Having attended a 2004 lecture in which Jack Horner told us that dinosaurs were probably all covered in feathers I find that all and any new dinosaur discoveries must be taken with a grain of salt. Though scientists have been studying dinosaurs for 100 years or so, it seems increasingly likely they've barely scratched the surface of really knowing what that era was like.

There's a comment at the end of the video, something along the lines of:

  Our old image of dinosaurs as ruthless killing machines 
  may be wrong, [...] some of them were caring parents, 
  with teenagers and younger offspring; they may have lived
  in herds; they could have been highly social.

That makes sense to me. There are evolutionary advantages to being social and sociability occurred in mammals in a relatively short time after the extinction of the dinosaurs. It would be pretty amazing if, in all the time the dinosaurs had, they never evolved to be social. Don't you think?

I've never seen a social lizard, and they've had 65 million years longer to evolve than dinosaurs.

Just yesterday, I was watching a pack of Komodo Dragons ... well, eat the corpse of some mammal. It was a video - I didn't see it in real life. But the point is: they weren't eating each other.

Well the adults are cannnibalistic, known to eat the young ones.

As far as I know ALL carnivorous snakes and lizards are potentially cannabalistic, that's why herpetologists have to be extremely careful not to keep different sized snakes or lizards together.

Have you ever seen a social bird?

Just a good example of how hard it is to tell what things are like based on remains rather than actually being there.

It's a good example if it's right.

But if it is wrong they it would also be a good example, no?

The proper definition of a species is that individuals can interbreed and have fertile offspring. It's impossible to test whether two fossils are the same species. The usual assumption is that if they look different they were different species, but that can be wrong for many reasons. Dogs, for example, are all one species, but you might well classify Chihuahuas and Labradors separately based on their remains

I think your definition is a bit off. If I recall correctly, a species is a group that does interbreed naturally and produce fertile offspring, not just can.

For full discussion, including various proposed definitions and the problems of identifying species from fossils, see http://en.wikipedia.org/wiki/Species

This is a fairly common problem in taxonomy, where to draw the line between species and subspecies. That is why you will occasionally see mention of "lumpers" and "splitters", those who include everything possible into one species, and those who break things up into multiple species rather than subspecies. (Lumpers and splitters also occur at higher levels of organization - genera, families, and even orders - in some classes and phyla.) Of course, in paleontology you also have the problem of different aged animals looking different and single oddities getting written up as species because of the extremely small sample size available.

paleontology need to get help from CSI people :)

what they do currently is no more than a educated guess, this might also apply to evolution theory in general, until we have a time machine there is no way we can prove they are/were right.