Re: More on garbage

From: Nick Maclaren (nmm1_at_cus.cam.ac.uk)
Date: 06/11/05 

Date: 11 Jun 2005 21:35:56 GMT

In article <6b0sn2-4f2.ln1@news.rcn.com>,
Jon A. Solworth <solworth@cs.uic.edu> wrote:
>>
>> That is not the case in the real world, and the distinction is yet
>> another way that modern computer science has pulled up the drawbridge
>> of its ivory tower. In accounting and finance, auditability is
>> critical - and Baring's bank went belly-up because it didn't have
>> adequate accountability.
>
>Baring's bank went belly up because it didn't have adequate mechanisms
>for integrity. Lessing performed both trading and accounting tasks,
>very bad system design.

Let's agree on the latter! If his actions had been reported to his
management in an appropriate and timely fashion, he would have been
stopped before the bank went broke (that's the accounting). But, if
he had been prevented from committing the bank's entire finances,
the same would have happened (that's integrity). Baring's bank
lacked both, and is no more.

>> In safety-critical engineering, black
>> boxes and similar are regarded as essential (and there really is
>> no difference between reliability and security).
>
>On the contrary, there is a difference between security and reliability.
>Security has an intelligent adversary where as reliability deals with
>independent random events.

Grrk. All right, I am happy with that (and with both models, incidentally).
However, when it comes to designing for those, there is far more that is
in common than there is different. What is more, a breach of reliability
is a common cause of a breach of security, and vice versa, so it isn't
a good idea to separate them if you want to achieve either.

>You fail to understand the technical meaning of safety and liveness
>in the well established Lamport sense. This is just another example
>of the overloading of words. Because you don't know all the definitions
>of a word, it doesn't mean that my usage is somehow flawed.

I didn't say that your usage of the word is flawed, and didn't mean it.
Sorry if I caused offence that way.

What I meant is that the definitions are an example of how and why
modern computer science has diverged from the real world in general
and engineering in particular. That definition of "safety" is just
plain cuckoo, from a real-world engineering perspective, however
well-established it is in theoretical computer science.

Regards,
Nick Maclaren.

Relevant Pages

  • Re: Is MSIE dead as a browser - if Microsoft does not patch it then it is as far as I am concerned!
    ... > currently 3 identified IE vulnerabilites. ... IE has grown up a lot in speed, functionality, and reliability. ... years the patches were by and large all about more features, better speed, ... improved security simply was not what the public was crying for and the OS ...
    (microsoft.public.security.virus)
  • Re: Were losing the battle
    ... Wake me up, if you can, when the non-MF platforms can multi-task with literally thousands of tasks and still get reasonable work done in a reasonable time frame. ... And whether we like it or not, the MF still has very high reliability, excellent security and a pretty D*** high degree of recoverability. ...
    (bit.listserv.ibm-main)
  • Re: IBM fingerprint reader
    ... My new TP 43p is my second IBM machine with a fingerprint reader and works as designed. ... Not only the reader has turned out to be reliable but the security suite that comes with it is ironclad as well. ... IBM aside, I've been working with various forms of biometrics for a long time - oh, last 12 years or so, before the word "biometrics" meant anything to GP - and can vouch for its reliability. ...
    (comp.sys.laptops)
  • Re: Folder Permissions Question - Server 2003
    ... that deny all users access? ... As for security and how to, using GROUPS and then adding members to the ... lets say you have a network share called ACCOUNTING: ... On the ACCOUNTING SHARE you uncheck Inherit permissions, select COPY, ...
    (microsoft.public.windows.server.general)
  • Re: More on garbage
    ... > That is not the case in the real world, and the distinction is yet ... there is a difference between security and reliability. ... > design, a lot of logic goes on the diagnosis. ...
    (sci.crypt)