Cube wrote: FAQs should exist but they should continually mold the official manual. Many program authors rely 100% on forums to provide the users with the operating instructions, and that's a messy and awkward and bad approach. Keyword searches in forums are a very flawed method if the user doesn't happen to know the right terms or jargon, and sometimes they just don't work anyway.
There should be FAQs but every FAQ represents a failure in the original documentation, and those failures should be continually addressed.
As far as my own FAQ requests, I've had a bunch of questions and I would never remember any of them. Probably most people wouldn't remember either. Monitoring FAQs, hence, should be a constant.
Thanks for your feedback. Sometimes I find that we have the answers to user's questions, but they are buried in the existing docs somewhere. This is *not* a criticism of our existing docs. I am simply pointing to the reality that different people may use different wording and or have a particular question that leads to the same answer already existing somewhere in our docs. Thus, I think this is the value of having multiple types of indexes that help people find answers.
However, you also make a good point that we need to have a good mechanism for quickly addressing information not already in our docs.
I tend to think Jane is very good about continuously looking for where our docs can be improved--she has put a lot of very solid effort into that. However, I also realize that sometimes we don't get direct feedback on doc improvements from our users... The "feedback" is in what FAQs people ask that they are not finding the answer to in our existing docs (and existing index) for whatever reason.