Design is critical for open source projects

In an apparently "not very popular" post, I spouted about the bad, and total lack of, design of many technically-inclined websites. I also said that design is an essential part of the "selling process", and that if you're trying to persuade me to use your project - you're selling. The cost has nothing to do with the persuasion; the biggest challenges in selling are getting someone to pay attention and then making the purchase. For open source projects, that translates to getting noticed, getting downloaded and, finally, being used. So the whole thing is actually more difficult for open source projects! No marketing budget means no marketing - and that's just the beginning. If I've paid you money, I'm inclined to use the product; if I've downloaded it from GitHub, you just have to hope that I actually remembered I did so, what your efforts do and then hope I can persuade myself that they're worth using!

Well, the first part of that equation, that process, is called "your website". With no marketing staff you really need a good website. It's a persuader. A GitHub page is okay, but if you've got a large project, it absolutely does need its own website. And that's where a lot of projects fall down.

Sourceforge, which I was surprised to discover still exists, has poor web design. The information is hard to find; it's not presented in a visually appealing manner. It's a shoddy interpretation of minimalism. GitHub is a bit like Facebook. The design is okay, but I can't help but notice it's a 'lowest common denominator' design. Nothing stellar, just usable, visually inoffensive, but very welcoming. At least Sourceforge isn't exactly hostile to users, that honor doesn't even go to Fossil-SCM, another open source repository. It goes to Google Code. Their vaunted "data-driven" design process either produced something only its (clearly not artistic) designers could love, or, more likely, wasn't considered. Mind you, Google's data-driven design doesn't produce good design, anyway. It produces innocuous design, which is arguably slightly worse than "no design". So which one of these is the most popular? GitHub.

I don't care which one has the most repositories, or users or what-have-you. GitHub is popular, it's the clear winner and it has the developer's attention. Because it is well designed and welcoming. Good design is important. Look at Pinterest - that's superb design. (Take a look, in particular, at their login screen.)

Now, some might say that my design skills aren't exactly stellar. They're not. I couldn't come up with anything like Dan Flavin's lights, or Jasper John's flags. But that doesn't mean I can't appreciate the art - it simply means I don't have their unique vision and abilities. In the land of technical skill and in the technical meritocracy of open source, where the yardstick of design includes only technical prowess, it is often argued that "if you can't do it, don't criticize it". I'm not an architect, but I know that most houses are shoddily designed. I know that many malls are blandly architected (this one, in Santa Monica, could never be called "bland"!). I'm not a user-interface designer, but I know that MS Windows has very little aesthetic design (MS Office having less than that...), Linux has no aesthetic design and OS X is (usually) a delight because it is designed and engineered so well.

When it comes to design, the argument that you shouldn't criticize what you can't do is an old ignoratio elenchi, a non sequitur on steroids. Most people can't sew, but they still appreciate well-sewn clothing! I can sew, but I'm never going to argue that you can't critique clothing because you can't put cloth together like a haute couture clothier! (I can't do that, either. And, by the way, haute couture designers hire expert sewers - their own talent being in design.) Design is a skill that needs a large dose of talent; you can learn the basics of design by picking up a book or two, but you're not likely to be the next Elliot Jay Stocks without that natural, innate ability to "see" visually.

All things considered, if you have little aesthetic sense, it's best to use a framework for your website. Like a technical framework, you trade some "freedom" for "that bit's done, already". These days, the go-to framework is Twitter's Bootstrap. Reuven Lerner, over in Linux Journal, notes that it has led to some lamentation about "sameness" over on Hacker News (not exactly an exemplar of good design itself, by the way), but that's easily ignored. What's important is that you impart a sense of confidence to those you are trying to persuade. A website that consists of text doesn't do that - it tells the reader, the potential user, that you don't care about design. And that, in turn, informs a lack of confidence in other aspects of your product's design. A basic design framework, such as Twitter's Bootstrap, gives the reader some confidence that you appreciate design, that they don't have to hide the website of a tool they use. When you're trying to persuade someone visually, design is critical. And with open source tools, you're not just trying to persuade someone - you're trying to persuade them to persuade themselves that your tool, system or what-have-you is worth their time. That it's a solution, that it is worth more to them than sitting down and writing their own solution. In the open source world, good design also tells the reader that your solution probably has legs. And that's worth a lot more than any argument about how design shouldn't be necessary.

Carolyn Ann

Added: Anna Powell-Smith wrote an interesting article about design.