The Design of Software (CLOSED)

A public forum for discussing the design of software, from the user interface to the code architecture. Now closed.

The "Design of Software" discussion group has been merged with the main Joel on Software discussion group.

The archives will remain online indefinitely.

Great design - should it comply with user expectations?

That's great question I think. Joels generally writes and I fully agree that in almost all cases we should keep up with users expectations. In the other words: the user interface model should comply with the typical users model. Usually changing this will lead to the users confusion. Such UI will not be easy to use, becouse a user expect something different.

But what if we have better solution than this commonly used? Should we just forget about that or try to force users to use it anyway and fear that it will prevent the product from selling?

The good expmple I know of that problem are washing mashines in US. I'm living in Europe, but as far as I know most of washing mashines in US have a cover opening up. And there is a drum inside, which revolves on a horizontal plane. There is also a stick in the middle with several pins which help to move clothes. Actually they do, but also they tend to destroy them as well.

On the other hand we have quite different washing mashines in Europe. They have a hatch on the side and the drum revolves on a vertical plane. The drum is profiled to wash clothes well without destroing them. I live in Poland and we have been using US style washing mashines during old 70s and communizm, when we had no other choice. There were actually one washing mashine model and your only choice was: you buy it or not (if you were lucky enough to be in queque short enough for you to get it).

This I think illustrates the problem. If you want to sell washing mashines in US, you have to offer those old style ones, which destroy clothes, because people are used to them. If you would like to sell only the better european ones, then you are most likely going out of business.

Is that the same with the software? I think that it depends. But it's easier to include just one new revolutionary thing in the software, than propose a software that does everything different.

The good software example is C1 compared to Photoshop in terms of processing many digital photographs. C1 has a completely new UI model - it allows to batch everything and this can be easily done. Photoshop forces you to process photos one by one, which can take forever if you have couple of hundreds of them. But I've heard from many people that C1 has strange UI, which they don't understand and they use Photoshop. C1 can save hours, but it's different and that's why many people will choose PS. But it changes - C1 interface is more and more popular and now the same idea is built in RawShoooters.

Probably the next one or two years it will become standard, which would be not possible if Phase One would not risk entering this new model to it's product. Then the Photoshop will be a choice for a few professional only, when C1 UI will be a standard for photo processing. Which will then work in advantage of Phase One of course.
Greg Oleksy Send private email
Friday, January 27, 2006
I'm not trying to destroy your analogy, but front-loading washers are now taking a *premier* spot in marketing.  The Maytag Neptune, for example.
Caffeinated Send private email
Friday, January 27, 2006
Well, to steal a line from Joel, it’s about making tradeoffs, and the resolution of the tradeoff, like everything in UI design, “depends.” The cost of inconsistency depends on the specific design and the context in which it is used. Take your washing machines: front-loading washing machines can and are be marketed in the US, and the main limitation is that they tend to be more expensive than top-loading. Yes, it violates expectations, but the cost of such inconsistency is minor when the inconsistency is obvious: you can easily *see* the door is in front. I can’t believe there are users who approach a front-loading washer and claw pathetically at the top trying to open it.

Context is important, however. People entering an appliance store may not immediately recognize they’re looking at a washer, so you need advertisements, signs, and salespeople to explain. Lord help you if you put front-loaders in a coin-op Laundromat. You’re sure to get a minority of users complaining that the “dryers” (front-loading in the US) don’t work. If users will undergo training anyway (e.g., custom business software), you can afford more inconsistency. If users only use the UI once in great while (e.g., an e-commerce web site) inconsistency is a greater concern.

A while ago on this board a GUI designer asked how Find should accessed for an application that had multiple kinds of Find. How about a Find cascade menu under Edit? The designer eventually settled on putting Find as its own top level menu on the menu bar beside Edit. Good tradeoff. The inconsistency is obvious to the user --as the pointer moves to Edit to find Find, wait, there it is right there. The benefit outweighing the minor inconsistency cost is that it is easier to use than making the user fuss with those nasty cascade menus.

A bad tradeoff may be putting Find under View, like MS Windows Explorer sort of does (technically, that’s Search, not Find) --high inconsistency cost, and little or no benefit. A really bad tradeoff might be “Find” being used to open a file rather than “Open” –-that’s like going into a Laundromat and some of the washing machines look like dryers. Making things that look the same act very differently is entails a very high inconsistency cost. Making things look different but act like they look is a low inconsistency cost.
Michael Zuschlag Send private email
Friday, January 27, 2006
Incidentally, you can make a top-loading tumble washer. In such a machine, the cylinder has a hatch in its side for adding and removing clothes and the ends of the cylinder are solid. The cylinder is mounted as in a side-loading washer, so the hatch in the cylinder can be accessed via a hatch in the top of the machine. These appliances are typically as narrow as 60 cm.

In addition to tumble washers and agitators (the ones with the stick and pins) there is also a third type of washers, called pulsators. Pulsators are very gentle to the clothes, but don't wash as well and aren't as energy and water efficient as tumble washers. Also, it is harder to make an agitator or a pulsator with a powerful integrated drying feature.

For some reason, in Finland you only have a choice between front-loading and top-loading tumble washers. Manual pulsators were used but I think the move from manual washers to automatic washers also meant a move from pulsators to tumble washers. I'm not sure if semi-automatic pulsators and agitators weren't available or if they just didn't gain traction.
Aapo Laitinen Send private email
Sunday, January 29, 2006
Yeah, in the US we have front loading washers. Everyone knows they are more efficient, use less energy, less detergent, less water, and your clothes last longer.

They are considered premium washers because of this. You can buy a front loader for $800, or you can buy a top loader for $120. So, almost everybody, knowing the design superiority of the top loader, says "I ain't paying $800 for no fancy washing machine."
Art Wilkins
Sunday, January 29, 2006
--> knowing the design superiority of the front loader
Art Wilkins
Sunday, January 29, 2006
Why then is nobody selling more affordably priced tumble washers in the US? I would image that a modern tumble washer with microprocessor logic and 1,000 rpm drying could sell at $250-350 and offer most of the benefits of the $800 machine. What occupies the mid-range marked now?
Aapo Laitinen Send private email
Monday, January 30, 2006

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
Powered by FogBugz