Skip to: Navigation | Content | Sidebar | Footer

Full Archives


February 24

Every 6 months or so the perceived lack of diversity in the world of web design comes up as an important discussion point. The most recent provocateur was Jason Kottke, who sat down and worked out the percentages of women speaking at recent and upcoming web conferences. And it's not pretty.

Since I just spent a bunch of time running a conference during which us organizers most certainly had discussions along this vein, now seems to be an important time to jump in with a few thoughts. I've been a bit tied up with, hey, yet another conference, so I'm doing this after having read multiple viewpoints from elsewhere. To save some time, here's a summary of some of the thoughts floating around out there.

I don't think any one individual's opinion really matters that much; it's a divisive issue and everyone has a different take. There doesn't seem to be a lot of consensus behind cause or effect, fingers are being pointed, and the overall tone feels excessively negative.

Since Web Directions North was mentioned by name in Jason's post, I'd like to start with a few stats about our conference in no particular order, followed by my personal take. (I've run this by my three fellow organizers to make sure I'm not putting words in anyone's mouth, and though I'll stop short of saying it's the official Web Directions response, there's certainly been strong agreement between us about this issue.)

  • With 24% female speakers, Web Directions North came out with one of the higher percentages compared to other web conferences on Jason's survey.
  • That we managed 24% was a direct result of our being consciously proactive about choosing more diverse speakers. It didn't happen by chance.
  • Our organizing team was 25% female.
  • Our volunteers were 44% female.
  • Our audience was somewhere between 35 and 40% female. (That's a direct count of the registration list, but it's approximate due to some ambiguous names.)
  • WDN was a Canadian conference; we had 19% Canadian speakers. (I personally think this was our biggest problem.)
  • Our speaker list featured few visible minorities, but those from (less visible) diverse backgrounds if you cast the net a little wider. I'm not going to call out specifics, but they're there.

Compared to most others, yes, we did well. But we're not above this problem. As an accurate representation of the industry in general, I think we could have done even better.

I believe that diversity begets diversity; if you fill a speaker roster with people of a fairly homogenous background, then your audience is going to reflect that. Placing a sea of similar faces at the front of a room delivers a very poor message, which is obviously never intended, but clearly implicit. If, however, you make efforts to mix it up and draw from a wider pool of experience, people from different backgrounds will feel more welcome. I think diversity for its own sake is a valuable end goal in itself. (Not to put too fine a point on it, but this is a basic marketing consideration: it makes sense to go after a much larger potential audience, no?)

Now, we're obviously seeing that it doesn't play out that neatly in reality. So, why not? And what are we doing to fix that?

Here are some things you may not see that most certainly shape this debate. I sympathize with one of Eric Meyer's points: as a conference organizer, you tend to be conservative. You need to ensure a speaker list that will fill seats. This isn't "we want to maximize profit" filling of seats either, this is "holy crap we just signed a contract that would put us out multiple tens of thousands of dollars if we don't hit certain numbers". When you book larger venues, you make commitments and really put yourself on the line financially. Those who haven't run conferences simply can't understand what a nerve-wracking experience this is. In the end Web Directions North worked out fine on this count, but ask me some time over beer how much the entire month of December sucked for the four of us who ran it.

Now, on the flip side, I'm seeing arguments being made that the current speaker lineups probably just reflect the reality of our male-dominated industry anyway. Sorry, I don't buy it. I travel around, I've been to a lot of different conferences the past few years. The industry is far more diverse than you get to see from the quiet side of a screen. (I briefly thought of starting a Flickr pool for photos of audiences taken from the stage — of which there are many — in order to highlight this. But while it would capture what the typical conference audience looks like at this moment in time, there's also the potential that audiences are somewhat self-selecting based on who's on stage, for the reasons I mentioned above.)

I'm also seeing it being argued that content is the most important point, and damn the consequences if the best content comes from a group of white males. Well, lately I've spent a lot of time not attending sessions at the conferences I go to. It's the same people, the same experiences, the same content, and it's getting old. Some of the most interesting presentations I've watched recently come from people providing fresh perspective, people who are "outsiders" coming in to our world to share their different takes. And I want more of it.

Okay, so as a conference organizer how can we reconcile all these views? That's the trick. They seem somewhat in conflict with each other. I can't speak for anyone else's event, but I'll tell you what we did for Web Directions North.

When we came up with our speaker list, we tried to include some old faces and some new. We brought in some familiar names who may or may not have previously spoken in Canada, people we knew that our audience would be interested in hearing talk. We filled the roster with a core of proven speakers, that was our safe bet. But, as we considered our lineup, we kept asking ourselves, can we go further? Who can we bring in that might not be an obvious choice? Are we overlooking any equally good speaker choices, people who maybe we should be inviting instead that would bring a more diverse viewpoint? And if our first choices didn't pan out (more common than you'd think), are we able to make our backup choices reflect the same goals?

