Annotation of OpenXM_contrib/pari/TODO, Revision 1.1.1.1
1.1 maekawa 1: For PARI/GP version 2.0.17
2:
3: This file lists the things that ought to be implemented / fixed. It is
4: hopelessly incomplete right now, but it's a start. Most items require major
5: work, and may not be that desireable after all. If you intend to work on
6: any of these things, have a look at CVS.DOC and drop an email to pari-dev.
7:
8: Priority classification (first column): [subjective and debatable!]
9: 5 as soon as possible
10: 4 should be done
11: 3 nice to have
12: 2 maybe not
13: 1 probably not
14: - unclassified
15:
16: *************************** DOCUMENTATION ****************************
17:
18: 5 complete/update the tutorial (esp. section 11 and 12)
19:
20: 4 add examples for all functions in Chapter 3
21:
22: 4 separate chapter 4 and 5 and document the innards of PARI (entree,
23: pariFILE, bloc ...)
24:
25: 3 clean up gphelp / TeX macros
26:
27: 2 HTML version (pod ? SGML ? yodl ???)
28:
29: *************************** BUGS ****************************
30:
31: - sum(i=1,3000, exp(-i^2)) --> overflow in R*R
32: The "negligible" terms should be neglegted, not overflow because of the
33: way the computation is done. Tough to correct, even with a way to trap
34: exceptions
35:
36: 5 nfgaloisconj still misses some conjugates. Symptom: try it on
37: polcompositum(quadpoly(11321), x^15-3*x^14-46*x^13+69*x^12+794*x^11-50*x^10-5226*x^9-4559*x^8+9682*x^7+15656*x^6+3342*x^5-5950*x^4-4386*x^3-1136*x^2-100*x+1)
38: [Hilbert class field extension]
39:
40: 5 silent overflows in codeword-handling macros: O(2^50000) --> O(2^-15535)
41:
42: 5 some doubles can overflow in rootpol.c, should use t_REAL everywhere
43: ? polroots(x^3 + 2^1024*x + 1)
44: *** underflow or overflow in a R->dbl conversion
45:
46: 4 gcmp is not transitive: [] == 0, [0] == 0, but [0] != []
47:
48: 4 mateigen([1,1;1,2]) won't succeed at default precision (28 digits).
49: Should use iterative methods.
50:
51: 3 polgalois should handle precision problems. Symptom: applied to
52: polzagier(11,0)/polzagier(1,0), gives wrong result with 38 digits
53:
54: 3 after allocatemem() / default(parisize, ...), the rest of the command is
55: discarded (ends by a longjmp).
56:
57: 3 recursive plot easily fooled. One could split intervals in 3 + make sure
58: size of neighbouring intervals don't differ too much.
59:
60: *************************** DEVELOPMENT ****************************
61:
62: 5 write decent (specialized) benches
63:
64: 3 switch to autoconf
65:
66: *************************** ALGORITHMS ****************************
67:
68: Kernel:
69: =======
70: 5 check NaN in dbltor and related routines
71:
72: 4 benchmark / profile basic functions and see what needs to be done
73:
74: 4 finish t_REAL Karatsuba and diviiexact
75:
76: 4 inline level0 routines should operate on _limbs_, not on words.
77:
78: 3 FFT for basic types and polynomials (say in A[X], for A = Z, F_q, ...)
