Are there any more liberally licensed libraries that get close to the 
functionality in MPFR, I know that there are some non-GPL BigInt 
implementations out there although I don't think any match GMP's 
performance.

On Thursday, April 10, 2014 10:13:20 AM UTC-4, Stefan Karpinski wrote:
>
> There's been a lot of talk about turning various bits of functionality 
> like GMP, MPFR, FFTW and such into packages that simply happen to be 
> pre-loaded by default, making it easy to get a much more spare basic Julia 
> version. This will definitely happen over the summer. Note that OpenBLAS is 
> BSD-license, so using MKL is not necessary for non-GPL Julia.
>
>
> On Thu, Apr 10, 2014 at 10:07 AM, Isaiah Norton 
> <isaiah...@gmail.com<javascript:>
> > wrote:
>
>> Rmath too, but I think that is on the way out. All of the licenses are 
>> linked here:
>> https://github.com/JuliaLang/julia/blob/master/LICENSE.md
>>
>>
>> On Thu, Apr 10, 2014 at 10:05 AM, Jake Bolewski 
>> <jakebo...@gmail.com<javascript:>
>> > wrote:
>>
>>> As readline is now removed I think GMP (BigInt's) and MPFR (arbitrary 
>>> precision floating points) are the only GNU ibraries left if you are able 
>>> to use MKL and don't require FFTW.  Steven also working on a branch where 
>>> he provides FFT support in pure julia. 
>>>
>>> Best,
>>> Jake
>>>
>>>
>>> On Thursday, April 10, 2014 9:26:55 AM UTC-4, Jay Kickliter wrote:
>>>>
>>>> There are bits and pieces in Github issues and posts, but can post a 
>>>> definitive list of what needs to be replaced/removed to make Julia non 
>>>> GPL? 
>>>> Will any functionality be missing? From what I understand I can use MKL 
>>>> for 
>>>> some stuff. I've read that MKL has the ability to mimic FFTW, but will 
>>>> Julia use that interface?
>>>>
>>>> For the record I'm not anti-GPL. I'd like to pitch Julia to my company 
>>>> as alternative to Matlab and C++. But our customers can't accept a project 
>>>> built with GPL. It's not a problem now, but I'm looking down the road when 
>>>> Julia can be compiled in to executables.
>>>>
>>>
>>
>

Reply via email to