|
3162
|
1 <html> |
|
|
2 <pre> |
|
2330
|
3 Octave PROJECTS -*- text -*- |
|
|
4 =============== |
|
|
5 |
|
5041
|
6 Check with maintainers@octave.org for a possibly more current copy. |
|
|
7 Also, if you start working steadily on a project, please let |
|
|
8 maintainers@octave.org know. We might have information that could |
|
|
9 help you; we'd also like to send you the GNU coding standards. |
|
2330
|
10 |
|
|
11 This list is not exclusive -- there are many other things that might |
|
|
12 be good projects, but it might instead be something we already have, |
|
5041
|
13 so check with maintainers@octave.org before you start. |
|
2330
|
14 |
|
|
15 --------- |
|
|
16 Numerical: |
|
|
17 --------- |
|
|
18 |
|
|
19 * Improve logm, and sqrtm. |
|
|
20 |
|
3713
|
21 * Improve complex mapper functions. See W. Kahan, ``Branch Cuts for |
|
|
22 Complex Elementary Functions, or Much Ado About Nothing's Sign |
|
|
23 Bit'' (in The State of the Art in Numerical Analysis, eds. Iserles |
|
|
24 and Powell, Clarendon Press, Oxford, 1987) for explicit |
|
|
25 trigonometric formulae. |
|
2330
|
26 |
|
|
27 * Make functions like gamma() return the right IEEE Inf or NaN |
|
|
28 values for extreme args or other undefined cases. |
|
|
29 |
|
|
30 * Handle complex values in fread and fwrite. |
|
|
31 |
|
|
32 * Support for lp_solve for linear programming problems. |
|
|
33 |
|
6628
|
34 * Improve sqp. |
|
2330
|
35 |
|
|
36 * Fix CollocWt to handle Laguerre polynomials. Make it easy to |
|
|
37 extend it to other polynomial types. |
|
|
38 |
|
|
39 * Add optional arguments to colloc so that it's not restricted to |
|
|
40 Legendre polynomials. |
|
|
41 |
|
|
42 * Fix eig to also be able to solve the generalized eigenvalue |
|
|
43 problem, and to solve for eigenvalues and eigenvectors without |
|
|
44 performing a balancing step first. |
|
|
45 |
|
|
46 * Move rand, eye, xpow, xdiv, etc., functions to the matrix classes. |
|
|
47 |
|
2477
|
48 * Use octave_allocator for memory management in Array classes once |
|
|
49 g++ supports static member templates. |
|
|
50 |
|
2789
|
51 * When constructing NLConst (and other) objects, make sure that |
|
|
52 there are sufficient checks to ensure that the dimensions all |
|
|
53 conform. |
|
|
54 |
|
2330
|
55 * Improve design of ODE, DAE, classes. |
|
|
56 |
|
|
57 * Extend meaning of .* to include v .* M or M .* v (where v is a |
|
|
58 column vector with the same number of rows as M) to scale rows of |
|
|
59 M by elements of v. Similarly, if w is a row vector with as many |
|
|
60 columns as M, then either w .* M or M .* w scales the columns of |
|
|
61 M. |
|
|
62 |
|
3141
|
63 * Make QR more memory efficient for large matrices when not all the |
|
|
64 columns of Q are required (apparently this is not handled by the |
|
|
65 lapack code yet). |
|
|
66 |
|
2799
|
67 * Consider making the behavior of the / and \ operators for |
|
3758
|
68 non-square systems compatible with Matlab. Currently, they return |
|
7072
|
69 the minimum norm solution from DGELSD, which behaves differently. |
|
2799
|
70 |
|
5164
|
71 --------------- |
|
|
72 Sparse Matrices: |
|
|
73 --------------- |
|
|
74 |
|
5610
|
75 * Improve QR factorization functions, using idea based on CSPARSE |
|
|
76 cs_dmsol.m |
|
5164
|
77 |
|
5610
|
78 * Implement fourth argument to the sprand and sprandn, and addition |
|
|
79 arguments to sprandsym that the leading brand implements. |
|
5164
|
80 |
|
5282
|
81 * Sparse logical indexing in idx_vector class so that something like |
|
|
82 "a=sprandn(1e6,1e6,1e-6); a(a<1) = 0" won't cause a memory overflow. |
|
5164
|
83 |
|
|
84 * Make spalloc(r,c,n) actually create an empty sparse with n non-zero |
|
|
85 elements? This allows something like |
|
|
86 |
|
|
87 sm = spalloc (r,c,n) |
|
|
88 for j=1:c |
|
|
89 for i=1:r |
|
|
90 tmp = foo (i,j); |
|
|
91 if (tmp != 0.) |
|
|
92 sm (i,j) = tmp; |
|
|
93 endif |
|
|
94 endfor |
|
|
95 endfor |
|
|
96 |
|
|
97 actually make sense. Otherwise the above will cause massive amounts |
|
|
98 of memory reallocation. |
