This morning there was an interesting question in the Windows Vista Security Newsgroup. The poster had written an application that users were downloading. However, when they ran the application they received a warning dialog, like this one:
The poster wanted to remove this warning dialog to avoid confusing users.
This dialog is created because Internet Explorer, and some other applications, add a bit to the file to mark it as being downloaded from the Internet. It serves as a warning that this may be untrusted content. If the file is digitally signed, the warning does not have the red shield, and the publisher is listed in the dialog, but otherwise it stays the same. The poster asked if getting a digital certificate and signing the executable would get rid of the warning. It will not. This warning is there to warn the user. I think it is an important safety mechanism, and that, rather than trying to remove the warning, which is possible, we should help the user understand it. Therefore, here is my response:
You should definitely digitally sign the application no matter what. However, that will not remove the warning. It just will have your (or your company's) name in the dialog and won't say "Unknown Publisher."
Technically, there is a way to get rid of this warning, but it is there as a warning to end users. If you remove it here, you would also remove it for all other executables. That would put your users at significant risk. If you programmatically remove that warning, you would be responsible for putting them at significant risk; a responsibility that I am pretty sure you do not want to accept.
Rather, I would suggest that you take the opportunity to educate your users. Teach them that the warning is there so that they can assess whether they want to accept the risk involved in opening applications off the Internet. In this case, you have digitally signed the application so they can trace it to you and have assurance that they are, in fact, opening a trusted application. Anytime they get a dialog like this they should evaluate it and see if they really want to accept that risk or not. If the publisher is unknown, they have no way to tell who wrote the application, and should consider it a higher risk.
Update, April 22, 2008:
Based on the comments, is quite obvious that I was not clear enough in the post.Yes, IE adds a flag to downloaded file through alternate data streams, and there are tools that can show you those streams, and even the built-in unzip tool in Windows adds the same flag if the archive that was unzipped has the flag set. The point, however, was not how a very technically savvy user can download an advanced tool and manually review the alternate data streams, and possibly remove them. If all you want to do is remove that flag it would be far simpler, in fact, to uncheck the box in the dialog for "Always ask before opening this file"; although maybe inspecting and twiddling with alternate data streams would be more satisfying for some segment of computer users.
The point I was trying to make was that a lot of people in the tech community focus on hiding warnings from the user so that the user is not bothered, ostensibly with data they are not competent to parse. That is wrong. There are very good reasons for these warnings in many cases. Rather than trying to prevent users from seeing them we all need to do our part to help users understand what they are seeing and make appropriate decisions based on that data. That would provide a savvier user base and a more secure eco-system in the long run. We cannot keep focusing on preventing people from making risk management decisions any longer. If we do, eventually, they will realize they do not have the skills to do so, and that nobody is willing to help them aquire those skills. At that point, the eco-system will be in danger of collapse.
