is_power and is_prime_power_with_data are inconsistent
#2352
Comments
|
On Sat, May 06, 2023 at 10:19:54AM -0700, Lars Göttgens wrote:
I would expect both methods to return _basically_ the same (up to the boolean flag). However, the order of base and exponent is different.
`is_power(::IntegerUnion)` gives first the exponent and then the base.
```julia
help?> is_power
is_power(a::ZZRingElem) -> Int, ZZRingElem
is_power(a::Integer) -> Int, Integer
Returns e, r such that a = r^e with e maximal. Note: 1 = 1^0.
```
On the other hand, `is_prime_power_with_data(::IntegerUnion)` gives after the boolean flag first the base and then the exponent.
```julia
help?> is_prime_power_with_data
search: is_prime_power_with_data
is_prime_power_with_data(q::IntegerUnion) -> Bool, ZZRingElem, Int
Returns a flag indicating whether q is a prime power and integers p, e such that q = p^e. If q is a prime power, than p
is a prime.
```
Finally, for `is_power(::QQFieldElem)`, the docstring specifies no behavior at all (although it again is exponent, base).
```julia
is_power(a::QQFieldElem) -> Int, QQFieldElem
is_power(a::Rational) -> Int, Rational
Writes a = r^e with e maximal. Note: 1 = 1^0.
```
I would propose adding the order into the docstring of `is_power(::QQFieldElem)`.
Furthermore, I think it would be less confusing to make these two functions consistent in their return semantics.
Well spotted - there is actually a 3rd point to consider... in
Oscar/example/PerfectPowers.jl
is a far superior implementation for large integers...
Do you happen to have any idea what the order is for is_power for ideals
and/or polynomials? I assume the is_prime_power_with_data is the black
sheep.
--
Reply to this email directly or view it on GitHub:
#2352
You are receiving this because you are subscribed to this thread.
Message ID: ***@***.***>
|
I guess, there is none... julia> R, x = polynomial_ring(QQ, "x")
(Univariate Polynomial Ring in x over Rational Field, x)
julia> is_power(x^3)
ERROR: MethodError: no method matching is_power(::QQPolyRingElem) |
|
Are there use cases for |
|
On Fri, May 12, 2023 at 03:02:26AM -0700, JohnAAbbott wrote:
Are there use cases for `is_power` (or similar) for polynomials or ideals?
Generic code that works over a range of (euclidean) rings.
For ideals there is a specific application
Related functions are "squarefree factorization" for polynomials (and maybe also polynomial decomposition), and radical for ideals.
Yep
I agree that the return order should be consistent! My instinct prefers returning the base first, and the exponent second. I'm not sure why `is_prime_power_with_data` returns a boolean as well as base, exponent -- that seems superfluous to me (_e.g._ set base to 0 to indicate not prime?)
because it is a "is_..." function hence return a boolean first.
The order of the other values is just wrong - I did not check when I
wrote it.
CoCoA has a function `StarRoot` which computes the `r` from `is_power`; given `n` and `r` it should be easy to compute `e` (right?) _Any suggestions for a better fn name are welcome_
it is called valuation(n, r) or, slight variation, remove
… Just for self-inconsistency: I prefer that functions whose name is `is_XYZ` return a boolean (as primary return value). I think that may be why I named the CoCoA function `StarRoot` rather than `IsPower`.
--
Reply to this email directly or view it on GitHub:
#2352 (comment)
You are receiving this because you commented.
Message ID: ***@***.***>
|
|
@fieker do you have a preference for the order of the return values? Base first or exponent first? |
|
On Sun, May 28, 2023 at 08:31:52AM -0700, Tommy Hofmann wrote:
@fieker do you have a preference for the order of the return values? Base first or exponent first?
No preference
… --
Reply to this email directly or view it on GitHub:
#2352 (comment)
You are receiving this because you were mentioned.
Message ID: ***@***.***>
|
I would expect both methods to return basically the same (up to the boolean flag). However, the order of base and exponent is different.
is_power(::IntegerUnion)gives first the exponent and then the base.On the other hand,
is_prime_power_with_data(::IntegerUnion)gives after the boolean flag first the base and then the exponent.Finally, for
is_power(::QQFieldElem), the docstring specifies no behavior at all (although it again is exponent, base).I would propose adding the order into the docstring of
is_power(::QQFieldElem).Furthermore, I think it would be less confusing to make these two functions consistent in their return semantics.
The text was updated successfully, but these errors were encountered: