Computer science/Catalogs/List of seminal concepts in computer science

From Citizendium
Jump to navigation Jump to search

This is a list of seminal concepts in computer science.

  • subroutines (originally implemented with self-modifying code; Whirlwind had a really cool instruction to help in that)
  • virtual memory - also memory management in general
  • 1952: the first compiler written by Grace Hopper[1] - also programming language, interpreter
    • Plus a bunch of language features/ideas, e.g. structures, which arrived in COBOL; different calling conventions (call-by-name, call-by-reference, call-by-value, etc)
  • operating system (invention of) - also batch processing, timesharing
    • Switching processes, and support for multiple processors, by switching kernel stacks
  • threading and local variable scope, requiring use of stack frames
    • stacks, in general, preceded some of the later elucidations, depending on exactly what you mean by 'threading'
  • hierarchical file systems
  • stored program architecture
  • Von Neumann architecture (having program and data in same physical memory) (EDVAC was the first instance in a design, the Manchester University 'Baby' was the first actual operating one, and the EDSAC was the first useful machine of this type)
  • Harvard Mark I: Harvard architecture (having program and data in different physical memories)
  • 1950's (need exact date); Konrad Zuse was likely the first to propose pipelining of processor instructions for faster throughput (invention of); but I don't know what computer actually did it first
  • all the other key processor design speedups, at least the biggest concepts (each has a multiple ways of being realized)
  • compiler optimizations
  • Cross compilation: allows support for new/experimental platforms, or platforms for which a native development environment is not readily available
  • Bootstrapping: Getting it all going. Myriad examples include having an OS (Unix) written 99% in a high-level language (C), so that to port the OS to a different processor required mainly writing a C compiler for that processor; along with the item above, a precursor to what open source and embedded projects do when they have to compile their code for various hardware/OS platforms
  • hardware abstraction: the BIOS does this in an IBM compatible PC; the measure of success for a hardware abstraction can be seen in part by the portability of the code dependent upon it - maybe a better example is the earlier IBM 360 architecture, where the same virtual machine was instantiated in a number of totally different physical implementations
  • relational database model first proposed by E.F. Codd in 1969.
  • the various ways (past and present) for distributing applications across a network, or making procedure calls across a network: EDI, COM, COM+, CORBA, RPC, XML-RPC, and now XML web services (so-called Service Oriented Architecture, or SOA)
    • networking in general - early networking support was, err, primitive!
  • advent of managed code languages, such as Ada and Java, that can provide "type safety" and other benefits such as substantial software reuse by means of shared libraries

References

  1. Grace Murray Hopper (proceedings of Grace Hopper Celebration of Women in Computing conference, 1994)