Before the review, a tidbit from "97 Things Every Programmer Should
Know" - #93, I think (page 186-7):
Write Code As If you Had to Support It for the Rest of Your Life
Now to our main story.
I love the new books section of any academic library. Today, I
checked out the interesting-sounding "Data Analysis with Open Source
Tools" by Philipp Janert. I hoped to find Numpy or R covered, but the
catalog entry was pretty terse, so I had to actually walk over to find
out.
Well, this is quite a book - I'll spare the superlatives or gripes
(though I'll note that the bibliography of each section, where it
intersects with my own knowledge of what is excellent, is right on,
though at a high level at all times). If you have access to it
(Google Books?), check out the 'Workshop' sections toward the end of
most chapters. There are mini-case study intros to the following Sage
components:
NumPy
matplotlib
scipy.signal
gnuplot (I think it's in Sage?)
GSL
Sage
R
SQLite
as well as a few other tools, several Python libraries.
I note that Sage is in fact one of the programs listed (see pages
184-188), primarily for its calculus and symbolic linear algebra
abilities. It is far too long for me to type in here, but I think it
would be interesting in the light of the "+100 FAIL points" thread for
those with access to read his thoughts on Sage.
Highlights are an amusing example:
var('a,b,c')
M = matrix([[a,b,a],[b,c,b],[a,b,0]])
M.eigenvalues()
<snip huge result>
and a somewhat harsh critique of the 500 MB unpacking to 2 GB and a
"heavy-handed and ultimately unsustainable" solution as being the big
umbrella, though he "sincerely appreciates the straightforward
pragmatism of this solution". I think he is looking at it too much
from the point of view of the book's title; of course, the right tool
for the right solution; Sage has a different goal.
The author is in fact pretty positive about these tools in general,
including Sage - otherwise why write a 500 page book, weighty even by
O'Reilly standards? - but even in Appendix A, where he discusses
various tools like Matlab, R, the Enthought suite, can't seem to
really be happy about any of the options.
So worth knowing about, especially from the point of view of what it
is that quants need from software, and that some people will be
introduced to Sage for the first time from this reference.
- kcrisman
--
To post to this group, send an email to [email protected]
To unsubscribe from this group, send an email to
[email protected]
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
