Most well-determined structures (protein or nucleic acid) are publicly available in more than one database. The two most prominent are:
The primary depository for protein structures, the Protein Data Base, has a site which can be found here.
The National Center for Biotechnology Information (NCBI) maintains a protein structure database, derived from the PDB, but with the structure information in a different file type. The front door to this data base can be found here.
Not all protein structures have actually been determined - modeling protein structures based on sequence data is an active business. The Swiss-Prot database of modeled structures was the first to make such models available here.
Somewhat later, the University of California at San Francisco set up a separate model database, which is found here.
The old workhorse for this course is Protein Explorer, and you can get to it through here
The NCBI structure viewer is called Cn3D, and can be obtained here. The program does not operate from this site - you download it to your own platform, and it is then called up locally when needed.
Rhodes likes the structure viewer from the Swiss-PDB site, known as DeepView, which can be obtained here. This program allows you to change the amino acid at particular positions in a protein, and then look to see what consequences (might) ensue.
The wave of the future is probably the Chimera software, available here. This is definitely not compact or efficient software (it operates at a crawl on my old laptop), but it is very powerful, with a variety of add=on capabilities (such as docking software), and very nice ray-traced images.