Trusting the internet is about as bad as trusting an old printed source. (IMHO :-) The best conclusion is based on finding all the available information, consider the source, and base a conclusion with a healthy dose of common sense. There are plaenty of "facts" out there that are simply myths or misinterpertations that have been repeated so often that they take a life of their own.
It is not the facts that change over time - just our understanding of them. (I wasn't aware of the SPng 400/455 issue until Brian posted it, but it does make sense.)
This doesn't just apply to railroad history - I have found it in genealogy research also. There is nothig wrong with making a best guess based on available information, as long as it is qualified as such.