I'd actually go one step further, and say that user interface is /more/ important than code. In almost all case with a start-up, you users don't really care about your code, they care about the functionality, and its usability. That means that HCI trumps code, to some extent.
Of course, interface matters only in so far as it exposes functionality; that's why craigslist, which is a monster of design, but highly usable, can do so well. The interface facilitates the user's interaction with the functionality.
Agreed. What users see is way more important than what they do not see (except maybe in the sense that the code is producing what they see, but, anyways).
As for craigslist, the author of that post seems to be using "design" interchangeably for "interface design" and "visual design". Yes craigslist is ugly, but they get away with it because the interface is designed quite well. I can go to craigslist with a task and figure out what I need to do and get it done relatively quickly. So while a appreciate your intent with "monster of design", I think it's better put "monster of visual design".
I agree. In addition, some interesting notes about HCI. First- if you haven't studied HCI or know what the acronym stands for, you may make it as a programmer for some company, but you're not going to reach your full potential creating new things.
Anyway, I was reading an HCI paper in one of the publications a few years ago and came across these interesting statistics:
75% of software does _not_ in fact increase the productivity of the end-user. (over a paper-based alternative)
Doing a single user-testing session results on average of fixes that take less than 2 days and that increase the usefulness of the product by 80% (that is it increases the end-user's productivity that much).
Here's a poor-man's usability test. Get a bunch of people who have never seen your product and don't know much about it, but that are in your target demographic (shouldn't be difficult- a brother, roommate, family of the employees, someone off the street). Sit them down in front of your product and tell them you're not going to be allowed to answer any of their questions because you're measuring the usability of the system. Then stand over their shoulder and let them do their thing. Don't make a sound- no guiding, no "oh- that actually works by doing this"- just observe and take notes. A few sessions like this will mean a world of difference in your UI.
A well planned project shouldn't have this problem since the UI should come before the code. The decisions for UI should be resolved before the code since it will dictate how the different functions will interact. Sometimes a good interface design will reveal that a there's no room for certain seemingly good features, thus eliminating a lot of potentially trivial work. Also, while designing the UI before the code won't necessarily affect the quality of the code, doing the code before the UI could substantially jeopardize the quality of the UI.
I've been trying to spend a lot of time on graphic design lately, and the link to the free icons will be very helpful. http://www.iconbuffet.com/freedelivery/packages
6 comments
[ 2.2 ms ] story [ 16.8 ms ] threadOf course, interface matters only in so far as it exposes functionality; that's why craigslist, which is a monster of design, but highly usable, can do so well. The interface facilitates the user's interaction with the functionality.
As for craigslist, the author of that post seems to be using "design" interchangeably for "interface design" and "visual design". Yes craigslist is ugly, but they get away with it because the interface is designed quite well. I can go to craigslist with a task and figure out what I need to do and get it done relatively quickly. So while a appreciate your intent with "monster of design", I think it's better put "monster of visual design".
Anyway, I was reading an HCI paper in one of the publications a few years ago and came across these interesting statistics:
75% of software does _not_ in fact increase the productivity of the end-user. (over a paper-based alternative)
Doing a single user-testing session results on average of fixes that take less than 2 days and that increase the usefulness of the product by 80% (that is it increases the end-user's productivity that much).
Here's a poor-man's usability test. Get a bunch of people who have never seen your product and don't know much about it, but that are in your target demographic (shouldn't be difficult- a brother, roommate, family of the employees, someone off the street). Sit them down in front of your product and tell them you're not going to be allowed to answer any of their questions because you're measuring the usability of the system. Then stand over their shoulder and let them do their thing. Don't make a sound- no guiding, no "oh- that actually works by doing this"- just observe and take notes. A few sessions like this will mean a world of difference in your UI.