Earlier quoted context omitted.
You know, I thought I had UI design mostly figured out, using a few simple rules: * Minimise the total number of clicks needed to access the sum of the program's functionality. * Make frequently-used functionality easier to reach than rarely-used functionality. * Don't have more elements in the UI than is easy to parse quickly. * No surprises. However, after reading this article, I get the feeling that it's a more ar…
One of the key things about UI design these days is that it's very easy to fall into the 'widget trap'. This is where you have some content you want to interact with, and instead of really analyzing what the best way to interact with it is, you just throw the data in a standard widget, like a listbox or textarea. It might be a really nice listbox or textarea, but there may also be a much better way of interacting wit…
Products For People Who Make Products For People
71–80 of 83 posts
Re: Products For People Who Make Products For People
#72Earlier quoted context omitted.
One of the issues is that current UIs for computers have very little context information about whats on screen, which would make language based interaction much more convenient. Also, editing a photo is a perfect example of where language interface would be useful: instead of hunting through Photoshop's millions of menus for the graphic effect you want, you could just say/type things like: Scale to 200% Increase cont…
But then you need to know the magic words (scale as opposed to "make bigger" or "zoom" or some other combination). Point and click context menus try to fix this -- "oh, it's 'scale'"
Re: Products For People Who Make Products For People
#73Great stuff. I think I'm developing a severe case of nerd crush on Zed. The part about linguistic experience was especially good. I have thought for some time now that all configuration files, command-line options etc. should have simple and clearly defined syntax. Some already do. My dislike of weird, inconsistent and complex syntax (and semantics) is so severe that I've semi-intentionally not learnt shell-scripting…
Re: Products For People Who Make Products For People
#74Great stuff. I think I'm developing a severe case of nerd crush on Zed. The part about linguistic experience was especially good. I have thought for some time now that all configuration files, command-line options etc. should have simple and clearly defined syntax. Some already do. My dislike of weird, inconsistent and complex syntax (and semantics) is so severe that I've semi-intentionally not learnt shell-scripting…
Graphic UIs show you what actions are available right there, linguistic UIs rely on you to think up a valid expression in that program's particular language to do what you need. GUIs will always be easier for simpler stuff, but less expressive than LUIs.
Re: Products For People Who Make Products For People
#75I agree with many of the things said in the article. I'm so annoyed with people who make open source projects that don't comply to the principles that Zed listed. (Clear code, documentation, easy to get started, reasonable defaults, support, etc.) It happens so often that I interact with such OSS projects and it's so hard to work with them. I feel bad for all the honest effort that these people put into the core part…
Re: Products For People Who Make Products For People
#76Earlier quoted context omitted.
One of the key things about UI design these days is that it's very easy to fall into the 'widget trap'. This is where you have some content you want to interact with, and instead of really analyzing what the best way to interact with it is, you just throw the data in a standard widget, like a listbox or textarea. It might be a really nice listbox or textarea, but there may also be a much better way of interacting wit…
The problem with that is that widgets are your language, so you tend to think in them and analyse the problem in terms of widgets. It's sort of the Sapir-Whorf hypothesis applied to UX, it's hard for someone to think in widgets they don't have.
Re: Products For People Who Make Products For People
#77There is really no mystery about UI. Yes, it requires skill and a deep knowledge of how humans process information, and even a touch of design genius. But the real problem is that most shops can't measure usability in a fine-grained way. And yet everyone does have some idea that it's really important. Consequently they are bamboozled by anyone who purports to have such expertise. For the most part, so-called usabilit…
You know, I thought I had UI design mostly figured out, using a few simple rules: * Minimise the total number of clicks needed to access the sum of the program's functionality. * Make frequently-used functionality easier to reach than rarely-used functionality. * Don't have more elements in the UI than is easy to parse quickly. * No surprises. However, after reading this article, I get the feeling that it's a more ar…
Usability is so much more about mental models than it is about efficiency, about helping the user feel confident & safe, than about a lack of surprises. Sensible hierarchy and visual relationships are a big thing that most people seem to forget.
Workflow, too, is critical. I've used a lot of software where every screen is mathematically, scientifically "usable" but the software overall was crummy because it duplicated a workflow from another piece of software which had got it wrong initially.
For example: 100% of email clients. Feed readers. Project management tools. They're all the same.
Surprises are actually very good in many cases. Even "ease of use" is something you don't actually always want.
My interfaces break all sorts of rules like "use standard widgets" and "conform to expectations" and "no surprises" -- but that surprise then results in an "AHA!", great enjoyment and power, and that's why people write love letters to our customer support account.
Re: Products For People Who Make Products For People
#78Earlier quoted context omitted.
Fortunately I think this is changing: tools are coming out that make running quantitative UX studies easier. Services like Optimizely, GazeHawk (disclaimer: that was a shameless plug for my company), UserTesting.com, etc are becoming more commonplace, bringing the barrier to entry for a UX consultant who is actually useful down. In the future I imagine there being plenty of UX consultants who, rather than having a st…
The problem here is that good UX isn't borne of qualitative analysis -- you have to build something worth analyzing to begin with. Once you do start analyzing, what do you do with the results? You can try to make small tweaks -- it can be worthwhile if what you have is already good. If what you have is bad however, no amount of rigorous analysis is going to solve the artistic problem of coming up with a cohesive visi…
The way I describe this to people is: "Usability science can only test a hypothesis. It doesn't give you a hypothesis to test."
Not to mention the problem of hill-climbing… you may think your test results are amazing, but you can't tell if you're really at a peak or in a valley with a lot of mist obscuring a potentially even higher peak..
Re: Products For People Who Make Products For People
#79Earlier quoted context omitted.
Graphic UIs show you what actions are available right there, linguistic UIs rely on you to think up a valid expression in that program's particular language to do what you need. GUIs will always be easier for simpler stuff, but less expressive than LUIs.
Not sure I understand correctly "LUI" but if it means something near command line, I dare disagree. Two examples of "simple stuff" that are way easier in CLI than GUI: redo last action (up - enter), list actions done recently (history).
You're right it can be easier once you are familiar with it, but there's nothing to suggest how/if "up" works if you are not.
Re: Products For People Who Make Products For People
#80Earlier quoted context omitted.
"I have come to think that GUIs as we know them are based on fundamentally flawed ideas - pointing and grunting - instead of using a language, the feature that separates humans from other animals." The dilemma is that human language is staggeringly ambiguous. This works because human beings have an extremely sophisticated model of other human beings in their head and a really good pattern matching facility, which mak…
One of the issues is that current UIs for computers have very little context information about whats on screen, which would make language based interaction much more convenient. Also, editing a photo is a perfect example of where language interface would be useful: instead of hunting through Photoshop's millions of menus for the graphic effect you want, you could just say/type things like: Scale to 200% Increase cont…