User:Petréa Mitchell/A Plea For Usability

From Citizendium, the Citizens' Compendium
Jump to: navigation, search

Usability is a very young part of the computer field, so young that very few people in IT, professional or hobbyist, have even started to take it seriously yet. As someone who has taken a deep interest in it, in the spirit in which this project is founded, I will try in this essay to explain why Citizendium should care about it, and how paying attention to usability can help us to make this project better.

Facets which are particularly relevant here


If someone hands you a book which they say is a reference work, and you see that it is printed with purple ink on pink pages, with the binding across the top rather than the side, you will immediately be dubious of its authority, because you know what a reference work is supposed to look like: white paper, black text, using a font with serifs, and so forth.

Similarly, if you present someone with a Web site which doesn't look like a reference site to them, it takes an immediate hit to credibility, before they read a single word. What are the precise attributes of font, color, and layout that tell someone they are looking at an authoritative online reference? Short of a formal usability study, there's no way to know exactly, because some expectations will be unconscious. But we can bet that it looks an awful lot like Wikipedia (see below).

Because we can't know what changes will cause credibility problems without that kind of research, any changes to the interface should be done with great caution and only with a really good reason.


The other reason that books have standardized to a certain look is that it enhances readability. Black on white gives you maximum contrast. Serifs help minimize eyestrain. Providing a consistent look page after page lets the flow of reading continue uninterrupted. Overall, the point is to not throw obstacles in the path of the reader. What would be the point?

Usability is so often framed as a struggle to make things usable at all that the benefits of ease of use are often forgetten. But what's wrong with making things easier for our readers? If that text up in the corner is hard to see, what's wrong with making it visible? What's wrong with not confusing them by moving the controls around? Why is conforming to their basic expectations of how things work a bad thing?


Being a reference site, an important part of our usability is, can people find what they're looking for? In this area, Citizendium has made a big step forward with the subpages project. Pulling together all the information on a subject without having a bunch of scattered "List of such-and-such" articles, and yet still be able to have a readable article giving an overview of the subject, without excessive wads of detail, is a terrific idea.

Having information available in context is such a good idea, in fact, that it should be applied more to the workings of this project. An example: When I got here, the Big Cleanup documentation said a stub was any article of 50 words or less. Some months later, I was writing articles with the aim of making them at least more than stubs. This came to a halt in part because I was suddenly informed that the limit had already been changed to somewhere upward of 100 words. It was a few months after that that I stumbled by accident across the page giving the real limit of 150 words.

Why wasn't there an announcement of a change of policy? Okay, there probably was, somewhere, in one of the forums or mailing lists. The only way to find out about stuff like this is to monitor all of them. I know there's now a mailing list which is supposed to include announcements. But let's back up a section. What would be wrong with having an announcements page within the context of this project, which is put on every new person's watchlist, rather than having them go manage an account in yet another place, and then only see Citizendium announcements mixed in with stuff from all sorts of other parts of their life? Why was it wrong to have this information easily available in context?


An even stronger reason to argue for not segregating important areas of discussion into other forums is that with every split-- no matter how easy you make it to sign up-- you will lose a certain number of people who feel that the requirement exceeds the amount of effort they are willing to contribute. As you make your group of participants smaller and smaller, it becomes less and less representative of the overall contributor base.

There is room to argue that certain subjects are suited to being discussed off-site among the people who contribute most. But remember that one of the biggest criticisms of Wikipedia is that it allows the person with the most spare time and stamina to win.

Wikipedia is the standard

Speaking of Wikipedia, let's get back to Web site usability. One of the first principles of it is that people arrive at your site with certain expectations that they have formed through using other Web sites, and if you break those, you lose. In our case, the default model is Wikipedia as it appears to the general public.

If you are reading this as a Citizendium contributor, I know how you probably feel about that. I'm sorry. But this is the reality that we are stuck with. If there is any detail where we do not present something that is at least as easy to use as Wikipedia-- and remember that "easy to use" also generally includes "works the same way"-- we suffer by comparison.

Because people will not stick around to learn new controls. They will not memorize new terminology. They will not stick around to be educated about how we think things ought to work. They will leave. And they will be twice as hard to lure back next time.

Changes that can enhance usability

So are we stuck having to look exactly like Wikipedia? Can't we distinguish our brand in any way? Yes, we can. Wikipedia is only the baseline that we have to meet. We can find ways to make Citizendium more usable, without breaking established expectations.

Subpages are, as I mentioned, a good step in that direction. There's a lot more to develop there. (I know "Catalogs" sounds all professional and grown-up, but when someone is looking at, say, an article on an athletic competition and want to see a list of winners, they're going to look for something that says "Winners", rather than thinking "I want to look at a catalog". See above about educating readers.)

Another area with vast potential for improvement is search. In this area, Wikipedia is already outweighed by the entire rest of the Web, which has established this standard expectation:

  • A keyword (or short phrase containing one) such as "Search" or "Find"
  • Followed by an empty box
  • Followed by one button, labeled "Search"
  • Which leads to a list of results found with intelligent fuzzy matching.

The reason even Wikipedia doesn't do the last, of course, is that intelligent fuzzy matching is really, really hard to get right. But if you want to use your programming skills to make Citizendium better, I can't think of anything that would improve the user experience more.


If you have come to this essay because you are interested in usability issues in general, or if it has somehow persuaded you that you ought to know more, here are some good places to start:

  • Alertbox is the biweekly column of Jakob Nielsen, one of the top people in Web-specific usability.
  • Web Pages That Suck takes the approach of teaching by bad example. If that doesn't appeal to you, there are also some tools to help you recognize and avoid bad Web design.
  • For more general issues, The Design of Everyday Things is the classic introduction, and still depressingly relevant today. Its author, Don Norman, also has a collection of essays available at his site.