|
|
99 |
|
|
100 The fact is that this doesn't make sense in any case as the assign |
|
|
101 function makes another copy of the sparse matrix. So although spalloc |
|
|
102 might easily be made to have the correct behaviour, the first assign |
|
|
103 will cause the matrix to be resized !!! There seems to be no simple |
|
|
104 way to treat this but a complete rewrite of the sparse assignment |
|
|
105 functions... |
|
|
106 |
|
|
107 * Other missing Functions |
|
|
108 - symmmd Superseded by symamd |
|
|
109 - colmmd Superseded by colamd |
|
|
110 - treelayout |
|
|
111 - cholinc |
|
6627
|
112 - bicg Can this be taken from octave-forge? |
|
5164
|
113 - bicgstab |
|
|
114 - cgs |
|
|
115 - gmres |
|
|
116 - lsqr |
|
|
117 - minres |
|
|
118 - qmr |
|
|
119 - symmlq |
|
|
120 - spaugment |
|
|
121 |
|
2330
|
122 ------- |
|
|
123 Strings: |
|
|
124 ------- |
|
|
125 |
|
2789
|
126 * Improve performance of string functions, particularly for |
|
|
127 searching and replacing. |
|
|
128 |
|
2330
|
129 * Convert string functions to work on string arrays. |
|
|
130 |
|
|
131 * Make find work for strings. |
|
|
132 |
|
2378
|
133 * Consider making octave_print_internal() print some sort of text |
|
|
134 representation for unprintable characters instead of sending them |
|
|
135 directly to the terminal. (But don't do this for fprintf!) |
|
|
136 |
|
|
137 * Consider changing the default value of `string_fill_char' from SPC |
|
|
138 to NUL. |
|
|
139 |
|
2330
|
140 ---------------- |
|
|
141 Other Data Types: |
|
|
142 ---------------- |
|
|
143 |
|
|
144 * Template functions for mixed-type ops. |
|
|
145 |
|
|
146 ------------ |
|
|
147 Input/Output: |
|
|
148 ------------ |
|
|
149 |
|
|
150 * Make fread and fwrite work for complex data. Iostreams based |
|
|
151 versions of these functions would also be nice, and if you are |
|
|
152 working on them, it would be good to support other size |
|
|
153 specifications (integer*2, etc.). |
|
|
154 |
|
|
155 * Move some pr-output stuff to liboctave. |
|
|
156 |
|
|
157 * Make the cutoff point for changing to packed storage a |
|
|
158 user-preference variable with default value 8192. |
|
|
159 |
|
2378
|
160 * Make it possible to load other image formats (ppm, pbm, etc. would |
|
|
161 probably be best since there are already filters to convert to |
|
|
162 these formats from others.) |
|
|
163 |
|
2330
|
164 * Complain if there is not enough disk space available (I think |
|
|
165 there is simply not enough error checking in the code that handles |
|
|
166 writing data). |
|
|
167 |
|
|
168 * Make it possible to tie arbitrary input and output streams |
|
|
169 together, similar to the way iostreams can be tied together. |
|
|
170 |
|
|
171 ----------- |
|
|
172 Interpreter: |
|
|
173 ----------- |
|
|
174 |
|
3162
|
175 * Allow customization of the debug prompt. |
|
|
176 |
|
|
177 * For the keyboard function, parse return (or quit) more |
|
|
178 intelligently so that something like |
|
|
179 |
|
|
180 debug> x = 1; return |
|
|
181 |
|
|
182 will work as expected. |
|
|
183 |
|
5008
|
184 * Handle multi-line input at the keyboard/debug prompt correctly. |
|
|
185 |
|
2601
|
186 * Fix the parser so that |
|
|
187 |
|
3081
|
188 if (expr) 'this is a string' end |
|
|
189 |
|
|
190 is parsed as IF expr STRING END. |
|
|
191 |
|
6628
|
192 * Clean up functions in input.cc that handle user input (there |
|
|
193 currently seems to be some unnecessary duplication of code and it |
|
|
194 seems overly complex). |
|
2330
|
195 |
|
|
196 * Consider allowing an arbitrary property list to be attached to any |
|
|
197 variable. This could be a more general way to handle the help |
|
|
198 string that can currently be added with `document'. |
|
|
199 |
|
|
200 * Allow more command line options to be accessible as built-in |
|
|
201 variables (--echo-commands, etc.). |
|
|
202 |
|
|
203 * Make the interpreter run faster. |
|
|
204 |
|
4325
|
205 * Allow arbitrary lower bounds for array indexing. |
|
2330
|
206 |
|
4325
|
207 * Improve performance of recursive function calls. |
