[clavalle]

"you may not have to code for a UX job, but it really helps you get a broader perspective and can improve your work"

If you are a UX designer and cannot code up your ideas it is like being an architect that knows nothing about materials and construction.

[arstin]

I like the analogy, but I'm not sure how you intended it since I think it can reasonably make different points. It can surely help if an architect knows about material and construction (see, for example, the work of Christopher Alexander's group who relied on knowledge of materials and construction to wildly reduce costs while increasing quality). But it's also plainly true that any architect is going to be dependent on skilled contractors. Might be fun to read or write an article exploring all the nuances of the analogy.

Speaking as a UX designer by title and inclination who has to spend most of his time as a full stack dev by economic necessity, I've come to the view that in an ideal world most designers ought not have to code. It works well for me, since I'd personally get bored only implementing or only designing. But every week I feel the tension between very different demands on my limited energy, ability, and time. I feel how I almost subconsciously compromise my design work to get it to something I can implement in a realistic time, how the fact I've invested effort into the implementation very subtly pressures me to not make serious design changes, how my figuring out a cool coding trick or library encourages me to add it to my design when it's actually a superfluous distraction (see the vast majority of products by UI engineers who care about design).

To be honest, I'm hard pressed to say how being able to build essentially whatever I want has helped my design work, especially now that there are pretty good, quick to use prototyping tools like Atomic. Maybe it has in such a deep way that I don't even notice it, but I'm skeptical! :)

That said, I do personally agree with the author that, all else equal, some coding knowledge is a good thing if only to help in communicating with and empathizing with developers.

[clavalle]

Having hired a UX designer that had never coded, it was a very frustrating experience.

I enjoyed the high-touch with customers approach that they brought to the table but when it actually came to implementation it was as if I brought an astrologer to an astronomy meeting -- a lot of the same words were said but neither knew what the other was talking about.

There was just no sense of proportion on the UX side.

I am not saying a UX designer needs to code in the course of doing their UX work, but I do think they need to know how and have had some experience in it. Otherwise, the distance between saying something and doing it is too great.

[arstin]

Yeah I can understand that. I've worked with plenty of designers who can't code, and to my taste the only part that can be a little frustrating is when they don't accept that some great ideas really can't be built. But that said, I do think that negative is usually outweighed by the value of keeping the ideal in front of the reality...it's just they need to be sure to also help with what needs delivered in two weeks!

As sort of a joke, I've been on the flip side of that---especially with more enterprise-y Microsoft stack style dev teams---where I propose a design and am told it's impossible and I respond with a detailed explanation of how to build it within their constraints. I think they like that even less! :)

[kpil]

I'm a developer with some interest in UX, design and perhaps most, graphical design and typography.

I'm fine with no-coding UX specialists however - I have notice that not being able to code at all while working in a rather "techy" role is often a proxy for not being able to create a coherent mental model of the the processes and features that they are building an interface for.

This might be a bigger or smaller problem - if the system complexity is low, then maybe "superficial" features like readability, flow and layout is more important. It's hard enough to get good at - I suspect that some inherent flair is required to get good at that.

If there is a true inherent complexity, then the more hard-core aspects of modelling and how well the users can understand and interact with the model is of more importance. It's another skill set, but when it's needed it's very important to get right.

Learning to code typically enhances ones modelling skills, and more importantly "failed" devs are probably not suited for UX either.

There are of course plenty of people that are intelligent and smart enough to get god on "modelling" on their own.

That said, there seems to be a lot of UX people around that have read a few books and are practicing some generalized rules they found on some websites, and could fall head first in big pile of good design without noticing it... I doubt it would be hard to compete with them. A cool haircut seems to be enough.

While I am whining, while Personas is a good tool - keep it to a minimum - no full size cutout dolls, CV, background story, accessories, etc. I suggest to keep it to the level described in User Story Mapping (Patterson) and stop there... Anchor the personas with the sales guys, support guys, and people that potentially have met a user.

[arstin]

I think your comment about modeling is a good one. It actually fits with how I view UX design. I think it's no accident that the best UX designers, in my experience, tend to come from a psychology background rather than an art school one. (And also why so many great UX designers can be poor at UI design/branding!)

To the extent coding helps train you to think clearly about complex and---especially in this case---ambiguous things, it probably could help your ability to design. But! I've often been baffled why most developers, who should be good at modeling, are so often unable to transfer those skills to UX design.

The best explanation I've been able to give so far is largely that "modeling" is somewhat equivocal when viewed in terms of day to day work. Think about the very different skills, knowledge, values, habits of mind, etc it takes to construct a useful data model for a relational database, a predictive dynamic model of neuron populations, an evocative persona model of users, and so on. One neat observation here is that if you try to build, say, a persona in a similar form and with similar standards as a database data model it will be a disaster...unable to fulfill it's purpose. (Not disagreeing with your comment about ridiculous persona models here!)

IDK, this is cool to think about and there's a lot more to say but I'll leave it there.

[kpil]

I feel that i need to balance my remark regarding incompetent UX specialists with stating that for every useless UX specialist, there are at least 10 useless developers.

Done.