, from oldest to newest:
What, no mention of Raskin in this article? This article could have been written a couple of years ago with hardly anything different. That tree browser by Rao looks like it's never moved past the programmer-graphics stage, and it's a pain in the ass to use.
Gelernter is on to something; the "chronological presentation of everything" seems to work basically like a blog-based desktop.
Posted by Andrew @ 11/21/2001 01:03 AM pst
~~~
chronological presentation is still doomed as a desktop interface...I mean how many people are going to remember what date or time they were working on a specific document?
Posted by dan @ 11/21/2001 07:14 AM pst
~~~
yeah, that's certainly one criticism. It seems to me that G. has addressed this somewhere, but I can't remember where. I think the fact that you can sort of continuously scroll forward and back in time makes it a little easier to find something specific.
BTW, G's book "Machine Beauty" is really terriffic, although the sections where he begins to describe the chrono interface are less interesting than the rest of it.
Posted by Andrew @ 11/21/2001 08:30 AM pst
~~~
Saying "X solution for everything" limits the effectiveness of the solution by applying it in non-optimal ways, reducing the overall utility of the solution.
Chronological views are great for lots of things, but not everything.
Trying to apply them everywhere will dilute their effectiveness, and limit the uptake of a new perspective.
Allowing different access methods for the same information (remember recent conversations on facets), constraining certain types of information to certain access methods, and NOT being dogmatic over one method or another allows us to build systems that reflect the realities of postweb complexity.
If we do that, our systems might even work well for the people who use them.
Don't get me wrong, Gelertner's a terribly smart cookie. No matter how smart and innovative though, no information retrieval method works well everywhere...something I need to remind myself of in the glee of kewl new IA geek toys.
Posted by Jess @ 11/21/2001 11:48 AM pst
~~~
I agree with Andrew. This article could have been written a couple years ago...at least to me, it's all old news.
I also agree with dan...and I'll add this: a chronological desktop is still flawed in that it doesn't take advantage of our remarkable spatial memory. Humans are really good at remembering the location of things. It upsets me that more interaction designers haven't taken advantage of this population independent human ability.
Side note: I used to subscribe to Technology Review, but found it was mostly uninsightful academic whining.
Posted by Brad Lauster @ 11/21/2001 12:09 PM pst
~~~
For some people chronological interfaces are very helpful, they may not remember the date and exact time (although I know a few people that are that chronologically adept), but they do know the chronological order of things and what was prepared prior to what. They can remember what project they were working on when they read an article. I know a few detectives that are insanely chronologically oriented and that is how they classify information. This is as easy as I know many people that are insanely UNspatially oriented. It sounds like we need to diversify our group of friends and realize that people have different cognitive approaches to information and how it is organized. That is why we try to understand the user and run user tests.
Different techniques have different purposes and knowing which one is best is based on understanding the user/audience of a project.
Posted by vanderwal @ 11/21/2001 09:15 PM pst
~~~
Vanderwal seems to be contradicting himself by arguing against an interface based on location, but at the same time, saying we need to "have different cognitive approaches to information."
We're talking about "The Next Computer Interface" here, not a project targeted at a group of people. The primary (OS) interface is used by everyone who uses the computer, not just people who remember things via chronology or spatial location (I'm hesitant to use both "physical location" or "virtual location" here, but is spatial location redundant?).
I'm not saying a chronological interface is bad, I'm saying that for a primary computer interface to be good, it should take advantage of all of our abilities (including recall via chronology and location).
Posted by Brad Lauster @ 11/22/2001 11:17 AM pst
~~~
I completely agree that we need new approaches to interface. I was responding to the sweeping generalizations that get most IAs, and particularly Jakob, into much trouble. I was showing there are always exceptions to the generalizations. It is not a contradiction to point out that not everybody thinks in the same manner and we must extend the interface options, while keeping what works for some people (like the chronological approach). To not understand the approach that there are broad expanses of cognitive understanding of information presentation.
I agree that many people combine spatial relationships and chronological relationships for memory, as well as sensory memory clues. Your agreement with Dan's broad dislike for chronological relationships for retrieval seemed to put your comment in line with spatial over chronological relationships. I was having problems digesting the suggestion of throwing out what we have that works for some folks as we add other elements that expand the options we have to help make a digital information repository more usable for others.
The items in the Next Computer Interface, reminded me of the early 90s and some of the proposed interfaces based on spatial and visual interaction with digital information. There were all sorts of samples running around that were overlays to DOS. These seem like there has been more thought and refinement. I spend chunks of time at Nooface, but find most of the interfaces impractical (for myself) in their current iterations. Many remind me of old microfiche machines and having to dig out information sliding from one idea to another and across slides. The spatial interfaces like antartica are current examples that are getting to spatial relationships, but those of us that think tangentially and would like to link and store information across ideas and thought threads. These spatial relationships don't have the holes yet to step easily through to tangential information. The steps that have been taken are fantastic and should be well applauded. But, at current time they leave me feeling like I am in London and finding I have left an out of print book by Erwin Panofsky at home in the USA as I am entering the National Gallery and thinking I would love to read his the background and descriptions as I look at Early Northern Renaissance art. The spatial interfaces do not allow me to port (at least yet) easily to get that information and have it easily at hand.
My other issue with the current new interfaces is that many of the interfaces are reading or visually oriented. Again, they will improve and get there. I can not grab the information to use to help support and idea in a document I am compiling. Pulling a quote from the current new interface is relatively difficult. Many of the amazing zoom tools that are now years old that have come out of the University of Maryland labs that Ben Shneiderman has been developing, fall into a similar category. These tools allow a nice interface, but I can not yet use the information in a manner that I would like to use it.
Posted by vanderwal @ 11/22/2001 08:09 PM pst
~~~
There seems to be an assumption running through this thread: that the current desktop implementation is equivelant to spatial organization. I do not necessarily believe it is.
The current file/folder metaphor is more taxonomical than spatial. I remember where I put something because the hierarchy of folders is (presumably) arranged to make it easy to find and classify.
Spatial organization would use, I imagine, a system of landmarks, but one that isn't abstracted. The OS wouldn't be a collection of nested directories, but instead a landscape with various geographical phenomenon...
By the way, Technology Review has really evolved in the last couple of years. While interface design is a rare topic, they do look at the convergence of new technologies like biotech and infotech. They're a bit more preachy than whiny, but still one of my favorite magazines...
Posted by Dan Brown @ 11/26/2001 01:45 PM pst
~~~
