Other Directory Sites: SeekWonder | Directory Owners Forum
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:18:03
Marcus Watts (mdw@umich.edu)
Wed, 13 Jan 1999 16:23:50 -0500
You wrote:
> In a response to a post of this list, I suggested an alternative
> term 'XYZ stream encryption sequence'. In sci.crypt, where one person
> vehemently objected to my terminology, I suggested an alternative
> term 'intended approximation to an ideal OTP'. In both cases I
> haven't yet obtained feedback. What's your opinion to these?
> Or do you have a better suggestion?
>
> BTW, what interests me personally more is the question: Would the use
> of an alternative term lead to substantial motivations of the readers
> to contribute discussions on the topic? I guess that this issue is
> probably at least as essential as the aforementioned.
XYZ stream encryption sequence - wordy, but acceptable
intended approximation to an ideal OTP - Bad.
Most of my encryption library is at home, but I think there's
actually a somewhat better term than "stream cipher". The term
"stream cipher" may be a bit generic to some, and may cover other
small block size fast running algorithms that do something more
complicated than just x'oring the data with a key stream. The
problem is, I don't remember what the better term is. "key-autokey
synchronous"? Perhaps someone else will remember.
The reason to be careful about terms is this: in english, words have 2
meanings: "denotative" - what they mean literally, and "connotative",
kind of the emotional "what they imply". This can be seen in words
such as "lady, girl, wench, female, woman, madam, wife" - all of which could be
applied to the same person, but would almost certainly produce a different
reaction in the listener. In different listeners, the same words
can produce a different reaction. So, to produce the desired reaction,
you have to know your listener, or at least have a pretty good idea how
they might react. It should be noted that English, with its large
vocabulary, is much more prone to this "problem" than is, for instance,
French, which has one word, "femme", which covers all of the above
meanings and is nearly always the only choice. (Writers tend to
regard this as a "feature" not a "bug" of English.)
With mathematics, some words have a second more strict "mathematical"
interpretation. It's tempting to think these aren't really "English",
but part of some universal language. Unfortunately, it's not quite that
simple. Different people can use the same term to mean different things,
so, to be really rigorous, it's always wise to define your terms.
When mathematicians talk about some common subject, it is usual to try
to arrive at a consensus regarding a common vocabulary of words and
what they mean, and for all of the participants to try to use that same
language. This is easiest when the subject is something that is old
and well understood. On the other hand, when something new is introduced,
it is very often the case that the person who first describes them uses
words and terminology that is different than what is ultimately accepted.
This can be see with Newton, for instance, when he wrote about calculus.
An "approximation to an OTP" may mean to you what "stream cipher" means
to many people. But, when you use the term "OTP', people instantantly
think "snake oil". That's the most common usage, and hence, that's the
connotative meaning of OTP. Mathematicians and cryptologists are people
too, and they react just as emotionally as anyone else, all appearance
of logic aside. If you want to talk to these people, and you don't
want to engage their emotional snake-oil reflex, you will want to avoid
that dangerous term, OTP.
-Marcus Watts
UM ITD PD&D Umich Systems Group
The following archive was created by hippie-mail 7.98617-22 on Sat Apr 10 1999 - 01:18:03