Someone believes that multiple skinning (the ability for a user to change an application's look-and-feel) is, at best, a complete waste of development time and, at worst, throwing away a key usability advantage of windowing environment (having a common look-and-feel across applications). The worst culprits are applications that do not even support the 'traditional' look-and-feel.
If you're considering developing software with the ability to change the skin, please make sure you ask yourself these questions: Will you put off some users who can't work out how to operate your non-standard interface? Do you really want to have to test multiple interfaces? Would your development time be better spent implementing new features or improving existing ones? Is this really a good idea?
It's not all bad. Allowing users to create their own look-and-feels can be a good way to create a dedicated community of users around your product.
But it doesn't have to involve the whole GUI. Just letting the customer place their company logo in a corner of the application can be a highly selling feature.
A Rebuttal:
Users want skinnable apps. The first successful skinnable apps were MP3 players. A skinnable interface makes it possible to fit the controller into a corner of the screen.
No, a custom-drawn interface makes it possible to fit the controller into a corner of the screen. Replaceable skins have nothing to do with it. The bog-standard Microsoft CD Player application has done this for years, without being skinnable.
The 'traditional' look-and-feel should be supported as a default. But skins are largely subject to natural selection. Those which don't fulfill a user requirement (find success in a niche) won't be used. Meanwhile, experimentation in user-interface design may end up changing creating new look-and-feel traditions.
There is an interesting paradox here. The skinnable MP3 players are by all standards very successful, but at the same time they often have very bad usability. Some niches of applications seem to survive with bad usability, while others are killed by it.
If a skin has bad usability, it would tend to fall out of use, don't you think? I mean, no one is forcing you to keep an unusable skin. I don't think skinnable apps are bad, but there should always be a "Default" that has a standard look-and-feel.
Then of course, there are games. Mods and Skins have been a big part of gaming since somebody changed the font in Net Hack. A game like The Sims takes it literally (not only setting the skin, but the hair, and clothes).
I think you are correct that some applications are not 'skin-worthy'. Microsoft Word (for example) would be silly to skin.
On the other hand, Power Point is arguably an 'outline'-skinner (even though everyone uses the same background).
Mozilla Browser is skinnable, but the skins that people use are basic variations of the normal browser. We've yet to see a wildly divergent skin.
If a skin has bad usability, it would tend to fall out of use, don't you think? No. Lots of bad things achieve popularity and stay popular. Hell, cigarettes are popular.
Cigarettes have a great interface (Teach Me To Smoke notwithstanding). Anyone who can hold something between their index and middle finger and can flick a lighter can be smoking in moments. People who smoke habitually can easily take out cigarettes and light them without thinking about it at all. Though I've never been a smoker of cigarettes, friends who are say one of the things they like about cigarettes is always having something to do.
I've heard the same argument used to defend alternative medicine: if it didn't work, people would stop using it. One could make the same argument about gambling fallacies. No matter how much money people lose at casinos by following gambling fallacies, they don't give up the fallacies.
My crappy ghetto blaster from 1988 has better usability than any CD-player application I've seen on Windows. You'd think that sucking would create opportunities for better CD-player apps to take over the market, but it hasn't happened.
Skinnability is a distraction. It sates the masses by dazzling them with shiny objects. A product that sucks can still dominate a market. Does anyone here seriously think that the best product tends to be the market leader?
See also: www.suck.com 
Compulsive Customizers are major victims of Skinnable apps.
Skins are bad. (If you must have skins, bland is good.) Designing good UIs is a hard problem and few end-users have the ability to improve on the basic widget set, which is usually of high quality, designed to promote inter-operability and decrease the time taken to learn new applications. Most people attempting skin design "paper over" all the subtleties which would be present in the basic widget set (drag handles etc.) and proudly produce applications that stick out like a sore thumb and are unusable to anyone other than them. And the user who downloads it and splatters Britney across their application won't realise that the lack of subtle visual cues is slowing them down. The native widgets at least put an upper limit on the terribleness of UIs, even if it is a very high limit. But of course, skins are popular because they serve the same role as T-shirts and ringtones, i.e. letting people show off. They also boost people's sense of control over their applications and thus their computer, which is a good thing, but not sufficiently good. -- An Aspirant
But see also Customizable User Interface
See original on c2.com