
  • Tags are used for bottom-up organizing (so that you can find it later), while categories are used for top-down organizing (putting stuff in collections). They have somewhat different affordances, but tags can easily become categories and vice-versa.

This all sounds good, but I asked how much of this was driven by user observations. Mimi said the biggest source came from looking at people’s e-mail folders to try to figure out what their organizational schemes were. So I’m still not sure how much of Chandler’s design is driven by what people actually do versus the Chandler team guessing. I hope it’s more the former.

Chandler is particularly interesting to me because it’s trying to address many of the same issues as the IBM research project I’m in, Unified Activity Management.

The second half was an absolutely hilarious talk by Merlin Mann about modern life in general and dealing with the deluge of information. In fact, he manages a whole web site about this problem called 43 Folders. One organizational framework that he discussed in particular is called Getting Things Done (which Mimi also touched on in her talk). Instead of rehashing what Merlin said, take a look at his intro. Suffice to say that geeks seem to have gravitated to it, so I’ll have to take a look.

I just got back from a BayCHI talk by Jensen Harris, the lead designer of the new Microsoft Office 12 user interface. He’s actually already blogged a lot of what he talked about, so I won’t repeat it here — take a look at his “Best Of” list on his blog for an overview. Instead, here are some high-level impressions.

  • Jensen is an excellent speaker. He’s clear, funny, and not afraid to poke fun at Microsoft’s previous attempts at “improving” the user interface of Office.
  • UI designs are actually driven by data that is collected anonymously, and with permission, from current Office 2003 users. It may not be perfect data, but it’s a lot better than guessing.
  • The new UI is most easily adopted by novice users. Power users already know Office well, so they have the most to loseWeight Exercise in the new UI. The biggest thing Jensen feels the new UI is lacking is customizability for power users.
  • Office 12 is about a year away from release, but Microsoft is already talking very openly about it. They haven’t been nearly this open in the past, and other companies certainly aren’t as open today. I believe it’s from a combination of starting from a position of strength and feeling the pressure to show that they are innovative, that the next version of Office really is worth buying.

I mentioned last month that Microsoft is overhauling the user interface of the next version of Office. Now member of the Office user experience team, Jensen Harris, has a blog all about Office’s UI, both its past and its future. It contains some good insight into how Office’s UI has evolved, how they are designing the new UI, and what issues they’ve already run into while testing it.

From the New York Times:

A ubiquitous city is where all major information systems (residential, medical, business, governmental and the like) share data, and computers are built into the houses, streets and office buildings. New Songdo, located on a man-made island of nearly 1,500 acres off the Incheon coast about 40 miles from Seoul, is rising from the ground up as a U-city.

Sounds like a ubicomp researcher’s dream.

A constant theme in HCI and computer-supported cooperative work (CSCW) is how hard it is to get a group of people to adopt and support new technology — being technically superior is not nearly enough. One good example, interestingly enough, is in the area of web application frameworks.

For example, in my favorite language Python, there are a bajillion web frameworks but no clear leader. Meanwhile, one open source web application framework written in Ruby has been getting a lot of buzz lately. The users of Ruby on Rails rave about its ease of use and simplicity. This is causing some consternation in the Python community, to the point that some are creating a clone of Rails in Python called Subway.

But simply copying the framework itself won’t generate the same excitement. There are other reasons Ruby on Rails is getting so much attention. The creators of RoR have evangelized it very effectively and have created good documentation so that others can jump in. Also, the library was extracted from an existing application, so the creators had a real sense for what was useful and what wasn’t. It reminds me of Papier-Mâché, a toolkit for creating tangible user interfaces by Scott Klemmer, who designed and implemented two tangible UI apps before designing the toolkit.

However, all is not lost in Pythonland. Thanks to Robert Scoble I just found out about a Python web framework that seems to hold great promise. Django was also extracted from an existing application, and its creators seem to understand the importance of good marketing and documentation. I’m hoping it succeeds.

The two features I find most interesting about Mac OS X 10.4 (“Tiger”) are not the ones getting the biggest hoopla, Spotlight and Dashboard. I’m more intrigued by improvements aimed at programmers. Core Data helps the developer manage the data within an application. Core Data, along with Cocoa Bindings, promises to make it much easier to write applications using the flexible model-view-controller pattern. Having just written a good-sized MVC app myself, I appreciate anything that makes it easier. The other feature I’m keeping an eye on is Automator, which allows end users to create their own scripts to automate tasks. Anything from renaming a bunch of files to rotating, cropping, and e-mailing a collection of photos is possible.

There has been a real lack of end-user programming tools built into operating systems lately; the last one was HyperCard, which was last bundled with a Mac in 1990. Windows had the primitive Macro Recorder, which simply recorded raw keystrokes and mouse events, and even that was removed from Windows when Windows 95 was released. So Automator is a welcome change. In some ways, Microsoft is moving in the other direction. It plans to include a new command-line shell and scripting language with Longhorn called MSH. That’s good for sysadmins, but doesn’t do much for end users.

« Previous Page