79:
80: 2 add support for different multiprecision kernels (a la LiDIA)
81:
82: Misc:
83: =====
84: 5 (easy) modify invmod to get Jebelean extended gcd
85:
86: 4 decent finite field package (remove all the nfxxxmodpr functions)
87:
88: 4 round 4 (too slow, use modular techniques)
89:
90: 4 rnfkummer (doesn't work)
91:
92: 4 nfsubfields [use known subfields to discard blocs right away] (current uses
93: far too much memory)
94:
95: 4 LLL over R (not stable). Clean up bibli1.c (remove duplicate code)
96:
97: 4 rethink the element_mul type of functions: multiplication table is O(n^3),
98: polynomial multiplication + remainder is O(n^2) [+ can use Karatsuba/FFT]
99:
100: 3 have quadclassunit return bnf structure
101:
102: 3 have quadclassunit treat reliably non fundamental discriminants (at
103: least, make relevant checks before entering infinite loops)
104:
105: 3 2-adic initell
106:
107: *************************** LIBRARY DESIGN ****************************
108:
109: 4 a way to trap PARI errors, esp. all kinds of overflows (exponent, degree,
110: stack, ...). Something like:
111: trap(errpile, &stack_overflow_handler)
112:
113: 4 allow generic functions to process and produce t_SMALL's
114:
115: 4 remove all dependences on types ordering [if (typ(x) < t_POL) ...]
116:
117: 3 integrate Ilya's code for yet bigger prime tables. Unify functions using
118: diffptr
119:
120: 3 remove global variables gpi, geuler: their precision is unpredictable
121: (at least as much as requested in last const[pi|euler](), possibly much more).
122: Explicit call to mp[pi|euler] should be required. Will break existing code...
123:
124: 3 rename library functions following GP names
125:
126: 3 rewrite the generic functions (gadd/gmul/gdiv plus a couple of others) to
127: use (two-dimensional) table lookups instead of switches
128:
129: 3 introduce "special" types for objects which are really containers and
130: should be defined by a secondary type (e.g number fields, finite fields,
131: p-adic fields, elliptic curves, ...). Would keep list of types small (and fast)
132:
133: 3 better representations for multivariate polynomials
134:
135: 3 type "element in" ([number|finite] field...), "point" on elliptic curve
136:
137: 3 find a way to deal (generically) with "integral object + its content"
138: Application (e.g): fractional ideals much more efficient.
139:
140: 2 "mute" variables for t_POLMOD. Should have Mod(x,x^2+1) == Mod(y,y^2+1).
141:
142: 2 have some header magic (transparently) prepend some prefix (e.g "pari_")
143: before all exported functions to prevent name conflicts. Build alias files
144: for debuggers
145:
146: 2 should use vsnprintf instead of vsprintf in pariputsf. Requires GNU libio.
147: [paripusf uses static buffer which should be ok for "most" purposes]
148:
149: *************************** GP DESIGN ****************************
150:
151: 5 remove limitation to 8 arguments for static functions (argvec[9])
152:
153: 4 use divide & conquer approach in string / integer conversions
154:
155: 4 a way to trap PARI errors (esp. stack overflow). Something like:
156: trap(errpile, "GP_command_to_be_executed_when_stack_overflows")
157:
158: 3 split matrix_block in 2 parts: exploration + assignment. Would allow
159: Math::Pari to use complicated x[][][][] constructs
160:
161: 3 changevar with explicit (incomplete) permutation. E.g
162: changevar(p, [x,y], [a,b]) for x -> a, y -> b
163: add a flag to substitute in succession or in parallel (e.g if a involves y)
164:
165: 3 possibility to save and load a session (variables, functions, defaults)
166:
167: 3 a type t_FILE [current: stream re-opened/flushed/closed after every single
168: write(): disaster when one wants to write often to the same file]
169:
170: *************************** TOOLS ****************************
171:
172: 4 a script converting prototype to parser code (e.g GEN f(GEN,GEN) --> "GG")
173:
174: 3 a script to translate "legacy" GP code into something using GP2 function
175: names
176:
177: 3 write a gp-to-C compiler (adding #defines in anal.c + modifying the
178: table of functions in init.c, etc, to include the explicit function name).
179: Should be easy: the code is mostly there.
180:
181: 2 write a GP scripts debugger (cf above)
182:
183: 2 write a GP scripts profiler (cf above)
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to TODO CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [local] / OpenXM_contrib / pari