Gary Neitzke and I both discovered the same "bug" with OneNote last week. We both entered a bug with repro steps (the step by step directions on how to get a bug to occur) like this:
The body of the page is white
The color selected in the dropdown to be applied.
Other details: I was using WinXP, Gary Vista. We both had flat screen LCD monitors, and I had one machine with a GeForce video card, and another with Nvidia. Screen resolutions varied, and color depths tested were 16 and 32 bit color.
Pretty straightforward. And since the bug was seen by at least 2 different people on at least 2 different machines, I was pretty sure this was going to be easy to track down. Color bugs depend on human eyesight to some extent, and it was doubly reassuring that 2 different sets of eyes were seeing the same behavior.
The dev who was looking at this could not see the problem. He had to swing by my office to take a look, and it was pretty easy to see there. Thank goodness for the solid repro. Easily the most frustrating aspect of testing is intermittent bugs: they won't always happen, but they DO exist and are very hard to trace. It's the exact same thing as taking your car to the shop because it is making a noise and, naturally, having it work perfectly once there.
The odd thing that stood out was step (4) above saying that teal was visible. The dev was stumped, but since was running Windows 2003 server, he thought maybe the video was different enough that he needed more information. He came back and I switched over to the computer which I was using to show the bug. He remarked that the screen was very bright, and decided to adjust the brightness to better suit his eyes. Voila, the "bug" went away. The brightness setting of the video cards was so high on our machines that combined with the LCD monitors, we could not see the colors at all. Turning down the brightness fixed the problem, and we resolved the bug as "External".
Up to this point, this is a good tip for OneNote users: double check the brightness setting of your video card to get better color settings.
But the testing continues. I was still unsure if this is a bug or not. We have a "workaround" to the apparent problem so we can keep working, but still have not determined why the problem exists. We can safely say it is not a code bug but it could be a design problem. We might have picked 15 of the worst colors to use for page coloring that just happen to be most susceptible to being washed out on modern monitors.
So I went to a CRT monitor next. Things looked good there, but that was one single test on a single video card and montior. I went to our config lab verified our colors on a much wider variety of hardware. The colors looked a washed out on some, but most were good. I had to look for commonality on the problem machines vs. the machines that worked as expected. Getting into the properties dialogs of the display settings showed that on the machines that worked best, a monitor driver had been installed, and on the machines that did not, the monitor drivers had NOT been installed.
Back to my office I went, and sure enough, there was no monitor driver installed on my test machine. Off to the Dell site to install the driver, and it made another small difference. Now my colors looked nice and distinct on each machine.
Lesson learned: update all my drivers before entering video related bugs. Then double check to make sure the machine is in the state expected.
The last thing I have to do is check the Nvidia and GeForce support forums to see if this is known, happens only to us or is wider spread, and if they are starting to turn down the default brightness as more people switch from CRTs to LCDs.
Questions, comments, concerns and criticisms always welcome,