Poser is primarely a human figure composition tool. With poser you buy the ability to make human figures, dress them, put hair on their heads and props in their hands and then pose them in a certain way. This gives you as result a "tin soldier" that you can export and use in another more mature program such as 3DS Max or Lightwave. Building a human figure is what Poser excels at, not at rendering or animating.

Poser is easy to learn if you stick to this human figure approach. Poser is more difficult if you wish to dig into the more advanced technologies that it covers: file formats, UV-mapping, CR2-editing, Python scripting etc.

Another often overlooked difference between Poser and other 3D modelling tools is that poser does not implement a true XYZ-world as e.g. 3DS Max. The approach of Poser is a bone-figure-relative-addressing method. This means that movements of the bones of the figure are dictated by relative position of a master-bone (slave-master as in a hierarchy). This can give odd results when trying to set-up a complex scene with many props. That is for me a big reason to split activites in Poser and other 3D applications based on my previously mentioned "tin soldier" approach.

Hope this makes sense...

BTW, did you known that Maya is used by ILM for Star wars II ?