|
2330
|
208 |
|
|
209 * Improve the way ignore_function_time_stamp works to allow |
|
3167
|
210 selecting by individual directories or functions. |
|
2330
|
211 |
|
|
212 * Add a command-line option to tell Octave to just do syntax |
|
|
213 checking and not execute statements. |
|
|
214 |
|
|
215 * Clean up symtab and variable stuff. |
|
|
216 |
|
|
217 * Input stream class for parser files -- must manage buffers for |
|
|
218 flex and context for global variable settings. |
|
|
219 |
|
3162
|
220 * make parser do more semantic checking, continue after errors when |
|
|
221 compiling functions, etc. |
|
|
222 |
|
2330
|
223 * Make LEXICAL_ERROR have a value that is the error message for |
|
|
224 parse_error() to print? |
|
|
225 |
|
|
226 * Add a run-time alias mechanism that would allow things like |
|
|
227 |
|
|
228 alias fun function_with_a_very_long_name |
|
|
229 |
|
|
230 so that `function_with_a_very_long_name' could be invoked as |
|
|
231 `fun'. |
|
|
232 |
|
|
233 * Allow local changes to variables to be written more compactly than |
|
|
234 is currently possible with unwind_protect. For example, |
|
|
235 |
|
|
236 function f () |
|
|
237 local prefer_column_vectors = something; |
|
|
238 ... |
|
|
239 endfunction |
|
|
240 |
|
|
241 would be equivalent to |
|
|
242 |
|
|
243 function f () |
|
3758
|
244 save_prefer_column_vectors = prefer_column_vectors; |
|
2330
|
245 unwind_protect |
|
|
246 prefer_column_vectors = something; |
|
|
247 ... |
|
|
248 unwind_protect_cleanup |
|
|
249 prefer_column_vectors = save_prefer_column_vectors; |
|
|
250 end_unwind_protect |
|
|
251 endfunction |
|
|
252 |
|
|
253 * Fix all function files to check for bogus inputs (wrong number or |
|
|
254 types of input arguments, wrong number of output arguments). |
|
|
255 |
|
|
256 * Handle options for built-in functions more consistently. |
|
|
257 |
|
|
258 * Too much time is spent allocating and freeing memory. What can be |
|
|
259 done to improve performance? |
|
|
260 |
|
|
261 * Error output from Fortran code is ugly. Something should be done to |
|
|
262 make it look better. |
|
|
263 |
|
|
264 * It would be nice if output from the Fortran routines could be |
|
|
265 passed through the pager. |
|
|
266 |
|
|
267 * Attempt to recognize common subexpressions in the parser. |
|
|
268 |
|
|
269 * Consider making it possible to specify an empty matrix with a |
|
|
270 syntax like [](e1, e2). Of course at least one of the expressions |
|
|
271 must be zero... |
|
|
272 |
|
|
273 * Is Matrix::fortran_vec() really necessary? |
|
2862
|
274 |
|
2330
|
275 * Add a command that works like bash's `builtin' command. |
|
|
276 |
|
|
277 * It would be nice to have an interactive debugger. |
|
|
278 |
|
2746
|
279 * Rewrite whos and the symbol_record_info class. Write a built-in |
|
|
280 function that gives all the basic information, then write who and |
|
|
281 whos as M-files. |
|
2439
|
282 |
|
2799
|
283 * On systems that support matherr(), make it possible for users to |
|
|
284 enable the printing of warning messages. |
|
|
285 |
|
2862
|
286 * Make it possible to mark variables and functions as read-only. |
|
|
287 |
|
3092
|
288 * Make it possible to write a function that gets a reference to a |
|
|
289 matrix in memory and change one or more elements without |
|
|
290 generating a second copy of the data. |
|
|
291 |
|
5228
|
292 * Use nanosleep instead of usleep if it is available? Apparently |
|
|
293 nanosleep is to be preferred over usleep on Solaris systems. |
|
|
294 |
|
2330
|
295 ------- |
|
|
296 History: |
|
|
297 ------- |
|
|
298 |
|
|
299 * Add an option to allow saving input from script files in the |
|
|
300 history list. |
|
|
301 |
|
3092
|
302 * The history command should accept two numeric arguments to |
|
|
303 indicate a range of history entries to display, save or read. |
|
|
304 |
|
5026
|
305 * Avoid writing the history file if the history list has not |
|
|
306 changed. |
|
|
307 |
|
|
308 * Avoid permission errors if the history file cannot be opened for |
