Make your web application's interface elements obvious

March 28, 2009

This is a true story:

I periodically browse a few people's Twitter pages to look in on their tweets. Until a few days ago, when Twitter choked on one of its periodic load issues and failed to serve me any of their usual CSS styling, I did not know that the 'in reply to <whoever>' piece of the tweet status information for replies was clickable and went directly to the original tweet. This discovery was just short of electrifying, because suddenly conversations on Twitter became several orders of magnitude easier to follow.

(Previously I'd been clicking on the username of who the reply was to and browsing back in their tweets to try to find the original tweet based on tweet timestamps, which didn't always work and was in general a pain.)

There's no clue in the visual appearance of the status information that it is an active part of Twitter's user interface. To the extent that the design offers any clues, the fact that the entire status information is in small text and subdued colours signals that it is not very important. Yet something vital for practical usability is lurking there.

Presumably your web application's interface elements are there to be used (if not, one wonders why they're there at all). If you want people to use them, people need to know that they're there to start with, so you need to do something to draw attention to them; not just when the user gets their mouse pointer near enough, but all of the time.

(This is related to my old entry on disappearing links, but there what I was talking about was links in things like article text, not interface elements.)

Sidebar: my theory on why Twitter is doing this

In a spirit of charity, I can see a reason for Twitter to do this: they don't want to draw your attention away from the actual tweet text, because the tweet text is the most important thing. The tweet text is short and (almost) plain text, so anything that they do to make other things nearby more flashy will probably pull people's attention away.

(I still think that they could and should find some way to mark those interface elements; there must be some way to draw attention to them, just not too much attention.)

Written on 28 March 2009.
« The theoretical advantage of a separate /boot filesystem
Why I wrote my own bulk IO performance measurement tool »

Page tools: View Source, Add Comment.
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Sat Mar 28 00:29:33 2009
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.