Wikipedia:Reference desk/Computing: Difference between revisions

 

: @Steve: the Colossus didn't really have any specialised hardware for code breaking at all. Bletchley had a very clever system whereby they broke the tasks of analysing the settings of a geimschreiber link into a few very simple (very mechanical) tasks. One was a very simple (by modern standards) letter-group frequency analysis (which is just a few counters). Another was "dragging" a crib over a ciphertext (testing a range of cribs against possible wheel settings), and another was "boxing", which involved exploiting statistical similarities between related machine settings, and was again another kind of counting. Before Colossus these jobs were done by an oddbod variety of bespoke machines (one rightly named a Robinson after Heath), which had no real stored program (they just did what a loop of punched paper tape did them, and thus didn't even have any kind of real conditional logic). The Colossus proved to be faster and more flexible, but even then it was heavily attended by wrens who feverishly reconfigured it and did a lot of the functions we'd expect even the most basic computer to do itself. Fundamentally the colossus reproduced the same statistical hacks that the earlier machines had, which in turn had (briefly) been done on paper by people. It made no attempt to simulate the target machine's operation, so it was entirely unlike a dedicated crypto-breaker box like the EFF's DES-cracker. It really didn't do much more than a bit of counting and some strcmp. [[Special:Contributions/87.114.128.88|87.114.128.88]] ([[User talk:87.114.128.88|talk]]) 01:02, 11 December 2008 (UTC)

 

== Read/Write rating of CF card ==