Is there any source where the basic facts about orbifolds are written and proved in full detail? I found the article by Satake "The GaussBonnet Theorem for Vmanifolds", but I'd like to have a more complete and modern source.
There is a very nice short paper by Andre Henriques: http://arxiv.org/abs/math/0112006 He explains different possible definitions of orbifolds and some relations between them. He also gives many good example. Ieke Moerdijk also has a nice paper, but it is a bit longer and has less examples http://arxiv.org/abs/math/0203100 One should note that there are a few ways of thinking about orbifolds:
The second way of thinking is the more modern approach and my references above are more in this line of thought. In either way of thinking, they often arise as quotients X/G where G is a compact Lie group acting on a manifold X, with G acting locally freely (all stabilizers finite). 


A great place to start is Peter Scott's beautiful paper "The geometries of 3manifolds". It goes through tons of examples of 1 and 2dimensional orbifolds; this includes those coming from spherical/Euclidean/hyperbolic triangle groups, and he gives a complete classification of 2dimensional Euclidean and spherical orbifolds. Along the way he covers all the usual tools like orbifold fundamental group, van Kampen's theorem, and Euler characteristic. Finally he uses orbifolds heavily, in the context of Seifert fibered manifolds, to explain which manifolds admit one of Thurston's eight geometries, and how. The other topics touched upon in the paper (Lie groups, connections and holonomy, group actions, foliations, group extensions, etc.) are all well worth knowing, and presented very explicitly and clearly here. (A scan is available from Scott's webpage, but it is oriented sideways and the filesize is large. Anyone with access should get the paper from the journal directly.) 


One standard and beautiful source is chapter 13 of Thurston's notes on 3manifolds, which you may find here. 


There are some recent lectures by Michael Davis called "lectures on orbifolds and reflection groups." They are available on his webpage here They seem good. 


The reference I would use is ThreeDimensional Orbifolds and Their Geometric Structures by Joan Porti, Michel Boileau, Sylvain Maillot, Panorama et Syntheses 15 (2003). It contains references for what it doesn't prove, and is extremely readable. Another choice (which I like less) is Threedimensional orbifolds and conemanifolds by D. Cooper, C.D. Hodgson, abd S.P. Kerckhoff, MSJ Memoirs, 5. (2000). 


Since Ilya mentioned orbifolds in Physics, the earliest reference I am aware of is the 1985 paper Strings on orbifolds by Dixon, Harvey, Vafa and Witten. This is one of the landmark papers of the socalled "first superstring revolution" because of the realisation that one could get a "realistic compactification" (meaning a model with 3 generations of quarks and leptons) out of a $\mathbb{Z}_3$ orbifold of a sixtorus. This orbifold admits a CalabiYau resolution. It should be mentioned in that most uses of the word "orbifold" in string theory refer to global orbifolds, so riemannian manifolds of the form $M/G$ where $M$ is a riemannian manifold and $G$ a finite subgroup of isometries. In most applications, $M$ is actually a euclidean space or a torus. The philosophy behind the use of orbifolds in string theory is that if one knows how to describe string propagation on $M$ then one knows how to describe it on $M/G$, but there are very few $M$ for which one knows how to do this. As the above $T^6/\mathbb{Z}_3$ example shows, smooth CalabiYau (and not just CY) manifolds may have orbifold points in their moduli space and it is at those points that one can obtain information from string theory calculations. 


Eh, it depends on what you need to know about orbifolds. Here are the basic facts you certainly need to know: orbifold is a smooth manifold X together with a very good finite group G action — it should be similar to the free action, meaning that it may have some isolated fixed points, but no more than that. The idea of orbifold is that it allows to do computations with singular X/G as if it was smooth. E.g. you can define a differential form on an orbifold — that would be a Ginvariant form on X, you can define Euler characteristics, etc. Now I'm not sure where you wan to go from there. Since orbifolds first appeared in physics (there's a more abstract mathematical notion of stack) you might want to read some physical papers to learn why they are relevant to string theory. Alternatively, you can search for orbifold on arxiv and find many math papers that prove some specific and sometimes generally interesting things about orbifolds. 


In terms of introductory material, Peter Scott's paper is pretty good, although as noted above, it is less modern in its exposition. There is also the first part of the book "Orbifolds and Stringy Topology" by Adem, Leida and Ruan which covers both approaches to some degree. It also covers the orbifold cohomology ring, which depending on your interests, might be of use. 


A word of warning: the "paths" on an orbifold are subtle. 

