Annotation of OpenXM_contrib/pari-2.2/TODO, Revision 1.1.1.1
1.1 noro 1: This file lists the things that ought to be implemented / fixed. It is
2: hopelessly incomplete right now, but it's a start. Most items require major
3: work, and may not be that desireable after all. If you intend to work on
4: any of these things, have a look at CVS.txt and drop an email to pari-dev.
5:
6: Priority classification (first column): [subjective and debatable!]
7: 5 as soon as possible
8: 4 should be done
9: 3 nice to have
10: 2 maybe not
11: 1 probably not
12: - unclassified
13:
14: *************************** BUGS (general) ****************************
15: Unless specified otherwise, examples assume 32bit machine and default precision
16:
17: 5 setrand(5);thueinit(x^3 + x^2 - 43690*x - 3529208)
18: *** non positive argument in mplog. \\ precision problem
19:
20: 4 precision problems in polgalois:
21: setrand(7);
22: polgalois(x^8+162644002617632464507038884216211529274267271168000002)
23: gives wrong result at \p28 [32,-1,17], correct at \p38 [32,-1,15]
24:
25: 4 gcmp/gegal not transitive: [] == 0, [0] == 0, but [0] != []
26:
27: 4 rnfpolred is numerically unstable:
28: ? rnfpolred(nfinit(quadpoly(904,y)),quadray(904,1))
29: *** division by zero in gdiv, gdivgs or ginv
30:
31: 4 qfbclassno(x < 0) may give a wrong result [e.g x = -2878367]
32:
33: 4 t_SER computations much too slow when number of terms increase, even
34: when applying a trivial formula, e.g cos(x) at \ps 10000
35:
36: 3 many functions assume MAXVARN is a free variable, yielding bogus
37: results if the input involve it (only affect library programming).
38:
39: 3 quadclassunit not reliable when fed non fundamental discriminants
40: (infinite loop [e.g quadclassunit(-352)], wrong result)
41:
42: 3 ellinit over Q_2 doesn't work (AGM loses too much precision)
43:
44: 3 ellpointtoz over Q_p makes no difference between P and -P (corrected over
45: C by a cheap trick)
46:
47: 3 recursive plot easily fooled. One could split intervals in 3 + make sure
48: size of neighbouring intervals don't differ too much.
49:
50: *************************** BUGS (GP specific) ****************************
51:
52: 5 life of GP variables too short in pathological cases:
53: v = concat(v, [v=1,v=2]); concat(v, [v=1,v=2]) --> SEGV
54:
55: 4 changing primelimit from within forprime loop yields unpredictable results
56:
57: 3 after allocatemem() / default(parisize, ...), the rest of the command is
58: discarded (ends by a longjmp).
59:
60: 2 under GP after an error, memory is only recovered from "entire variables".
61: Individual components of lists/vectors/matrices are left alone if the
62: GLOBAL object wasn't modified during the last cycle (i.e only v[x] = ...
63: occured)
64:
65: 1 sum(i=1,3000, exp(-i^2)) --> overflow in R*R
66: The "negligible" terms should be neglegted, not overflow because of the
67: way the computation is done. [ Tough to correct! ]
68:
69: *************************** DOCUMENTATION ****************************
70:
71: 5 complete/update the tutorial (esp. section 11 and 12)
72:
73: 4 add examples for all functions in Chapter 3
74:
75: 4 separate chapter 4 and 5 and document the innards of PARI (entree,
76: pariFILE, bloc ...)
77:
78: 2 HTML version (pod ? SGML ? yodl ???)
79:
80: *************************** DEVELOPMENT ****************************
81:
82: 5 write decent (specialized) benches
83:
84: 3 switch to autoconf
85:
86: *************************** ALGORITHMS ****************************
87:
88: Kernel:
89: =======
90: 5 check NaN in dbltor and related routines
91:
92: 4 benchmark / profile basic functions and see what needs to be done
93:
94: 4 finish t_REAL Karatsuba and diviiexact
95:
96: 4 inline level0 routines should operate on _limbs_, not on words.
97:
98: 3 inline more assembler kernels, alpha in particular (ix86 and sparc are
99: already inline).
100:
101: 3 FFT for basic types and polynomials (say in A[X], for A = Z, F_q, ...)
102:
103: 2 add support for different multiprecision kernels (a la LiDIA)
104:
105: Misc:
106: =====
107: 5 don't assume that nf.zk is HNF. Allow using LLL-T2-reduced bases (with
108: first element 1) --> huge improvement to all idealred computations (esp.
109: bnfisprincipal, bnfinit)
110:
111: 4 graphics: allow FIG output (besides PostScript) [easier to edit!]
