“Defaults are arguably the most important design decisions you’ll ever make as a software developer.”
You’re on that slippery UI slope … the better chosen a default is, the less often a user will have to change it. And the less often they’ll have to change it, the less exposure it needs in the UI. And as UI exposure drops, the difficulty in finding it and using it increases, leading to the misconception by the masses that the option cannot be changed.
Therefore, the better the default value is, the worse the UI is.
Okay, okay, just kidding on that last part. But my point is really that there’s flipside to this. Some designers/engineers add options or preference because they DON’T understand what the best default value is; while others feel that once you understand what the best value is, you don’t NEED the option to change it any more, you just stick with that optimum value, and you’ve simplified the UI.
And the country field is a perfect example of what Hanford is describing: why have a country drop-down in the first place if we can automatically detect where the user is coming from?
I mostly agree with you but I don’t get the comment about linux/unix. Surely the different distributions prove your point almost completely?
For instance, when it comes to the more server oriented distros one of my favourites is Debian. Why? Because it offers sane defaults. Once you install something, you’ll know it’ll work almost exactly like you expect it. When it comes to desktop systems, debian offers a bit too much choice I suppose, which brings me to the other distros.
If you look at the different distros out there, every time a new version comes out it is compared to a) the previous version, b) the current version of other distros. When making that comparison people most always look at how easy common tasks are compared with a and b. In other words, how good are the defaults in this distribution. In the desktop space, it is clear the defaults are what matter.
I don’t see how you arrive at the “there is no baseline” conclusion, since distributions constantly take cues of each other, i.e. they all started introducing desktop search, integrated note taking, compositing, etc at the same time because others did. Distro’s converge and diverge all the time, and interact just the same way as Mac OS X and windows do, and in mutual interaction with them of course. I think it is much better to view distributions separately (albeit part of the same software ecosystem), instead of trying to conflate them as one group of functionally slightly different software.
Sorry to go off on a diatribe about such a small point in your post, it’s too good to comment on any of the major points.
Right on. This is what killed Microsoft in Vista in the inexperienced users mine. They went from being one default (easy to use for the non-techy) to “secure by default” (“Yes, elevate me to admin so that I can install this already!!!”)
A lot of the backlash was changing too many defaults for the end user!
It is not a good idea to abolish the country field in favour of detecting via the IP address. There are proxy services that users might use that might give misleading IP addresses to your software, there are also IP addresses that aren’t specifically assigned to a country, and your users might be using IPv6 addresses that might not yet be accounted for by your IP-to-country code.
I think the Apple comparison is important here. Desktop search has been around for years on XP, through external apps. On Apple it’s there by default. FlyAKite OS makes Windows XP look really nice. Apple OSX looks good by default. eXpose clones exist for XP, on OSX it’s there by default. http://sanaris.ru
Jeff, your point about the power of an app or feature being installed by default is a good one. A couple of related points:
I’m a heavy cmd.exe command prompt user, and when I first heard about Powershell, I was initially very excited about it as a cmd.exe replacement. However, my enthusiasm for learning PowerShell waned significantly after I read that PowerShell wouldn’t be shipped with Vista – this means that even after Vista becomes the prevalent business desktop OS a few years down the road, it still won’t be installed on 99% of the machines that I sit down at other than my own, and I’ll have to fall back on cmd.exe anyway when I want to work from a command prompt. Thus, the value of knowing PowerShell is diminished.
Similarly, I still use notepad.exe for many simple text-editing tasks. 3rd-party Notepad replacements like Notepad2 and Notepad++ are well and good, but they aren’t going to be installed on machines other than my own that I sit down at, so does it make sense for me to rely on their features? (For non-lightweight text editing on my own machine, I generally just use Visual Studio, which at any given time I usually already have one or more instances of running. I find that notepad.exe itself works poorly, yet is still “good enough,” for lightweight tasks.) The value of Notepad2 is diminished simply because it isn’t on every machine that I sit down at.
The argument that “they can’t ship a better notepad” is rubbish. I addressed some of the arguments in my post, but here’s a summary:
quote
Read the court’s judgement (http://www.usdoj.gov/atr/cases/f200400/200457.htm). The restrictions on bundling are mostly on what Microsoft can require of OEM’s, and if OEM’s can add or remove bundled software. Fine, let them remove it. It’ll sell as well as Windows N.
And an open source project like Paint.NET would completely sidestep the “secret API” issues, since the source is publicly available.
Look, the “we’ll probably just get sued” way of thinking is out of hand. If we followed that line of thinking, Vista would have no new features. Windows XP has notepad and paint program, neither of which were mentioned during the anti-trust case. No one said they couldn’t update these applications.
/quote
