3 ms·
I'm going to review this from a very subjective standpoint. If you want, you can just announce that I'm not the intended audience, that's OK. So, I'm a lead dev
by golergka 11y ago
I'm going to review this from a very subjective standpoint. If you want, you can just announce that I'm not the intended audience, that's OK. So, I'm a lead developer of a mobile project (usually, a game), who wants to improve collaboration with designers.
They work in a professional WYSIWYG design tool like Photoshop and Illustrator (most know both). Then I take their designs and recreate them in the WYSIWYG design tool — Unity3d UI engine, or iOS Storyboards, or this Android XML thingy (I never actually worked with it, but from a couple of tutorials I got a feeling that it's not that different). Re-creating these designs is work-load problem as well as a communication one.
Now, this service allows me to solve this communication problem by asking designers to learn this tool — which is new to them, which requires time, and also isn't as powerful as Photoshop.
Or I can just ask the designers to learn the tool that WYSIWYG tool that I use to actually create their designs — like Unity UI or iOS storyboards — and solve both of these problems at once. Of course, they will also need to learn the scary words like "Pull", "Commit" and "Push", but after long experimentation I found out that designers are actually not brain-dead (as many engineers erroneously think) and are capable of understanding the basic VCS concepts.
This leaves me with a question: how is the first solution better than the second?
And also: how soon will we finally understand that the 00s hipster/indie-kid aesthetic is completely dead, and using it in a product video is not "cool" anymore — and haven't been for the last 5 years? Seriously, "corporate suits and elevator music" image would look much more fresh.
- jasim 11y agoWere you actually successful in getting the designers to ditch the WYSIWYG tools and write code instead? Agree strongly that designers are capable of using version control, navigate the command line, and even write scripts that make their life easier (the Sketch plugin ecosystem is a good example). However, it was my observation that to do the thing that they care most about, which is building great interfaces, they do not need to know any of these things, and they do it only for the benefit of the developers. Which is a good reason why tools like these exist. I'm curious to hear your thoughts around this. Thanks. (I'm not at all affiliated with Figma and in fact I'm building a tool in the same space)
- golergka 11y ago> Were you actually successful in getting the designers to ditch the WYSIWYG tools and write code instead? No, this would be a stupid thing to do. I suggested moving them to the WYSIWYG tool that is directly coupled with IDE instead.
- endergen 11y agoI like your point about having designers a the particular platforms UI tool/system. It's sad though that they can't learn a platform agnostic tool so that skills are more shared and portable across platforms. So I like this kind of solution for tackling that. Also, making it easy for designers to mockup interactions in more than just series static mockups is much better for articulating their designs. This is why so many design prototyping tools are coming out of Facebook, Apple, etc these days. Having it online, collaborative, and platform agnostic seems like a strong basis for a successful production provided they can provide tooling to make it easy to use on popular platforms. I think an embeddable UI runtime you can drop into projects would be the ideal technical approach.
- IshKebab 11y agoTo be fair those tools only existed very recently, at least on Android. And it's only very recently become good enough for anyone sane to use.
- tptacek 11y agoThe Photoshop-to-code problem is also well known in web design, where a third solution, presumably applicable here as well, is the emergence of transactional PSD-to-(whatever) services; designers make PSDs, and you pay $50 to have them converted to markup, which adds only marginally to the cost and only a day or two to the turnaround time.
- golergka 11y agoBut this is not Photoshop-to-code problem. I thank the gods I have little to do with web and javascript almost every time I read something about that field. This is Photoshop-to-another-designer-friendly-WYSIWYG-tool problem.
- jasim 11y agoThere have been a couple of tools that have attempted to automate this. A quick google for "automatic PSD to HTML" will turn up many. But almost all of them produce markup that is tightly dependent on the organization of the layers in the source document. Also the elements are usually absolutely positioned and the code in general needs a ton of work to get to a good shape. There is no reason for this to be so. This can be automated - a developer should be able to convert a design into well-written HTML and properly factored CSS (using SASS mixins) in under an hour. This should make much faster and cheaper iteration on the front-end possible. We're building a tool to do this for Sketch - https://protoship.io https://protoship.io. It is early days, but we've been dog-fooding it to do conversion projects ourselves and I can't imagine going back to manual HTML/CSS coding again.
- wizzzzzy 11y agoJust out of curiosity, what kind of design do you expect a developer to be able to convert into well written code in under an hour?
- jasim 11y agoAll sorts of web pages. The tool will create HTML and CSS, the elements will use DOM flow and positioning, and grid columns if that information is annotated through our UI. The developer has to give meaningful names through our UI so that we can create CSS classes that follow BEM conventions.
- pavlov 11y agoAs you've worked in games, perhaps the designers you've worked with are more adept than average at taking up new tools...? At least that is my experience. Game artists often have experience with all sorts of in-house tools with UIs of highly varying logic and consistency, whereas most "traditional" designers have used Adobe tools and nothing else. It seems to me that the success of Sketch lies in how little it diverges from the Adobe formula. Sketch doesn't really add any new significant mental model: it adopts Illustrator's terminology (such as artboards) and some of Photoshop's bitmap capabilities, and cleans up the three decades of cruft that obstruct the Illustrator and Photoshop interfaces. That has clearly been a good recipe for something that's easy to pick up for designers. The UI layout tools in Xcode and Android Studio are not like that. The learning curve is very steep. Interface Builder is a tangled mess of quirky NeXT legacy and complex new concepts like auto-layout constraints. Android Studio's UI editor is worse because it's "property editor hell" -- editing layouts by twiddling Android's endless and confusingly named properties will drive designers mad. (I used to think that Microsoft's Visual Basic was the epitome of this type of tool, but Android is much worse.) I've worked on a tool called Neonto Studio that tries to provide a designer-friendly cross-platform mobile UI solution: http://neonto.com http://neonto.com It generates native code for both iOS and Android, there's no framework. It has been really difficult to figure out a layout model that is both understandable to designers with Adobe/Sketch background _and_ can translate reasonably well to iOS and Android classes! Our current effort is pretty good, but it definitely has room for improvement. Still, I would suggest that Neonto Studio is a much gentler way to introduce mobile UIs to a traditional designer than throwing them into the deep end of Interface Builder or Android Studio :)
- dclowd9901 11y agoIn my experience, getting most designers I work with to use any new tool or adapt to any new process is about as easy as preventing the sales staff from selling features before we've written them. At my last job, we couldn't even pin designers down to a pattern library. Why am I the one who has to convince designers a pattern library is a good idea? I digress. This tool is a good thing because specialization is good, and if I, someone who uses git on the reg, can still monumentally fuck up its usage, I'm not going to force it on anyone who doesn't absolutely need it.