112:
113: 4 decent finite field package (remove all the nfxxxmodpr functions)
114:
115: 4 rnfkummer (doesn't work)
116:
117: 4 nfsubfields [use known subfields to discard blocs right away] (current uses
118: far too much memory)
119:
120: 4 LLL over R (not stable). Clean up bibli1.c (remove duplicate code)
121:
122: 4 rethink the element_mul type of functions: multiplication table is O(n^3),
123: polynomial multiplication + remainder is O(n^2) [+ can use Karatsuba/FFT]
124:
125: 3 \ and % (euclidean division) could accept more arguments (e.g t_REAL)
126: x \ y := floor(x/y) if y > 0 and floor(x/y) + 1 if y < 0
127: x % y := x - y(x\y) [ hence 0 <= x%y < |y| ]
128:
129: 3 zetak is very inefficient and unable to handle most fields of degree > 6
130:
131: 3 have quadclassunit return bnf structure
132:
133: 3 2-adic initell
134:
135: 3 p-adic ellztopoint
136:
137: *************************** LIBRARY DESIGN ****************************
138:
139: 4 a system of DEBUGLEVEL classes [e.g \g 5 "LLL" ] (such that the user can
140: easily define new ones)
141:
142: 4 allow generic functions to process and produce t_SMALL's
143:
144: 4 remove all dependences on types ordering [if (typ(x) < t_POL) ...]
145:
146: 3 integrate Ilya's code for yet bigger prime tables (pari-dev-435). Unify
147: functions using diffptr (if possible correcting the forprime bug above)
148:
149: 3 remove global variables gpi, geuler: their precision is unpredictable
150: (at least as much as requested in last const[pi|euler](), possibly much
151: more). Explicit call to mp[pi|euler] should be required. Will break
152: existing code...
153:
154: 3 rename library functions following GP names
155:
156: 3 rewrite the generic functions (gadd/gmul/gdiv plus a couple of others) to
157: use (two-dimensional) table lookups instead of switches
158:
159: 3 introduce "special" types for objects which are really containers and
160: should be defined by a secondary type (e.g number fields, finite fields,
161: p-adic fields, elliptic curves, ...). Would keep list of types small
162: (and fast)
163:
164: 3 better representations for multivariate polynomials
165:
166: 3 type "element in" ([number|finite] field...), "point" on elliptic curve
167:
168: 3 find a way to deal (generically) with "integral object + its content"
169: Application (e.g): fractional ideals much more efficient.
170:
171: 2 "mute" variables for t_POLMOD. Should have Mod(x,x^2+1) == Mod(y,y^2+1).
172:
173: 2 have some header magic (transparently) prepend some prefix (e.g "pari_")
174: before all exported functions to prevent name conflicts. Build alias
175: files for debuggers
176:
177: 2 should use vsnprintf instead of vsprintf in pariputsf. Requires GNU libio.
178: [paripusf uses static buffer which should be ok for "most" purposes]
179:
180: *************************** GP DESIGN ****************************
181:
182: 5 remove limitation to 8 arguments for static functions (argvec[9])
183:
184: 4 use divide & conquer approach in string / integer conversions
185:
186: 4 let forprime(n=a, b, ...) work even if a,b above primelimit
187: (maybe with a Warning), via simple sieve + pseudo primality test
188:
189: 3 split matrix_block in 2 parts: exploration + assignment. Would allow
190: Math::Pari to use complicated x[][][][] constructs
191:
192: 3 changevar with explicit (incomplete) permutation. E.g
193: changevar(p, [x,y], [a,b]) for x -> a, y -> b
194: add a flag to substitute in succession or in parallel (e.g if a involves y)
195:
196: 3 possibility to save and load a session (variables, functions, defaults)
197:
198: 3 a type t_FILE [current: stream re-opened/flushed/closed after every single
199: write(): disaster when one wants to write often to the same file]
200:
201: 3 add a possibility to increase the maximal recursion depth (need to
202: increase the GP process stack: use setrlimit(RLIMIT_STACK,) + fork)
203:
204: *************************** TOOLS ****************************
205:
206: 4 a script converting prototype to parser code (e.g GEN f(GEN,GEN) --> "GG")
207:
208: 3 a script to translate "legacy" GP code into something using GP2 function
209: names
210:
211: 3 write a gp-to-C compiler (adding #defines in anal.c + modifying the
212: table of functions in init.c, etc, to include the explicit function name).
213: Should be easy: the code is mostly there.
214:
215: 3 write an analog of the 'echo' default which would print code as it is
216: executed, not as it is read
217:
218: 2 write a GP scripts debugger (cf above)
219:
220: 2 write a GP scripts profiler (cf above)
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>