RE: Every time ..., was Re: General formula

[Alan Lloyd <Alan.Lloyd@xxxxxxxxxxxxxxxxxxxx>]

Thanks - comments in line.

> -----Original Message-----
> From:	David Chia 
> Sent:	Thursday, June 03, 1999 4:01 PM
> To:	ietf-pkix@imc.org
> Subject:	Re: Every time ..., was Re: General formula
> 
> > 
> > Ed - this is all too hard for me ... I cannot even predict how long
> my
> > wife will be on the phone tonight or the number of times my mobile
> will
> > drop out when using it (hands free of course) in the car.
> > 
> > As to certificate usage and lifetime - theory is useful - but
> practice
> > makes perfect. 
> > 
> > What is the lifetime of a Bus or Train ticket... It depends on the
> > attitude and hunger of the damn machines that read them...QED
> > 
> 
> You seem to have missed the point. 
	That some times happens when the point is hard to see or there
is no purpose for hitting it.

> The aim is not to predict 'exactly'
> the life time of any 'particular' certificate.  The exercise is to
> determine if collectively the phenomon exibits predictable behaviour
> and to estimate the 'expected' or the statistical maximum likelihood
> value (together with the standard error of estimation if possible) so
> that cost can be evaluated and alternatives can be compared.
	I understand that - and being involved with this stuff for the
last 100 years or so - I will always state that a looking at a data
structure in isolation is a very limited exercise. I often present to
companies  - and one line I use is that a certificate represents a
currency (within a domain) and it is verified through the use of
directory services. A currency has a lifetime according to treasury
policy.. but and instance of the currency has a lifetime that is
affected by use and application. 
	Eg. an Australian $ note (being plastic) does not last long if
hot water is poured on it.

	Ie.- its makes little sense to talk about the theory treasurey
policy when in real life - lifetime is a usage issue. 
	I have pointed out that the maths associated with the validity
of a certificate is theory. And that, in operational systems, a
certficate being valid or not - in order to deliver customer services -
is a small part of the picture. So any theoretical formula without its
application is  just theory.

	 
> When I buy hardware I usually also check the mean time between failure
> value beside cost and warrantee.
> 
	The MTBF of my car is 3 years - however when I break it -
operationally - is costs me heaps or its lifetime is shortened.
	 MTTR and Cost of repair is also useful - Shall we discuss
Integrated Logistics Support systems and Life Cycle Costs ?

	I hope this helps and regards alan

> David Chia, RMIT University