Date: Sat, 16 Mar 2002 13:11:48 -0800 From: Terry Lambert <tlambert2@mindspring.com> To: Chip Morton <tech_info@threespace.com> Cc: FreeBSD Chat <chat@freebsd.org> Subject: Re: Free BSD Message-ID: <3C93B514.9AB4BB7E@mindspring.com> References: <4.3.2.7.2.20020315181331.01b26160@threespace.com> <20020314204235.L152-100000@pogo.caustic.org> <15505.28725.937368.158235@guru.mired.org> <20020314204235.L152-100000@pogo.caustic.org> <4.3.2.7.2.20020315181331.01b26160@threespace.com> <4.3.2.7.2.20020315190230.01b2a4f8@threespace.com> <4.3.2.7.2.20020316100234.01b21638@threespace.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Chip Morton wrote: [ Jef Raskin ] > Oh I read the article, all right. And I disagree with most of it. And the > only thing that I've "suggested" here is that he ante up and actually > *implement* some of his ideas. I really take issue with his blanket > assertion that (paraphrasing here) "All windowing systems/GUIs are screwed > up, and I could do it so much better--but I won't." First of all, he specifically asked for programmers to contact him, as part of the article, so the "but I won't" doesn't hold water. Second, his design philosophy is that you build the interface, and the build the underlying system to support it. In this philosophy, it's impossible to glue a UI on an existing system and meet the design goal, unless you design the UI, and then go looking for a system that matches it and can be forced into supporting it, and/or build a new system from scratch. > And let's be real here about the rest of your/Raskin's points. All > windowing systems with significant market share operate under a de facto > standard. Regardless of whether you think that standard is the best or > not, most people use WMs that are functionally very similar. And switching > from one to the other doesn't require a degree in astrophysics to figure out. He specifically stated that the learning curve was not the porblem, it was the fact that the differences were in your face to the point that they required conscious acknowledgement in the decision making process. The point of interfaces is to abstract complexity, so that you can concentrate on tasks. Any time you concentrate on the interface itself, you detract from the attention that can be committed to the task. A good analogy here is the leaky vs. non-leak boat. If I am in a boat, and am there for the task of transportation, I concentrate on getting from point A to point B. On the other hand, if the boat has a slow leak, I am distracted from the task for which the boat was designed -- getting from one place to another -- and caught up with burning at least some of my time bailing water, rather than the task for which I'm using a boat in the first place. > Now if Raskin wants to change that de facto standard, he's essentially > going *against* the standardization that he argues for. The > standardization of the interface is essentially done. Let's leave it be > and move on to bigger issues. I guess this means you run Fvwm95, and every place it does not match the standard of Windows, you chip away at it until it more closely resembles Windows. Or this is your strawman argument, to argue against the merits of standardization, by implying that a bad standard is worse than no standard. The answer is that this is not true. A bad standard is in fact better than no standard. A standard, no matter how poor, means that I can save the training costs for a new employee, or that it's possible to hire a temporary worker, and have them be productive within the window of temporary need. I think Jef is arguing for good standards, more than poor standards, but primarily for standards over none. > I read an article in which an engineer at one of the major American car > manufacturers said that response time could be improved by placing > braking/acceleration controls on the steering wheel. He argued that the > time required for the driver to move his feet was significant in decisions > requiring split-second responses. But you know how many car makers were > interested in adopting his idea? Zero! Because nobody wants to relearn > how to use his hands rather than his feet to drive the car. I sure > wouldn't rent/buy one. THere are a couple of assumptions implicit in this example: 1) The engineer was right, not wrong 2) That improved response time is desirable 3) That it's impossible to implement dual controls 4) That the base cause was an unwillingness for people to relearn things #1 is an urban legend, without references #2 is arguable, from a safety standpoint (one has to wonder how many people go to step on peddles, only to reconsider before actually doing so, or how many accidents are caused because the car in front has anti-lock brakes, and the car behind has ordinary brakes) #3 is both an economic and a human factors issue; the cost of dual controls is going to be higher, but perhaps not very much so, if the controls are electronic, since a fly-by-wire system doesn't really care where the inputs to the wire are from, so floor and wheel controls become