|
|
309 writing. |
|
|
310 |
|
2330
|
311 * Fix history problems -- core dump if multiple processes are |
|
|
312 writing to the same history file? |
|
|
313 |
|
|
314 ------------------------------ |
|
|
315 Configuration and Installation: |
|
|
316 ------------------------------ |
|
|
317 |
|
2473
|
318 * Add an --enable-pathsearch option to configure to make it possible |
|
|
319 to configure and run without kpathsea. |
|
|
320 |
|
2330
|
321 * Makefile changes: |
|
|
322 -- eliminate for loops |
|
|
323 -- define shell commands or eliminate them |
|
|
324 -- verify distclean |
|
|
325 -- consolidate targets |
|
|
326 |
|
|
327 * Make it possible to configure so that installed binaries and |
|
|
328 shared libraries are stripped. |
|
|
329 |
|
3069
|
330 * Create a docs-only distribution? |
|
|
331 |
|
2330
|
332 ------------------------------ |
|
|
333 Documentation and On-Line Help: |
|
|
334 ------------------------------ |
|
|
335 |
|
|
336 * Document new features. |
|
|
337 -- history-search-{back,for}ward. |
|
|
338 -- Other stuff mentioned in the NEWS file. |
|
|
339 |
|
|
340 * Improve the Texinfo Documentation for the interpreter. It would |
|
|
341 be useful to have lots more examples, to not have so many forward |
|
|
342 references, and to not have very many simple lists of functions. |
|
|
343 |
|
|
344 * The docs should mention something about efficiency and that using |
|
|
345 array operations is almost always a good idea for speed. |
|
|
346 |
|
|
347 * Texinfo documentation for the C++ classes. |
|
|
348 |
|
|
349 * Make index entries more consistent to improve behavior of `help -i'. |
|
|
350 |
|
|
351 * Make `help -i' try to find a whole word match first. |
|
|
352 |
|
3130
|
353 * Clean up help stuff. |
|
|
354 |
|
2330
|
355 * Demo files. |
|
|
356 |
|
|
357 * As the number of m-files with octave grows perhaps a 'Contents.m' |
|
|
358 file for each toolbox (directory) would be appropriate so one |
|
|
359 knows exactly what functions are in a toolbox with a quick look. |
|
|
360 It would be best to generate information for each function directly |
|
|
361 from the M-files, so that the information doesn't have to be |
|
|
362 duplicated, and will remain current if the M-files change. It |
|
|
363 would also be best to do as much of this as possible in an M-file, |
|
|
364 though I wouldn't mind adding some basic support for listing the |
|
6627
|
365 names of all the directories in the load path, and the names of all |
|
2330
|
366 the M-files in a given directory if that is needed. |
|
|
367 |
|
2787
|
368 Also make it possible to recursively search for Contents files: |
|
|
369 |
|
|
370 help dir -- Contents from dir |
|
|
371 help dir// -- Contents from dir and all its subdirectories |
|
|
372 help dir1/dir2 -- Contents from dir2 which is under dir1 |
|
|
373 |
|
2330
|
374 ----- |
|
|
375 Tests: |
|
|
376 ----- |
|
|
377 |
|
|
378 * Improved set of tests: |
|
|
379 |
|
|
380 -- Tests for various functions. Would be nice to have a test file |
|
|
381 corresponding to every function. |
|
|
382 |
|
|
383 -- Tests for element by element operators: |
|
|
384 + - .* ./ .\ .^ | & < <= == >= > != ! |
|
|
385 |
|
|
386 -- Tests for boolean operators: && || |
|
|
387 |
|
|
388 -- Tests for other operators: * / \ ' .' |
|
|
389 |
|
|
390 -- Tests from bug reports. |
|
|
391 |
|
|
392 -- Tests for indexed assignment. Need to consider the following: |
|
|
393 o fortran-style indexing |
|
|
394 o zero-one indexing |
|
|
395 o assignment of empty matrix as well as values |
|
|
396 o resizing |
|
|
397 |
|
|
398 * Tests for all internal functions. |
|
|
399 |
|
|
400 ----------- |
|
|
401 Programming: |
|
|
402 ----------- |
|
|
403 |
|
2475
|
404 * Better error messages for missing operators? |
|
|
405 |
|
|
406 * Eliminate duplicate enums in pt-exp.cc, pt-const.cc, and ov.cc. |
|
|
407 |
|
|
408 * Handle octave_print_internal() stuff at the liboctave level. Then |
|
|
409 the octave_value classes could just call on the print() methods |
|
|
410 for the underlying classes. |
|
|
411 |
