{science, python} newb seeks {calculation, formatting} advice

Basically I'm looking for help with my homework  which involves
outputting formatted scientific equations and their solutions. I'm
wondering how to more productively use python to create documents with
equations (including units), and their solutions. What I mean:

I'm starting grad school, and they're expecting me to hand in stuff
(basically chemistry and physics problems) in PDF. I didn't hafta do
that as an undergrad (which wasn't *that* long ago), so I'm whacking
together a solution. Being too busy/lazy for TeX at the moment, I'm
using OpenOffice:

+ it ain't M$ Office

+ OO Writer generates PDFs easily/cleanly OOTB

+ OO Math's equation editor is easy, and it integrates pretty well
  with Writer

- no integrated solver that I can see. (If you know how to make OO do
  this, please lemme know. I'm lazy enough that I want the word
  processor to solve my equations for me, and, though I did search a
  bit, possibly too lazy to find out how to make that happen.)

So my workflow now is like

0 setup an equation in OO (including any required unit conversions
  <yuck/>)

1 bail to python for the calculations. That I can basically cut/paste
  out of OO Math into the shell dulls the pain somewhat:

>>> import math
>>> d=1e-6
>>> V=math.pi*(d**3)/6
>>> rho_p=1000
>>> m=rho_p*V
>>> m_t=1e-11 # converting units in my head--"danger, Will Robinson!"
>>> n_p=m_t/m
>>> print '%3.2e' % (n_p)
1.91e+04

2 paste answers back into OO Math

until I have all the equations done, and I can generate my PDF. I'd
appreciate suggestions regarding easier/better ways to do this,
especially (in increasing order of appreciation)

* format coercion: e.g. is there a way that I can set something (like
  'scale' in `bc`) once so that I could thereafter just do

>>> m_t/m
1.91e+04

  and get the results in the format I want without repeated print's?
  Extra credit for making it do the calculations with specified
  significant digits.

* automating unit conversions

* doing more serious math (coming down the pipeline before too long):
  matrices, numeric differentiation and integration.

* automating solutions

TIA, Tom Roche <Tom_Roche@...>