>>Hey all,
>>
>>Here's a thought (not addressed to anyone in particular)...
>>
>>Why not just mock up a couple of methods (option group, three-way checkbox, combo) and ask the users what they would like to see? Also ask someone who does training to see what their input is as to what method would be easiest to teach to newbies...
>>
>
>And do that for every logical field in your application? Seriously, the research has already been done for you, and the controls that we have are considered the clearest for their designated purposes...

Boy, I am afraid I couldn't disagree with you more here...I couldn't care less about some ambiguous "standard" written up in a Microsft MSDN library book (MS is not the only company to build a GUI operating system, so I don't necessarily care for Windows standards)...

Just addressing your statement:

- "Research has already been done for you": Really? Though I have seen plenty of standards books written by MS and others, I have never seen a single study done on what makes a GUI efficient, or widget A vs. widget B (with T-tests, distributions, etc, etc.) What exactly are you calling "research", and please point me to some if you can find it.
- "The controls that we have are considered the clearest...": Considered the clearest by whom? Well, the folks who established the standard, probably. I have _always_ been able to create something that users like better by _deviating_ from the standard. Adding a quirk here and there that, while non-standard and definitely non-glitzy, makes the users more efficient so they can go home on time -- custom programming (as opposed to sticking with generic Windows GUI elements) almost always results in a better fit for the end-user, which makes sense considering it takes longer to research and build.
- "for their designated purposes": Well, this is the kicker, ain't it? How do you propose to find the exact designated purpose without talking to the users (and yes, maybe asking them about each and every occurence of a Yes/No) and finding out what all might be needed? Not to bring a whole 'nother thread in, but in a recent discussion about .NET you stated that an app that eventually needs to go onto the Internet but was architected without the capability was probably architected incorrectly to start with. Well...then doesn't a GUI need to be talked through in-depth so that controls can meet current functionality as well as future needs?

Sure, if the purpose is crystal clear and the standard is crystal clear, then by all means, use whatever you deem to consider a standard that matches your purpose. But if either is not-so-clear, then the whole thing is wide open to interpretation as far as I see it. From what I remember of the original posting, the need is not for a simple Yes/No -- there may be additional choices like "No Answer" or "Maybe". And from what I remember of the "standard" posted from an MSDN book, the standard was even more fuzzy than the user's need.

JoeK