Re: "Perfect" or "Provable" security both crypto and non-crypto?

From: Paul Rubin (//phr.cx_at_NOSPAM.invalid)
Date: 09/14/04 

Date: 14 Sep 2004 13:09:21 -0700

rpw3@rpw3.org (Rob Warnock) writes:
> You've already gotten several other answers to this, but generally,
> the techniques tend to get used occasionally (a fraction of the time)
> by those familiar with them (a fraction of coders) on difficult algorithms
> (a fraction of problems), when the coder really, *really* wants to
> get it right. For example, I used it once on a portion of embedded
> firmware having to do with allocating ATM time-slots for a digital-
> media server. It took a while, but the resulting code "just worked"
> [and the last time I checked, was still in use 7 years later].

The fact is, "difficult algorithms" comprise about 0.001% of the
actual code in real-world software. And oddly enough, those
algorithms are often the easiest to use formal methods on, since you
can succinctly express what they're supposed to do (e.g. a sorting
algorithm is supposed to make an output vector with a particular
ordering relation between the elements). It's a heck of a lot harder
to succinctly express what a digital media server (taken as an entire
body of software) is supposed to do. In fact the specification is
likely to be as buggy as the code.

Relevant Pages

  • Secret Sharing/Key Splitting
    ... private keys, obviously with redundant information where you only need ... a certain fraction of the total number of splits to recreate the ... However, I can find very little information on these algorithms, can ...
    (sci.crypt)
  • Re: How to dissect algorithms.
    ... counting loop iterations and multiplying if they are nested. ... There are a variety of techniques that apply in particular ... At the serious algorithms geek level, ... recursive algorithms is one for "divide and conquer" algorithms that ...
    (comp.lang.scheme)
  • Re: Structural Dynamics
    ... My question is should a student learn any solution techniques like ... and what practical relevance it has. ... is not necessary to know the algorithms for calculating it. ... have enough knowledge in practical engineering techniques. ...
    (sci.engr.civil)
  • Re: can i learn OOD without any procedural 1st
    ... algorithmic processing on computers because they grew up in academia ... techniques, more than for applying procedural techniques in a computing ... That's because the OO paradigm is focused on problem space ... individual algorithms to mountains of data. ...
    (comp.object)
  • Re: can i learn OOD without any procedural 1st
    ... algorithmic processing on computers because they grew up in academia ... My point here is that there is a substantial learning curve in adopting OO techniques, more than for applying procedural techniques in a computing environment. ... That's because the OO paradigm is focused on problem space abstraction rather than hardware computational models. ... IT is characterized by applying lots of quite trivial individual algorithms to mountains of data. ...
    (comp.object)
Quantcast