First, let's all thank Robert for the work he did on the 5.9.2 release.
Second, I have placed a new version of nummod on
http://www.unk.edu/facstaff/profiles/willisb/
Get the file nummod(4).zip. There is a usage file, but no texinfo user
documentation. Here is a short tour of some new functions:
The function 'rationalize' replaces all floating point numbers with
exact rationalizations. Richard Fateman wrote the code that
does big float to rational conversion.
(%i2) rationalize(cos(0.5 * x - 0.25));
(%o2) cos(x/2-1/4)
There is a new maxima-integerp function that knows a bit more than
the 5.9.2 maxima-integerp function:
(%i3) declare(ne,even);
(%o3) done
(%i4) floor(ne/2);
(%o4) ne/2
If you are even, you are an integer:
(%i5) featurep(ne,integer);
(%o5) true
The largest addition to nummod is a simplifying version of max
and min:
(%i6) max(x,2*x,3*x);
(%o6) max(3*x,2*x,x)
(%i7) put(trylevel,10,maxmin)$
(%i8) max(x,2*x,3*x);
(%o8) max(3*x,x)
(%i9) apply('min, makelist(1/k,k,1,10000));
(%o9) 1/10000
Finally, there is a fix for the 2.5b-1 - 0.25b0 # 0 bug.
Richard Fateman provided me with the fix:
(%i10) fpprec : 18;
(%o10) 18
(%i11) 2.5b-1 - 0.25b0;
(%o11) 0.0b0
There is a compare function (this isn't new)
(%i12) compare(5,7);
(%o12) <
(%i13) compare(5,5);
(%o13) =
(%i14) compare(x,x+1);
(%o14) <
The functions max and min do a lenient test for realness:
(%i16) max(%i,5,7,8);
Nonorderable argument(s) in 'max' or 'min.' Returning a noun form.
(%o16) max(%i,8)
It can be tricked:
(%i17) max(acos(abs(x)+1), acos(abs(x) + 1) + 5);
(%o17) acos(abs(x)+1)+5
(The real domain of acos(abs(x) + 1) is empty.)
There are 10 rtest files with about 476 tests. I've tested the
new version of nummod with Maxima 5.9.2. It also works with 5.9.1,
but you'll need to install nset 1.203.
Barton