the cost of the switch components themselves. #4 could be completely wrong. It could be that humans are just hard-wired incorrectly to be able to deal with the controls on the whel. We don't know *what* process was behind the decision... assuming there even was one, and we aren't talking urban legend, here. > The point is that once the population at large has settled on a particular > method--whether through conscious decision-making or lack thereof--getting > them all to switch to another method, even a better one, is damn near > impossible. Believe me that if the music industry hadn't shoved CDs down > our collective throats, some of us would still be listening to cassette > tapes or 8-tracks. I guess that explains why most of the world, besides the US, hasn't been able to switch to MKS units? Or why the Euro never happened? There are tons of counter examples. Frankly, the idea that a switch can't occur is wrong in so many ways, it's hard to pick only a few of them to list. > The actual look of a window manager (or car, or woman, or anything else) > only matters very early up front. You may be wowed by the look of the > windows and widgets early on, but after that it really doesn't matter to > you while you're working. Sure, I don't like the look of twm, but I would > be no more/less productive by using it. In fact, I might argue that the > pleasure I get out of having an attractive, colorful windowing system with > my girlfriend on the wallpaper would actually make me more productive on > the whole. Productivity isn't just about the milliseconds saved in > dragging the mouse from one corner to the next. For large purchase decisions unlikely to be repeated, it's a motivating factor, and it's an important part of the consideration in deciding on "the whole package". If it weren't, people would not pay to run advertising. Your argument about your wallpaper is incorrect, since we are not talking about machines limited to a single user, we are talking about transportation of productive skills between machines. This could be hindered for you by the lack of the background you are used to seeing during your operation of a machine. Or it could be hindered by the presence of the picture of your girlfriend creating a hostile work environment for someone else who is looking over your shoulder in order to collaborate with you in a particular problem. It's a matter of perspective: "one man's mead is another man's poison". You seem upset that someone else might be able to dictate your UI. The answer is that it's not dictation, it's a consensus. The government you live under right now is consensual; not everyone can install the "society manager" of their choice, or an entirely different "society system", if the current one doesn't suit them. Basically, you are arguing that anarchy pleases you. 8-). > I really think that the shortcomings in Raskin's arguments become apparent > when he mentions the ridiculous prospect of using a wallpaper that looks > like open windows. Who the hell would do something that stupid? Besides > him, I guess. Someone who likes to screw with people would do it. Someone who doesn't want someone else to be able to casually look at their screen and see what's really going on there. Someone with poor taste in backgrounds. Someone who doesn't have a copy of that picture of your girlfriend. THere are a lot of bad background choices; the argument is called a "reductio ad absurdum". > Because frankly, I think that he deserved to get his ass > beat for changing somebody's color scheme to red on red and forcing them to > have to reinstall it. If that's his best argument for why we don't need > color in the GUI, then I rest my case. No, that was also a reductio ad absurdum argument. Consider a public library computer. What harm is it, if someone sets the color scheme to read and green during the Christmas season? None, I guess, unless you happen to be red-green color blind. The red-on-red example, though, is more of an example of a customizable interface that permits stupid choices, as well as permitting good ones. It's pretty obvious that it should not have permitted indiscernable choices for colors that are required to be contrasting for the UI to be usable. Even if you agree with none of his other premises, you have to agree that an interface that permits converting the product from its intended use to a doorstop is a poorly designed interface. > Like I said, he can develop his 1-bit WM and then he can have it. It's not about monochrome. Just because he designed the most critically acclaimed user interface of all time on a machine incapable of displaying color, doesn't mean that he would have limited himself to monochrome, had the choice been available. In fact, we see that he didn't; when the first color macintosh came out, the background was blue. In any case, developing a new window manager would not serve to solve the problem he's pointing at, any more than denying the existance of the problem solves it. -- Terry To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3C93B514.9AB4BB7E>