Wednesday, November 18, 2009

Three Pictures

What do these three pictures have in common?

The first is a US Dept of Homeland Security warning. The second is a Proposition 65 warning from California. The third is from Firefox when we add _any_ add-on.

Obvious questions are:
- does anyone feel safer when they see these?
- do these actually inform one of anything?
- why are these being displayed? for whose benefit?


Fritz said...

The first is too vague to inform us about a threat that we have no control over and therefore cannot really inform our actions

The second is too ever-present and since it can be anywhere from hotel restaurants to parking garages it does not inform our actions.

The third is more useful because although it does not help inform us if the add-on is problematic it gives us a clear path to act on: 1) think about the source of the add-on 2) decide if that source was likely safe, press install if you believe it was, otherwise cancel. It also has the benefit of making stealth installation less likely.

I understand the comparison and the Firefox add-on notifier can be improved I am sure. But I think it at least provides us with a course of action because of the notifier unlike the other two signs. That being said, I have never not installed an add-on because of that notification.

John J Barton said...

None of these provide value to the reader. They all exist to embarrass the intermediate organization (airport, facility, addon author) in to taking actions they would do anyway if they are responsible and would not otherwise, warning or not. The enforcing organization feels powerful and effective when all they have done is annoy.

The first two can be ignored in less than a second. The third one jams up in your face for 5 seconds and then requires you to click a button to get rid of it.

Alex said...

But without these messages how could we subsequently blame the victim for anything that goes wrong?

Gerv said...

The first isn't useful because it doesn't change behaviour.

The middle one isn't useful because it's too commonplace and it's incomplete. If there's asbestos in the walls in an office block, no-one actually cares. If that sign only appeared on a few buildings, and instead said:

"The State of California has determined that this building has, present in the air, an overly-high percentage of birth-defect-causing chemicals"

then that might make people sit up and take notice - because it's clear that they could actually be exposed to it.

The third one could be improved, but the alternative to some sort of dialog is just directly installing the software - which makes drive-by installation a lot more likely. So we could just have a "Do you really mean this? If you did, wait 3 seconds and click OK" dialog, but that would be even more irritating. (Right, John? :-)

Jason Lustig said...

As other comments have noted, the first picture does not give clear instructions on how to change one's behavior. It is rare that a terrorist threat is equally dangerous in all places, i.e. there may be a known terrorist plot against trains but not planes.

The second picture does not give the consumer enough information either. It does not inform the consumer as to which items or elements contain harmful chemicals. If it said that item X contained the items, then the store would be better off because the customer would know that other items are OK to purchase. Otherwise, the consumer may not purchase anything at all from the venue because they do not know if an item contains dangerous chemicals or not.

The third picture also does not give the user enough information. How is a user to determine whether an extension is "trusted"? It is the same challenge that faced when one is attempting to verify the identity of an ecommerce website. One solution is to allow the signing of extensions in the same manner as websites, so you can see that it is actually produced by a particular company (and provide its contact information).

John J Barton said...

Gerv's comment points at exactly the problem in these three examples: "...could be improved...".

All three have some small value. The first can tell me if the security line at the airport will be extra long. The second could lead to caution in poking around in a facility. The third can prevent me from installing an addon by accident.

All three fail their goal because the designers did not engage with users and iteratively improve their design. But even more, all three fail because their goals are unclear or unrealizable in the real world.

In terms of cost, only the third is relatively easy to improve.