|
HN users, on the whole, are terrible UX critics who have no idea what they're talking about on the subject. The average HN user probably: - knows their way around a terminal - understands default browser behavior in depth - has the programmer's mentality of 'everything should be governed by universal logical rules that apply equally everywhere' The truth is, this is not how most users think. Good UX for a cli application is not the same as good UX for a website. If you think of websites in terms of what programmers think good UX is, you get... HN and Craigslist and all the other things that HN users typically think are fantastic user experiences, because they meet their expectations. These applications are also almost uniformly ugly. This actually makes perfect sense, and for these users, this is fantastic UX. But what makes UX good is that the target audience can easily use the software. That's it. There's no such thing as an 'intuitive user interface'. Clicking on stuff in computers is not natively intuitive to humans. There are only familiar and unfamiliar interfaces. Every interface we haven't seen before is unfamiliar and 'terrible' at first glance. People will train themselves to use and love extremely bad interfaces, or refuse to use very good interfaces that are difficult to learn. The HN user is someone who has typically put a lot of effort into learning how computers work and how web browsers work by default, so they develop intuitions like, "If a website breaks the default browser behavior, it's bad UX." And for them, that's true, but what makes the default browser UX better than an alternative, other than that you already know it? Now maybe it actually is better, and I can think of many examples of websites that change default behavior in ways that are absolutely worse. But I've also seen websites do cool things with scrolljacking that aren't inherently wrong because they defied expectations for 3 seconds. Every application has to be learned by the user. How easy it is to learn it depends on the background knowledge and experiences of the user, so it's very very hard, if not completely impossible, to develop a UX that will be good for every possible user. But if some UX isn't good FOR YOU, that doesn't mean that it's bad, it might just mean you aren't the target audience. |
So would you argue that the Craigslist experience is only good for those with the "programmer's mentality"? "Ugly" is perhaps a smaller factor in UX than designers account for.
> And for them, that's true, but what makes the default browser UX better than an alternative, other than that you already know it?
Nothing, that's just it: you already know it. What you call "programmer's mentality" is actually fundamental to human cognition. In making sense of things we use what we've already made sense of. The more we can rely on our existing knowledge to figure something out, the less cognitive friction there will be.
The cost of surprise can certainly be outweighed by other factors, or surprise in itself can be exploited to usefully convey something, but in my experience it is not generally used to these effects. For every zooming interface or slideshow where scrolljacking might make perfect sense there are hundreds where those three seconds of defied expectations are wasted to implement something that is either useless or entirely detrimental to usability.
> But if some UX isn't good FOR YOU, that doesn't mean that it's bad, it might just mean you aren't the target audience.
Or I am the target audience and the designers have failed to design for the target audience. Or I am not the target audience only because the designer has failed to identify and characterize their target audience correctly. Or the designers primarily work with making business-required anti-features bearable. Or they're optimizing for first impressions and not long term usability. Or they're optimizing for generating more work opportunities ahead of themselves.
To decidedly say that you, our user, is not our intended target audience seems like a conclusion that must be backed with a lot of data, something which IME not a lot of organizations can muster. From that perspective these alternatives seem more likely.