Prev: two definitions of RSA?
Next: It's not science, it's sci.
From: Simon Johnson on 31 Jan 2010 17:18
> What are you shooting up here?
>

A post like this is what happens when you decide to read a badly
written paper after consuming a large quantity of vodka.

I suggest you try it some time!

Cheers,

Simon
From: Richard Herring on 1 Feb 2010 05:47
In message <slrnhma0nj.fc0.unruh(a)wormhole.physics.ubc.ca>, unruh
<unruh(a)wormhole.physics.ubc.ca> writes
>
>Can you write a whole book in English never using the letter e. That
>would be an easy job compared to what you are asking for.

The trick is to write it in French first, then translate ;-)

--
Richard Herring
From: Tom St Denis on 1 Feb 2010 08:37
On Jan 30, 6:33 pm, Simon Johnson <simon.john...(a)gmail.com> wrote:
> There are many people in this group that will defend, until death, the
> notion that dense mathematical notion in academic papers is not only
> required but _necessary_.

I think a large part of it is they're just easily impressed with the
level at which they can obscure thought. I'm always reminded of
polynomial multiplication as an example. I've read Bernsteins papers
on the subject (don't get me wrong, the man is crazy smart) and
compare to my humble attempt at teaching the subject. My approach
uses much lower mathematical vernacular, not only because I was trying
to reach a different audience but because I *AM* less educated.

That being said we both teach the subject yet my approach means that a
high school student could understand the algorithm, whereas his
approach to the SAME material is suited for solely graduate students
in honours math.

I don't think it's that Bernstein [in this case] couldn't teach it
using a lower basis of vernacular and technique, I'm certain he could
teach circles around me in that dept. It's that it's a point of pride
to think of yourself as "the guy working on something few people
understand." He does have a reputation to maintain afterall.

> However, I want to argue that it's the modern day equivelent of
> holding mass in Latin. To those who don't know Latin it's completely
> mysterious. To those that known Latin, most of what is stated is often
> uninteresting and obvious.

Latin is nice as it's a non-political language to base things out of.
If we named things using English or French words it'd become highly
political. No political region speaks Latin so it's outside the
problem area.

> So I ask, can we not translate the existing body of the mathematics in
> to such a language and show proofs of the relevant theorems using this
> grammar? This would make mathematics accessible to people like me.

Well more so I'd love to see math papers actually explain things with
nouns and verbs (e.g., me vs. Bernstein on polynomial basis
multiplication). All too often I pick up a paper on eprint talking
about some complex multiplication or whatever and they write 3 words
then drop into a nonsense sea of symbols and what not. I usually stop
reading at that point. I mean there is a certain requirement for
vocabulary to make things practical. For example, if I had to write
"the set of vectors for which a function produces the zero vector" all
the time as oppose to saying "the null-space" or "kernel" it might get
tedious, but that doesn't mean we need to replace ALL ideas with
single words or symbols. Sometimes a bit of prose can explain what
you want to say a lot better than 30 pages of dense mathematical
diarrhea.

> For example, if I wanted to understand the proof of the theorem that
> AES is resistant to differential cryptanalysis, I could recurse
> through the proof tree until such a point that the result is obvious.

For AES it's fairly reducible once you understand branch theory.

> I suspect I will get a lot of replies saying this constitutes a
> "dumbing down" of mathematics. However, this is simply wrong. The
> proofs didn't become any easier by doing this. We just wrote them in a
> language that was machine verifiable. Each step is unambigious and un-
> bogged down in hundreds of years of notational evolution. It is
> simpler, purer, and more powerful.

I think part of the problem is there IS a gap [which I think we're
arguing doesn't need to exist] between the scientifically established
academics and potential students. It's not that we have to make
science stupider, it's that we have to stop gaming the system for
preservation of the seemingly fittest. There is a lot of value in
being able to teach, but what really makes an academic valuable is the
ability to discover. So their job preservation should be tied to that
more than their ability to hide in obscurity and appear relevant.

Ironically, imho, I think it shows better how you understand something
if you can reformulate explanations without regurgitating what you
yourself were taught. That is, if you *could* teach [or explain] what
you're working on to a high school AP student that'd show that you
really know the subject more than just hiding behind obscure
terminology and explanations that people won't dare call you on out of
a sense of decorum.

Tom
From: Kristian Gj�steen on 1 Feb 2010 16:57
Simon Johnson <simon.johnson(a)gmail.com> wrote:
>However, I feel that some people write the papers in such a way as
>they don't want them to be understood.

I don't think it's intentional. Certainly not for me. But most people
could put more effort into presentation.

--
Kristian Gj�steen
From: Tom St Denis on 2 Feb 2010 08:22
On Feb 1, 4:57 pm, Kristian Gjøsteen <kristiag+n...(a)math.ntnu.no>
wrote:
> Simon Johnson  <simon.john...(a)gmail.com> wrote:
>
> >However, I feel that some people write the papers in such a way as
> >they don't want them to be understood.
>
> I don't think it's intentional.  Certainly not for me.  But most people
> could put more effort into presentation.

There is pressure due to page count limits in journals so you can
kinda understand the terseness there. But for the extended copies
people post online to places like eprint and what not there is really
no excuse.

And of course some people are just better authors than others. I tend
to like Schneiers, Rivest, Daemen, and Vaudenays style while I find
quite a few others harder to follow (the group here knows one person I
pick on so I won't repeat his name hehehehe). I kinda wish people
wrote more like Knuth. That man is a genius, yet he can explain
things in terms of math, abstract algorithms, and even his own
platform independent MIX code.

Tom
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5
Prev: two definitions of RSA?
Next: It's not science, it's sci.