5. Just Obscure...

• The Jargon File, v3.0(html)

• The Decompilation Page

• Kurt Godel Society (KGS) Webpage

  More obscure than one can readily imagine. Godel set up a representation of a "general theorem construction system," the Propositional Calculus, and provable manipulations of expressions using multiples and powers of prime numbers. From this, he established that not all statements that can be represented in the Propositional Calculus can be proven to be correct by "provably correct manipulations." There are true theorems where proofs cannot be proven within the system. The addition of the necessary axioms then augments the Calculus, providing a whole new set of statements that are correct but which cannot be proven to be so within the augmented calculus. We can then reiterate, and extend this to basically say that: "Any statement system will contain statements that cannot be proven correct within that system. "

  This put the end to such projects in mathematics as Bertrand Russell's Principia Mathematica that sought to do a unified set of proofs of everything that could be important in mathematics. Godel proved that the set of "useful" proofs could not be complete.

• Recent Computer Designers

  This web page links to information about computer CPU designs and their designers from the 1970s and onwards.

• TOP500 Supercomputing Sites

• Solving Rubik's Cube Using the Bestfast Algorithm and Profile tables

• True History Of The Net (Jokester...)

• The Programmer's File Format Collection

• Programming Pearls Estimation Quiz

• John Backus

  Inventor of FORTRAN, of the Backus Naur Form (BNF), a notation used to describe the syntax of many modern programming languages, notably used to define Algol, and many subsequent families of languages including C and its descendants and Pascal and its descendants. Backus also created one of the earliest (if not the earliest) functional language called FP.

• Mark Miller's Adages

• The 24th ACM International Collegiate Programming Contest World Finals 2000

• Easter Egg Archive

• How to Publish a Great User Manual

  The rules:

  • Supply a (real) manual.

  • Explain the problem being solved

  • Present the concepts, not just the features.

  • Give 'em more than they deserve

  • Make it enjoyable to read

• IDG Sending Out Legal Documents For Dummies

  • Corporate Standover Tactics For Dummies

  • CVS for Dummies

  • HTML for morons

  • Ulysses for Dummies

  • What Lawyers Do for Fun

• Three Wrong Ideas from Computer Science

  Namely:

  • The difficult part about searching is finding enough results;!!!

  • Anti-aliased text looks better;

  • Network software should make resources on the network behave just like local resources.

  I don't completely agree with him, and would expect that the "CS professors" of the world would disagree with the notion that these are strongly held beliefs in the realm of computer science. More likely they represent the misconceptions of CS students.

  He argues fairly persuasively in decrying these concepts...

• Using Decision Tables

• Computists International

• Tabs versus Spaces

• Gazebo - the Gazette of Geek Culture

• BCD - Binary Coded Decimal

  • BCD Instructions in Assembly Language

  • BCD instructions: RISC and CISC (1/3624) (John R. Mashey)

• Condorcet-DD

  A vote tabulation and ranking system for deciding single or multi-winner contests (elections) using preference or approval ballots. Dual Dropping will compute the Schwartz Sequential Dropping (SSD) and Tideman's Ranked Pair (RP) results for each contest round and select the combined SSD and RP contest outcome with the lowest overall dropping cost. It has options to compute just the SSD or RP outcome, various optional tie-breakers, and alternative measures of defeat and dropping cost (winning votes and/or margin).

• Garbage In/Garbage Out Columns from ACM Sigplan Notices

• The Absolute Minimum Every Software Developer Absolutely, Positively Must Know About Unicode and Character Sets (No Excuses!)

• Linkers and Loaders

  This book is on a largely unsung topic; linkers and loaders are typically nearly invisible pieces of computing infrastructure vital to allowing programs to be constructed as sets of interlinking modules.

  Unfortunately, there has been relatively little real progress in this area since the 1970s. The Unix "standard," ELF, is still largely tied to the structuring required to represent Position Independent Code (PIC) for sets of C functions, as well as static data.

  Representing C++ object code requires extending this:

  • Function overloading requires having some sort of "name mangling" arrangement;

  • There must be a collection of global object initializers and destructors run at the start and end of the main routine;

  • Virtual function dispatch table; templates and inline procedures define potentially infinite families of procedures specialized by a type. vtables provide a way of accessing the numerous "actual" procedures based on what type is passed in at runtime.

  Unfortunately, since there are no standards for this, and numerous differing implementations, even between compiler versions, it is tough to maintain binary interoperability of C++ applications.

• IBM Research | Ponder this

  A site at IBM Research that publishes monthly computing puzzles

• Al Zimmermann's Programming Contests