We considered it from a lot of angles, and what's clear is that no matter what mix you come up with, it's never going to be entirely representative. Did we feature enough women? Did we feature enough ethnic diversity? Did we feature enough religious diversity? Did we feature enough diversity of sexual preference? Did we feature enough Canadians? And so the sub-categorizing goes.

You can slice it up any number of ways and come up with many groups we didn't include. We'll never win if we head down that path.

Instead, the question to be asking is, was there enough overall diversity in our conference? Is there enough overall diversity in the industry? And if not, what can we do to encourage more?

I think we made some good strides this year, and hopefully next year we can go further. Hopefully, if everyone running a conference keeps this issue in mind as they plan their events, the overall speaker pool in this industry will shift over time to a much more inclusive group drawing from much wider experience.

Does this threaten me as a speaker? Not at all; the flip side to encouraging this sort of attitude shift means I might start going to conferences I wouldn't normally think of attending, and get introduced to people I might not otherwise meet. Not only would the speaker list in our particular corner of the industry diversify, presumably so too would other corners, which will crack open a much wider pool for those of us doing the rounds. That's exciting!

In the end, the most important point to me is simply that the debate is happening at all. Awareness of a problem is the key first step to fixing it.

(You may also want to see my fellow WDN cohort John Allsopp's take on this latest discussion, parts 1 and 2.)

Permalink › | 20 comments

Icon Design: Sizing

February 21

I've been thinking a lot about icon design lately. As a way to get the hang of writing on this site again after the post-WDN lull, I figure it might be interesting to share some of the things I've been discovering. All going well, I've got two or three more similar posts in mind.

One of the more deceptively time-consuming things you'll do when creating an icon is producing out size variations. If you require a single icon in more than one size, the time you spend designing the first size is only about two thirds of the work you'll end up doing; the other third lies in tweaking it for different dimensions.

Aha, you might say—that's what vector graphics are good for. Why not just produce the icon in Illustrator or Fireworks and resize it the quick and painless way? Because vector graphics work best with sufficiently large resolutions; scaling between 20cm and 1m at 300dpi doesn't really require much thought. When designing for the low-res pixel grid, it's a whole different ball game. Vector doesn't scale down to lower resolutions as gracefully as one might wish.

Icon Scaling Example

To help illustrate the problem, consider this example image. On the left is a large icon that needs to be scaled down. On the right are two possibilities, depending on whether the 'Scale Strokes & Effects' box is checked in Illustrator. The middle icon is clearly not an accurate version of the large icon, because the strokes have remained their large size while the rest of the proportions of the icon have been reduced. The right-hand icon is more true to the original, at a smaller size, so it would be a better choice.

But take a look at the strokes; they're quite a bit thinner now at the small size. Because they've been scaled proportionately, their thickness decreases along with the size of the icon. With this particular example, this isn't a problem. With a more detailed icon scaling to a smaller size, it is. The pixel is the smallest fundamental unit in an icon, and details that end up being smaller than a pixel cause a lot of problems.

Poor anti-aliasing when scaling icons

Figure: Calendar icon suffers from pixel blur when scaling to smaller sizes.

In the above example, I've got a calendar icon with some fine detail. Each box within the calendar has a 1 pixel wide grey border. When this icon is scaled from its starting size of 32x32 pixels to 24x24 and then 16x16 pixels, those 1 pixel wide borders end up as half-pixel wide borders, and Illustrator has to figure out how to anti-alias them. There's more explanation for why exactly this is so muddy, but I'll leave that for a follow-up post. For now, you can directly observe the effect: if you scale detailed icons so that their smallest details are less than whole pixel values, you end up with a messy result.

The solution is to start with the reduced version, and tweak it at the individual pixel level. Make the details fit within the pixel grid, remove extra detail that's causing blur, or even add extra detail if it helps you get to your end goal. Whatever it takes, the solution is to provide a critical eye to the problem and tweak until you get a result you're happy with, which is why the size variations are so much extra work.

Better anti-aliasing after tweaks

Figure: Calendar icon requires tweaks to scale gracefully.

In the calendar above, you'll notice what I've tweaked the two different sizes so the inner boxes end up with whole pixel values on either side. To do this I've had to reduce the size of the boxes at 24x24, and actually create more boxes at 16x16. I couldn't come up with a combination of 4 columns with a 1 pixel wide border that would fit within the space allotted at that smaller size, the only workable combination I found involved adding an extra column and dropping a row. The icon is a bit different than the 32x32 equivalent, but it's clearly derived from the larger one and works as an acceptable size variation.

Next time I write about this topic, I'll explain in a bit more detail the specific steps needed to produce variations and the challenges that go along with that. I wanted to set the stage first of all and describe why producing size variations was a necessity before I got into the mechanics of how you go about actually doing it. More soon.

Update: This article has been followed up with Icon Design: Anti-Aliasing.

Permalink › | 35 comments