As much of the primary literature and the specialized secondary literature is inaccessible to most deviants this blog post will distill the current knowledge in a cheat sheet of sorts to help interested people to hone their skills and knowledge. This will be updated as new finds and time allows.
Feel free to point out mistakes and propose additions. We would like them to be referenced though.
In palaeontography, the stance is one not unlike that of palaeontological studies in that a pertinently themed illustration is a series of hypothesis/predictions about extinct biota, though conveyed solely in a visual medium. As such, research is a necessary part of the process and the results are something one should abide by with a modicum of strictness which should be dependent on how convincing the arguments proposed in those works are. As in many things, your mileage may vary.
Be aware of the phylogenetic bracket and be sure use its inferential capabilities to determine what is or is not possible/plausible. Where there is no specific fossil material you can follow closely, try and find related species in which that fossil material is known and adapt it accordingly.
A more detailed account of the current method and a proposal for a new one can be found in this poster (*.pdf) by jconway.
This series of paragraphs details some of the anatomical points that escape attention to most.
Dinosaurs cannot actively pronate their hands, i. e., have the hand palms face down/backwards by the forearm's rotation. The primitive condition of dinosaurs (and archosaurs for that matter) has the hand palms facing each other (semi-suppinated) when the humerus (upper arm bone) is pointing downwards in a neutral body pose. Sauropods, ankylosaurs(?), stegosaurs(?), and Protoceratops have their hands passively pronated, which is to say the wrist articulation changed its shape evolutionarily to allow the hand palm to face backwards.
Dinosaurs (and archosaurs on the whole too) have only three clawed fingers on the hand, the thumb, index and middle ones. The ring and little fingers are clawless and splint-like. While the theropod digit frameshift muddles this somewhat in the case of ceratosaurs and alvarezsaurs it is of note that the resulting hand has a maximum of three claws.
Primitively the dinosaurian thumb is robust and heavily clawed and diverges from the other fingers when extended.
Archosaurian claws in life are at least a third larger than their bony cores.
The integument (skin cover) of dinosaurs is open to speculation given recent finds. Any mix of scales and feathers where fossil remains or ecomorphology are ambiguous is acceptable. Exceptions are titanosaurs (probably the remaining sauropods too) and thyreophorans which were completely scaled throughout their life. Stage I feathers (filaments or quills) can be found with certainty in all coelurosaurs, the quills of ornithischians and the pycnofibers of pterosaurs being of disputable homology with these. Stage 2 feathers ("hairy" down) can be found in all coelurosaurs. Stage 3 feathers (open pennaceous vanes and branched down) and stage 4 feathers (closed pennaceous vanes) can be found in the last common ancestor of paravians (dromaeosaurs + birds) and oviraptorosaurs. Stage 5 feathers (assymetric closed pennaceous vanes and afterfeathers) are found in paravians.
Pennaceous feathers in paravian forelimbs insert along the second finger and the forearm stopping more or less at the elbow. This is true even of a good number of extant birds. Pennaceous feathers in known oviraptorosaurs are found along the hand's second finger.
The dinosaurian tail is moderately flexible on the whole, more less so in tetanurans and ornithopods, but it's very doubtful the animal could voluntarily bend the tail sharply at any place in its two distal thirds.
No tetrapod can move its humerus forward in an arc larger than 90º in relation to the shoulder blade. The femur too, cannot be moved backward more than 90 degrees from the pelvic girdle. This in quadrupedal or bipedal animals without specialized limbs and with rigid backbones means the upper arm and thigh bones do not move in their respective motion arcs past vertical to the ground. Except for hominids, elephants and sauropods, the elbow and knee are always flexed: full extension would result in disarticulation.
The position of the fleshy/horny nostrils in tetrapods in relation to the bony external nares is, with little exception, on the zone of the latter closest to the snout tip and the mouth.
External ears with an ear duct are only found in mammals, dinosaurs and probably other archosaurs, and in some lizards. Visible tympani are found in non-mammalian cynodonts and probably other such theriodonts, also in most lizards and in frogs. All other tetrapods have no external ears.