-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathWishlist.txt
34 lines (22 loc) · 1.62 KB
/
Wishlist.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
* Upgrade to newgem 1.5.3 and improve the website generated by 'rake website'
* Exclude zero (null) polynomials from Polynomial#derivatives
* change tests accordingly
* Multivariate class
* document it better
* make its interface more clear and intuitive
* extend its functionality to make it rivals Polynomial (#to_s, etc)
* see commented test stuff in test_multivariate.rb
* Allow Polynomial#to_s order to be reversed (e.g., x^2+1 instead of 1+x^2)
* Let the user specify a precision in substitute in case of BigDecimal argument.
* Extend .parse_term so that it can handle:
* rational numbers in Rational#to_s notation (e.g. '(1/2)')
* complex numbers in Complex#to_s notation (e.g. '(1+1i)' and '(2 + 3i)')
* scientific notated numbers automatically outside the Float range, parsing them as BigDecimal.
* Write comments/RDoc examples with BigDecimal coefficients polynomials
* Make .from_string and #to_s more compatible (such that that p=Polynomial.from_string(q.to_s) always verify p.equal_in_epsilon(q, eps).
- Currently there is an issue when one or more of the coefficients are BigDecimal, because BigDecimal#to_s uses scientific notation.
- Possible solution: let .from_string digest numbers in scientific notation.
* Newton method root finding (also for non-polynomial functions, using numerical derivative)
* Compare to poly-ruby, in terms of features, precision, speed and interface
* FFT-based polynomial multiplication
* Update comments/documentation so that Chebyshev/Legendre/Ultraspherical time complexity is dependent on polynomial multiplication complexity (FFT multiplication will thus make them faster)