1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | ||
Editor: test1
Time: 2015/06/08 17:14:38 GMT+0 |
||
Note: |
changed: - check this list of Common [Axiom Problems]. check this list of Common [FriCAS Problems].
Frequently Asked Questions
You've got questions? We've got answers ... :)
Finally, if you can not find the answer to your questions here, then please send an email to:
"http://groups.google.com/group/fricas-devel?hl=en"
Note: please sign up before posting a message.
Or simply press the "edit":/FAQ/editform button at the top right of this page and enter your question to this site. Yes, edit this page.
Probably some hours later someone will have updated this page to include an answer to your question.
FriCAS? is free software, available under a BSD like license. For more details, please have a look in the "licences available in the SVN repository":http://sourceforge.net/p/fricas/code/HEAD/tree/trunk/license/
Contact FriCAS? "mailing list":"http://groups.google.com/group/fricas-devel?hl=en". Also look at FriCASProgramming?.
Future is in your hands. Developement plans include better algorithms, in particular handling of conditions in expressions, improvements to FriCAS? language, nicer user interface. Long term goal is to remove unsound computations (yes there are parts which produce useful, but not always correct results) and formal verification of the rest. What actually will be developed depends on you: code contributions, tests, bug reports and feedback help.
FriCAS? is developed mainly on Linux and should run on wide variety of Linux distributions. There is regularly build Cygwin and Mac OS X port. FriCAS? is in BSD ports collection and should build on many BSD flavours.
In the past people have compiled and run FriCAS? on Solaris, but there are no recent reports.
Look at "articles":http://axiom-wiki.newsynthesis.org/public/refs/articles.html on this website.
Click at _edit_ in the upper right corner of the corresponding page and follow the Wikipedia guideline of "being bold":http://en.wikipedia.org/wiki/WP:Be_bold_in_updating_pages in updating wiki pages. Yes, don't ask, just update (i.e. improve) the page.
Look at [Creating New Pages]?.
Look at FriCASBinaries?.
Report FriCAS? bugs in "bug tracker":http://sourceforge.net/p/fricas/bugs/ If you have found problen with this site, then the [Issue Tracker]? is the right place to go. When reporting bugs remember that the most important thing is _exact_ way to reproduce the problem, that is seqence of commands leading to wrong result. It is best to cut and paste commands or attach a file to avoid transcription error. Tell us about symptoms you see -- your command on different machine may produce different (correct) result. In such case others will not see the bug.
Post unified diff on FriCAS? "mailing list":"http://groups.google.com/group/fricas-devel?hl=en".
The short answer is that in a strongly typed system like FriCAS? handling definite values is a different problem from handling variables. To handle "indefinite" values you need expressions. FriCAS? currently have Expression domain, which can handle some problems. In particular Expression has quite advanced machinery to handle expression taking values in differential fields of characteristic 0 (in other words calculus expressions without inequalities). But for specific problem you may need something else, possibly to write a new domain (either based on Expression or entirely new).
Lists work differently than arrays. Array structure does not change, only differenet elements get assigned to given position. Each list node has link to next node. Destructive operations on lists may assign new values to links and change list structure. In particular new list may have different first node than old one. So after destructive operation one should always use returned result, because old list may look scrabbled.
Floating point arithmetic is inherently approximate. It is unusual for floating point operations to deliver exact result. Normally results are approximate and are considered correct if error stays within prescribed bounds. Most operations attempt to provide result with error less than half of least significant digit. However, when computing values of functions close to zeros accuracy is normally limited by accuracy of the argument and delivering more accurate result is simply wasted work. So FriCAS? uses more relaxed criterion: result should be withing half of least significant digit to some value of function when argument is varied by at most half of least digit. To achieve this the Float domain may increase accuracy in intermadate calculations. DoubleFloat? domain uses machine floating point operations, which have fixed precision. This limits precision of more complex operations, in particular error may by some (hopefully small) multiple of least significant digit.