I am an old fashined guy. I love paper books. But with a specific kind of material (i.e. programming references) I find it many orders of magnitude more proficient to use electronic books.

• Gamma, Helm, Johnson, and Vlissides's Design Patterns
• Scott Meyers' Effective C++ and More Effective C++
• SGI's STL manual
• C++: cplusplus.com, isocpp.org, reference, FAQ
• Valgrind manual
• Python documents
• Numerical Python manual
• The latest ISO C++ draft (December 1996)
• Stroustrup's Programming with exceptions
• Todd Veldhuizen: Techniques for Scientific C++ (PS)
• Saad's Iterative Methods for Sparse Linear Systems (PDF)
• Saad's Numerical Methods for Large Eigenvalue Problems (PS)

• Using the MEFPP_v2_0_FINAL CVS branch (french): utiliser-MEFPP_v2_0.html
• Using gprof for parallel applications: parallel_gprof.html
• Format of a .part file: part_file_format.html