Re: [Axiom-developer] magic, religion, and pamphlets

[C Y]

I'm not sure pamphlet are the wrong idea - they may not be the best of
all possible ideas, but that's something different.

Pamphlets are an expression of the "write for humans" model of
programming and documentation, and when the goal is long term usability
I think that model is correct.

"Write for humans" treats the most probable situation for someone
looking at the code - they don't know what they are looking at.  Even
as code is being written the numbers who can understand it directly are
few - this number drops sharply the older the code gets.  From this
standpoint, writing for humans (beginners if you prefer) is an obvious
default.

What has NOT been obvious, up until recently, is the need for code to
last indefinitely.  Early systems were bootstrapped using throwaway
code written once to get the machine running - most systems that
followed were so closely tied to the machines that they lived and died
with the hardware.  Abstracting the code beyond the architecture is
only a reasonable idea when the hardware resources are not the limiting
factor, and that is a recent development (historically).  Indeed, there
is a "new is better" philosophy in software that still runs strong even
today, in part because systems being replaced were not designed to last
- there was no point.

Now, with the benefit of hindsight, we see how much effort has been
spent and lost to posterity.  We see the costs of having to deal with
poor design decisions (or decisions dictated by hardware - I'm sure
both happened).  Re-inventing the wheel happens FAR too often.  Every
time a wheel is re-invented it means effort was spent which could have
been spent to solve some other, new problem.

TeX has shown us what happens when software is written to last.  TeX is
used (and unsurpassed) decades after its original conception.  No
serious competitor is in sight.  The initial investment put into the
TeX system has benefited countless people over the decades.

Axiom is converting to pamphlets because it intends to be to computer
algebra what TeX is to typesetting.  Not just a program to be thrown
away or rewritten after a few decades, but a repository of knowledge
and design that will stand the test of time.  It will evolve, like any
system, but the goal should be that changes should reflect developments
in mathematics.  There are many extensions to TeX, but they reflect
extensions to the core.  Likewise, Axiom should be a solid core on
which mathematical extensions can be built.  THAT is why pamphlet files
are important - because they serve as the detailed knowledge repository
behind the software.  They force us to consider what we are doing, and
explain it.  

Perhaps one thing that should be stressed - most of Axiom's pamphlet
files as they exist in the distribution today do not represent the
intended final form.  They are syntactically valid pamphlet files, but
they are not (with a few exceptions) literate documents.  A proper
documentation of the Axiom system should also be a working introduction
to the foundations of mathematics - as Axiom is a mathematical system,
this context is integral to its functionality and design.  The goal is
a long term one, but with any luck the benefits will be as well. 

There exists a wide variety of motivations among Axiom developers -
speaking for myself, I work on it because I see in it the best chance
for the Final CAS - a system sufficiently capable, extensible and well
designed that updating it and extending it will always be easier and
more desirable than replacing it.  It may even be possible to interface
it with formal proof systems to provide a level of rigor to rival or
exceed human correctness checking, at which point computer algebra will
assume a new importance in the research world.

Cheers,
CY

P.S.  Two unstated assumptions in the Axiom literate system are
functioning computers and understanding of the English language.  While
both almost HAVE to be assumed for the project to move forward, if we
look beyond 30 years to hundreds of years or more it becomes an
interesting question.  Will we still speak and understand English? 
Will we have functioning computers?  Outside the scope of the Axiom
project, but fascinating questions none the less.


      
____________________________________________________________________________________
Luggage? GPS? Comic books? 
Check out fitting gifts for grads at Yahoo! Search
http://search.yahoo.com/search?fr=oni_on_mail&p=graduation+gifts&cs=bz