"Sometimes I think a pretty good business model would be to copy the applications that 37signals makes, but make them more complex. More features, more promises—generally, just more complicated."
I would argue that KISS is actually a utilitarian choice. I have always interpreted KISS as how do I go from idea to product in the simplest manner, and keep both development, and the users interaction with the product simple.
If you look at new twitter, it is actually simpler than the old one, because now every tweet displayed only has one initial option, click to open, instead of the five per tweet that were displayed. They are giving us the user less initial options.
Also, you're site is pretty simple, and I would like to say that is a good thing, because all the user can do is read, they can't even attack you in a nice comments section. ;)
A stick is a simple tool. It's also a very useful and versatile tool that can be used in complex ways. Complicated in software would be like taking a stick, then adding an options panel, then putting a spade at one end, then a rake at the other, and 10 status indicators running the middle.
I think the UNIX philosophy goes a long way here. FP is also close to this. But it's not so simple... har har har.
Certain UNIX tools are fairly stick-like in their basic usage. Things like cat, grep, sed, sort, cut, paste, tee, less... they are all at least immediately useful with a basic command even if they have huge sets of options.
The UNIX grep vs. Windows Grep (see http://www.wingrep.com/resources/images/mainwindow1.png) is a great illustration, because the basic use case of grep is as simple as possible (`grep foo' for instance, operating on stdin). With the complex Windows Grep you get the whole UI (and more, with things like wizards and prompts) all the time, no matter what you need. You could never learn anything more about UNIX grep than `grep foo' and it would never get in your way, but with Windows Grep and other complex apps you get everything all of the time.
But, in the end, UNIX grep and Windows Grep can both basically do the same things.
His logic doesn't quite track for me. Take the Twitter example. He uses that as an example of why you shouldn't focus on simplicity because it makes twitter busier with inline media. But to me that's an example of simplicity because it makes sharing media much simpler.
Any new feature is going to add some complexity. The point of a K.I.S.S. philosophy is trying to lessen that complexity as much as possible and that's example what Twitter did.
The point is that Twitter started out with simplicity. My guess is if they started with their current design then there would be far less adoption because of the steep learning curve.
Before anything you should define what constitutes simplicity for the purposes of your post.
New Twitter is much more complicated than the old one.
You are confusing simplicity in the user experience with simplicity in the back-end. I'd argue the new Twitter is simpler. But its back-end is more complicated.
Great front-end simplicity often requires a lot of back-end complexity.
The argument that there is a tradeoff between simple/useful is something I don't agree with. An app or website can and should still remain simple to use whilst having many features if the UI is designed well. Simplicity should always be the end goal but this should not be at the expense of features!! There seems to be a certain mindset where people believe that minimal, or simple interface design is the 'easy' option. It is not, it's often the opposite. Good design ensures an end result always feels AND looks simple to use for the end user. The features added to a service should never be considered, or be resolved in a way where they could be described as 'clutter', as you label it. I'd consider that a fail if I looked at a new feature that way. If the features are implemented well by a good UX/UI designer, the products perceived 'simplicity' will not be compromised, in fact it would probably be enhanced if it were providing added flexibility.
10 comments
[ 3.7 ms ] story [ 41.5 ms ] threadhttp://www.joelonsoftware.com/items/2010/08/19.html
If you look at new twitter, it is actually simpler than the old one, because now every tweet displayed only has one initial option, click to open, instead of the five per tweet that were displayed. They are giving us the user less initial options.
Also, you're site is pretty simple, and I would like to say that is a good thing, because all the user can do is read, they can't even attack you in a nice comments section. ;)
Certain UNIX tools are fairly stick-like in their basic usage. Things like cat, grep, sed, sort, cut, paste, tee, less... they are all at least immediately useful with a basic command even if they have huge sets of options.
The UNIX grep vs. Windows Grep (see http://www.wingrep.com/resources/images/mainwindow1.png) is a great illustration, because the basic use case of grep is as simple as possible (`grep foo' for instance, operating on stdin). With the complex Windows Grep you get the whole UI (and more, with things like wizards and prompts) all the time, no matter what you need. You could never learn anything more about UNIX grep than `grep foo' and it would never get in your way, but with Windows Grep and other complex apps you get everything all of the time.
But, in the end, UNIX grep and Windows Grep can both basically do the same things.
Any new feature is going to add some complexity. The point of a K.I.S.S. philosophy is trying to lessen that complexity as much as possible and that's example what Twitter did.
New Twitter is much more complicated than the old one.
You are confusing simplicity in the user experience with simplicity in the back-end. I'd argue the new Twitter is simpler. But its back-end is more complicated.
Great front-end simplicity often requires a lot of back-end complexity.