|
|
412 * As much as possible, eliminate explicit checks for the types of |
|
|
413 octave_value objects so that user-defined types will automatically |
|
|
414 do the right thing in more cases. |
|
|
415 |
|
2330
|
416 * Only include config.h in files that actually need it, instead of |
|
|
417 including it in every .cc file. Unfortunately, this might not be |
|
|
418 so easy to figure out. |
|
|
419 |
|
|
420 * GNU coding standards: |
|
|
421 |
|
|
422 -- Add a `Makefile' target to the Makefiles. |
|
|
423 -- Comments on #else and #endif preprocessor commands. |
|
|
424 -- Change error message format to match standards everywhere. |
|
|
425 |
|
|
426 * Eliminate more global variables. |
|
|
427 |
|
|
428 * Move procstream to liboctave. |
|
|
429 |
|
|
430 * Use references and classes in more places. |
|
|
431 |
|
|
432 * Share more code among the various *_options functions. |
|
|
433 |
|
|
434 ------------- |
|
|
435 Miscellaneous: |
|
|
436 ------------- |
|
|
437 |
|
|
438 * Implement some functions for interprocess communication: bind, |
|
|
439 accept, connect, gethostbyname, etc. |
|
|
440 |
|
2454
|
441 * The installation process should also install octave.el. This |
|
|
442 needs to detect the appropriate Emacs binary to use to |
|
|
443 byte-compile the .el file. Following GNU Emacs philosophy, |
|
|
444 installation would be into $(prefix)/share/emacs/site-lisp by |
|
|
445 default, but it should be selectable. |
|
|
446 |
|
2330
|
447 * The ability to transparently handle very large files: |
|
|
448 |
|
|
449 Juhana K Kouhia <kouhia@nic.funet.fi> wrote: |
|
|
450 |
|
|
451 If I have a one-dimensional signal data with the size 400 |
|
|
452 Mbytes, then what are my choices to operate with it: |
|
|
453 |
|
|
454 * I have to split the data |
|
|
455 * Octave has a virtual memory on its own and I don't have to |
|
|
456 worry about the splitting. |
|
|
457 |
|
|
458 If I split the data, then my easily programmed processing |
|
|
459 programs will become hard to program. |
|
|
460 |
|
|
461 If possible, I would like to have the virtual memory system in |
|
|
462 Octave i.e. the all big files, the user see as one big array or |
|
|
463 such. There could be several user selectable models to do the |
|
|
464 virtual memory depending on what kind of data the user have (1d, |
|
|
465 2d) and in what order they are processed (stream or random |
|
|
466 access). |
|
|
467 |
|
|
468 Perhaps this can be done entirely with a library of M-files. |
|
|
469 |
|
3136
|
470 * An interface to gdb. |
|
|
471 |
|
|
472 Michael Smolsky <fnsiguc@weizmann.weizmann.ac.il> wrote: |
|
|
473 |
|
|
474 I was thinking about a tool, which could be very useful for me |
|
|
475 in my numerical simulation work. It is an interconnection |
|
|
476 between gdb and octave. We are often managing very large arrays |
|
|
477 of data in our fortran or c codes, which might be studied with |
|
|
478 the help of octave at the algorithm development stages. Assume |
|
|
479 you're coding, say, wave equation. And want to debug the |
|
|
480 code. It would be great to pick some array from the memory of |
|
|
481 the code you're develloping, fft it and see the image as a |
|
|
482 log-log plot of the spectral density. I'm facing similar |
|
|
483 problems now. To avoid high c-development cost, I develop in |
|
|
484 matlab/octave, and then rewrite into c. It might be so much |
|
|
485 easier, if I could off-load a c array right from the debugger |
|
|
486 into octave, study it, and, perhaps, change some [many] values |
|
|
487 with a convenient matlab/octave syntax, similar to |
|
|
488 a(:,50:250)=zeros(100,200), and then store it back into the |
|
|
489 memory of my c code. |
|
|
490 |
|
2789
|
491 * Add a definition to lgrind so that it supports Octave. |
|
|
492 (See http://www.tex.ac.uk/tex-archive/support/lgrind/ for more |
|
|
493 information about lgrind.) |
|
|
494 |
|
2330
|
495 ------ |
|
|
496 Always: |
|
|
497 ------ |
|
|
498 |
|
|
499 * Squash bugs. |
|
|
500 |
|
|
501 --30-- |
|
3162
|
502 </pre> |
|
|
503 </html> |
