Hold off on proposing solutions

This also holds true for user feedback. Users are great for identifying problems and pain points. They're absolutely terrible at recommending improvements, for a plethora of reasons. They can't see past their own issue, they can't see far beyond the present design, they have no idea of the internals, they don't realize or properly value unintended consequences of their suggestion, and they often quite simply don't have a good sense of design. Making the most of user feedback requires filtering out the feedback itself and identifying the underlying issue that's generating the feedback. Even if your solution ends up resembling one of the suggestions, it's most helpful to understand the issues driving that suggestion.

This is only tenuously connected, but Less Wrong is, for me, much like tvtropes.com is for some people. I usually find I'm opening an exponentially increasing number of tabs, as every post has several interesting links.

Here is the point in this comment where I would usually throw out a "I resolve that" or "what I should do" kind of solution for this issue... but that would seem quite counter to the message of this entry. Hmm.

This is a standard technique in B2B sales. Instead of trying to sell a predefined product (aka a solution), have a conversation with the customer about the issues that they want to address. It's a reasonable approach for complex problems because you have no idea whether your product is the right fit until you're aware of all the little details.

This attacks something that user researcher Steve Portigal calls "solutioneering" and the phrase has stuck in my head since I saw his ixdasf talk last year. He led the audience through an exercise to show a typical client project, where we went through several rounds of breaking down a consumer product and its uses. It was even harder than I thought it would be to stick to discussing the problem rather than jumping straight to possible fixes or improvements.

This is also discussed in Methods of Rationality, chapter 25 (Yudkowsky's Harry Potter fanfic): http://www.fanfiction.net/s/5782108/1/Harry_Potter_and_the_M...

I wonder whether it really would help with just one person. I suspect part of the problem with groups is that the one who proposed a solution will take a social hit if it is rejected. So, with just you, that shouldn't come up.

How does this relate to the TDD mindset?