Ok, I've seen a couple of posts regarding the productivity benefits of multiple monitors.  As Rory points out, though, some of the sponsors (NEC, ATI Technologies) have a vested interest in the claims that multiple monitors increase productivity.  For example, if everyone had 2 monitors, it would effectively double their market.

Anyway, I just thought I would like to shed some light on the subject.  As tech lead on a large project at Northrop Grumman I made it so that everyone on the team had at least two monitors.  The junior developers had dual 17" monitors, two senior developers had a 21" and a 17" (I would be in this group), two had a 21" and 2 17" monitors, and one had 3 17" monitors.  The division between junior and senior developers had more to do with length of stay at the company than with actual programming ability, but such is life.

So during this project, which ostensibly started in January of 2002 but didn't take of programming-wise until March, ran through the end of the year.  During this time, I kept metrics on the number of lines of code and defect levels with an in-house bug tracking system.  Yes it wasn't great, but it was infinitely better than shared Word documents on a file share!.  I clearly noted in the metrics spreadsheet (which made it legit to management, ha!) when multiple monitors were in use.  The project was iterative, so there wasn't the typical huge number of bugs found at the end of the project to skew the results.

After multiple monitors were introduced:

• Productivity in lines of code per day increased 10%.
• Defect levels decreased by 26%.

Also as part of my work I created a basic survey to measure some of the qualitative benefits of multiple monitors.  The survey used a Likert scale (answer 1 to 5 for each statement, with 1 being disagree strongly, 3 neutral, and 5 agree strongly) to measure respondents on 8 questions.  The most important results were:

• On average, people would much rather have 2 smaller monitors than 1 larger monitor.  Nobody answered that they preferred 1 monitor over 2 even a little bit.
• Multiple monitors were most useful when the application had palettes or when 2 or 3 windows needed to be open, such as for programming/debugging.
• The biggest complaint was desk space, since all of our monitors were CRTs (no LCDs).

One of the most important points to bring up is that we scavenged all the 17" monitors from unused computers or dev servers (it's much easier to use Terminal Services or RDP anyway!), so the monitors cost the company nothing.  The only other cost was the purchase of 4 $50 video cards for the computers that had integrated video cards without 2 video outs.

After the large project was over, I eventually scavenged my way up to three 21" monitors.  Let me tell you this was an awesome setup, except for the administrative assistant calling me up every 2 weeks wondering why I had 3 monitors and 2 computers listed under my name!  I could see all my code, toolbars, application, and look up information on Google groups.  I also started to notice when certain messages would fly by in one of Visual Studio's console windows that were out of place, and then I'd scroll back down through all the messages and figure out what was wrong.  I tried to go with a fourth 21", but it was just too much.  I had a big desk, but I was having to twist my head around to see what was on the end monitors.  At that point you have to be able to stack the monitors or have thin-edged LCDs.

So that's it.  I'm not claiming that my metrics-survey study was setup the same way as the University of Utah would have done it, but then again I did not get free LCD monitors or video cards to test it out either.  If you have any questions or comments, let me know!