Some thoughts on “open source design”

On Facebook, Máirín Duffy, a fellow designer in the community, posted a link to “Can truly great design be done the open source way?” I commented with a string of thoughts on the topic. I’m posting my response here for a wider audience.

Pandas Forever: a photo I snapped in London's Design Museum

The following is (a slightly edited form of) what I quickly wrote:

First, two points in direct response to the link:

  1. Apple’s design isn’t fully consistent, if one looks close. It is consistently high quality, however.
  2. Rhetorical question: What is “the open source way”?

There are many ways open source software is made, and some of it isn’t necessarily out in the open.

It is always assumed in this sorts of discussions that people code and software is churned out… But really, there are talented people with a vision controlling what goes into a codebase in many successful projects. (Linus and the Linux kernel, for example.)

A comparison is usually made between an entire project and a single design. That’s always ridiculous, and this is one of the reasons why these discussions are not so useful.

I’ve never seen “community coding” done by contests and voting for individual source files. Imagine if programmers were treated like that. How many would stick around?

The phrase “design by committee” is always brought up in these sorts of discussions, but nobody ever stops to think about “programming by committee” (and, obviously, that doesn’t happen in the open source world). Often there are too many opinionated, amateurish cooks in the design kitchen, spoiling the broth.

The Tango project and its offshoots, such as the Tango-styled Gnome icon theme, are examples of where open source (visual, in this case) design works well. There’s a standard, shared vision and a handful of talented designers work toward that goal.

Essentially, for programmers and for designers, there are some talented people doing stuff and later releasing it. The difference is that everyone has an opinion on UI and visual design, even if they are no good at it themselves. Not everyone using software is so opinionated about its source code. (People do not refuse to use resulting software because of indentation (tabs vs # of spaces) and coding style in the source code, for example.)

11 thoughts on “Some thoughts on “open source design”

  1. Good post, yet I digress on your last point. People (non-coders) do refuse to use software based on Mono (or say any other scripting language) with arguments that raise from patent FUD to performance FUD.

  2. When talking about any sort of design I like to draw upon the phrase “Comment is not contribution”, which links into Mark Shuttleworth’s recent comments about how the Ubuntu UI is not something to be voted on.

    I think maybe open source design can work, but it’d have to have everyone actually contributing the equivalent of code e.g. graphics, ui designs rather than just commenting on how good/bad something looks.

  3. I think another issue with that sort of design is that it’s often very small projects. The time required to do a new background for the next release of a distro or for inventing a logo is comparable to writing a tool like dos2unix. And small, easy projects invite bikeshedding. And that explains why everyone has an opinion.
    And if you have larger projects like Tango icons or a Blender movie, you automatically have a meritocracy with maintainers that rejct patches but also the ability to contribute small things via patches. And those projects seem to work quite well.

  4. Tango is also able to succeed because it is a highly modular project. It is more difficult to distribute the work involved in highly interconnected design tasks.

  5. When Mairin speak of the open source way, I assume she talk of following http://www.theopensourceway.org/book/ , which is a book written by the Redhat community architecture team. As explained on the wiki (http://www.theopensourceway.org/wiki/Main_Page ), they try to apply this to a wide range of topic. Too bad, the book is not finished yet, but I recommend reading it ( and seeing Max Spevack’s speechs too, like the one he did on fosdem ).

  6. +1 especially the comparison of “design by committee” to “programming by committee” (which, as you state, never happens). design isn’t a democracy, but if you don’t have strong leadership the passing of any new design ideas as law is impossible.

  7. @Allan: I very much disagree. Tango probably only seems highly modular from the outside. The goal of Tango is not to have an icon theme. It is to provide guidelines for a unified style across Linux. Having arbitrary people draw high quality icons in the Tango style, is quite hard to accomplish. It’s worked out ok so far, because there are a few core artists, all drawing the icons for the few projects that they support outside Tango, with artwork.

  8. I disagree with the idea that programming by committee doesn’t happen. Take a good look at -devel lists for GNOME and KDE, or at internal discussions at $company… It happens quite a bit in “open” source.

  9. “The difference is that everyone has an opinion on UI and visual design, even if they are no good at it themselves. Not everyone using software is so opinionated about its source code.”

    I am not sure I got this. Users do care a lot about the code I write. Namely when it produces crashes, erases their harddrive, or worse.

    So even if users don’t care about the source code per se, they sure do care about its effects. Similary, users might not care about the golden ratio or color theory, but they sure tell you whether a design is pleasant or not.

Leave a Reply