For all who are interested.
As far as I can tell:
1) The transition from nc_ini01.input to nc_ini03.input,and my
nc_ini03-a.input: breaks polynomfactor.input
2) Also the transition induces the following error: consider
z1 :=(1+z*p)* (p-1)
factorization(z1 ::NCP)
z1a : NCP := (1+z*p)* (p-1)
factorization(z1a)
With nc_ini01.input the "factorization(z1 ::NCP)" simply throws an
error; which, I guess, is okay.
With nc_ini03.input or nc_ini03-a.input it works and produces the wrong
answer for z1! Superficially it seems the ALS is also wrong.
3) My previous test harness ncpoly-a.tar.gz had a grossness in
test_init.input which I have fixed.
Attached is the corrected test_init.input
If anybody wants I will regenerate the tar.gz but, since it also passes
the item ")2", incorrect formating
is rampant. I am inclined to wait for a fix, even if it's a bailout
mode fix, but I would rather have the late conversion work. Makes a
nice commutation diagram (when closed).
Do not run polynomfactor.input in the same ws/time as nc_ini03-a.input
because I have taken "p" and used it as noncommuting variable in order
to make the entry of differential equations match the normal use of "p"
as a derivative operator. nc_ini03.input doesn't seem to have this problem.
RayR
--
You received this message because you are subscribed to the Google Groups "FriCAS -
computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
z1 :=(1+z*p)* (p-1)
factorization(z1 ::NCP)
z1a : NCP := (1+z*p)* (p-1)
factorization(z1a)
--
p_01 : NCP := a - b*x*y;
p_02 := p_01 + 2*p_01
p_02record := buildrecord(p_02 :: NCP);
p_02 :: XDP
p_02record.polFactors
p_02record.NCpass
--
--p_03 := addSTD(p_01, c*a*p_02);
p_03 := p_01+c*a*p_02;
p_03a := (p_01+p_02)*p_03;
p_03arecord := buildrecord(p_03a :: NCP);
p_03a :: XDP
p_03arecord.polFactors
p_03arecord.NCpass
--
-- Non commuting test!
nc_test := (c+a*x)*(d+b*y)-(d+b*y)*(c+a*x)
-- nc_test : NCP := (c+a*x)*(d+b*y)
nc_test_record :=buildrecord(nc_test :: NCP)
nc_test_record.polFactors
nc_test_record.NCpass
-- Confluent Hypergeometric DE
ConH := z*p*p+(b-z)*p+a
ConHrecord := buildrecord(ConH ::NCP)
ConH1 :=z*p*p +(1-z)*p -1
ConH1record := buildrecord(ConH1 :: NCP)
ConH2 :=(1+z*p)* (p-1)
ConH2record := buildrecord(ConH2 :: NCP)
ConH1record.polInput
ConH1record.polFactors
ConH2record.polInput
ConH2record.polFactors
)clear all
)read nc_ini03-a.input
-- )quietly option after read fouls up the read
--)read nc_ini03.input
--)display names
--
--
-- Factor test
--
--
factorCheck(x:XDP , y:LIST(XDP)): BOOLEAN ==
m:=1::XDP;
for i in y repeat m:=m*(i::XDP)
test(x = m)
--
--
-- NCPrecord is to be used as a template
-- Probably should be in library format
-- For now please operate on copy()'s
-- for instance see "buildrecord" below
--
NCPrecord : Record( _
polInput : Any ,_
polNCP : Any ,_
polFactors : Any ,_
NCprod : Any ,_
NCrep : Any ,_
NCpencil : Any ,_
NCpass : Any _
) :=['nil , 'nil , 'nil, 'nil, 'nil, 'nil, 'nil]
--
--
-- Put significant pieces of ALS into a list form
-- It returns a list in NPCrecord format
--
buildrecord(pol) ==
tmp_record := copy(NCPrecord)
pollyFact := factorization(pol)
pollyFacts := pollyFact :: LIST(XDP)
passFail :=
factorCheck(pol::XDP,pollyFacts::LIST(XDP))::BOOLEAN
prod :=1::XDP
for i in pollyFacts repeat _
prod :=prod*(i::XDP)
tmp_record.NCrep := representation(pol)
tmp_record.NCpencil := pencil(pol)
tmp_record.polInput := pol::XDP
tmp_record.polNCP := pollyFact
tmp_record.polFactors := pollyFacts
tmp_record.NCprod := prod
tmp_record.NCpass := if passFail then "pass" else "fail"
tmp_record
--
--)read tests.input
)read zx_test.input
