John Jones on 10 Jun 2002 16:19:13 -0700

[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: 22- polgalois()

>>>>> Karim BELABAS writes:

 > [ Cc-ed to pari-users since this might be of general interest, although it
 > discusses features of the alpha version 2.2.3 ]

 > On Sun, 9 Jun 2002, Karim BELABAS wrote:
 >> 22- INCOMPATIBILITY: polgalois(); changed 3rd component of result so that
 >> it gives the numbering among all transitive subgroups of S_n [ was ad
 >> hoc up to 7, then as described above for n >= 8 ]

 > Unfortunately, it will break existing scripts checking for a specific
 > 3-component output, and also those that check explicitly the old k to
 > distinguish between two groups, which occured only twice, in degree 6, for

 >   C_6   = [6,-1,1]     vs. S_3 = [6,-1,2]

 > and

 >   S_4^- = [24,-1,1]  vs. A_4 x C_2 = [24,-1,2]

 > Currently, the change is in effect whatever the value of the 'compatible'
 > default. It can be considered as a bug.

 > Any opinion ?

Since you asked for opinions, here is my two cents.  You are changing
a documented behavior which has been in effect for what, 10 years?
And, you are planning on doing it on a change in minor version, and
without a variable to control which form would be used.

IMHO, a change in minor version change should not break correctly
written scripts.  At least, that has been the philosophy of other
software projects I have used.  So, I think the right way to implement
a change like this would be to offer both naming schemes, default to
the old way, document the old was as deprecated, and offer a variable
to switch to the new behavior.  It would not be enough to put the
change into one of the existing categories of the compatibility
variable.  Correctly written scripts would break if they were written
to conform to the documentation.

                                 John Jones

Department of Mathematics and Statistics  Office:  PSA 729
Arizona State University                  Phone:   (480) 965-3725
Tempe, AZ    85287-1804                   FAX:     (480) 965-8119
URL:           e-mail: