As a start, it would include all well-established scientific truths such as evolution. It would also include recent history where there is little doubt about what happened.
It would not include any web pages that present the other side of the debate.
It would be WAY to arbitrary, and the harder you tried to make it work, the more it would look like what we have today.
Ultimately, a human (or multiple humans) would have to decide what was true, and what was false. I would almost be amused to see somebody ATTEMPT to do this job.
What about well-established scientific THEORIES (i'm not sure what is the definition of a scientific truth) that turn out to be false or inaccurate after decades of being well-established?
What if such a well-established theory is in the process of being disproven or replaced with a better theory - which happens to be "the other side of the debate" today?
But how do you decide that a theory has turned out to be false? Once a certain percentage of qualified scientists agree? What percentage? And how do you define a qualified scientist?
Think of all the FUD around climate change research. Do YOU know exactly which studies are valid and which ones aren't? Do you just hire climatologists to tell you? WHICH ones do you hire?
How do you define "good enough in practical cases"?
I could keep going like this.
I'm not even worrying about commercial viability, I'll make the assertion that what you're proposing to build is impossible to build. My reasoning is that you're trying to fit things without a universally accepted definition - "truth" or "scientifically accepted" - into an absolute, binary system - "include" or "exclude". That is impossible to do correctly or completely.
At best, you either end up with something fairly arbitrary, or you end up with a system which only includes the "facts" accepted by everyone, and excludes anything controversial. This ranges somewhere between useless and downright harmful, in my opinion (if this system is accepted as "truth", then new ideas have that much more of an uphill battle to be accepted).
Playing Devil's Advocate, (and for the record I believe in evolution)
"Evolution" is a big, broad weasel word. If you mean species variation due to natural selection pressures? Sure. Intra-species changes? Okay sorta, although such things are really hard to observe. The origin of life, the universe, and everything? Probably not. At least not in such broad and vague terms.
Evolution is more of a general idea of how things work which is then honed and applied specifically in different circumstances. It's not truth in the sense that Newton's Laws are truth. The broad term is much too vague to be included in a truth engine, in my opinion. It's not that it's false -- I believe the correct phrase is that it's not even false.
And you would understand this if you weren't part of the evil cabal that lives in my bathroom sink!
The search engine would return only those aspects of evolution that are known to be almost certainly true. See for example this video: http://www.youtube.com/watch?v=TUxLR9hdorI
Newton's laws are not truth but merely an approximation that suffices in most cases.
That Newton's Laws are not "truth" is exactly my point. I kind of set you up for that one, sorry.
Programmers and technicians deal in booleans -- yes or no answers. I know a couple of times I've heard people around here get frustrated when we start talking about the meanings of words. They just want simple declarative statements: evolution is true, detractors are wrong. Global Warming is true, detractors are wrong. etc.
Analysts, or people who have to translate from human languages to programming on a regular basis, understand that truth is such a slippery concept as to be almost meaningless. I can have three people in a room and have them discuss what a program is going to do until everybody is happy, and then still there are misunderstandings. Language is just too slippery for universal truth operations.
Or put another way, save universal truth for the philosophers.
