Re: [julia-users] New Package Looking for Contributors: VectorizedRoutines.jl

[Mauro]

On Fri, 2016-07-15 at 17:52, Chris Rackauckas  wrote:
> Thanks for pointing that out. Never heard about that one. It seems like its
> been dead for quite awhile and they haven't answered any issues in almost a
> year, and haven't committed in 8 months? If anyone is on that project,
> please join the thread.
>
> I think there are some major differences though. While they are targeting
> MATLAB porting, I am looking for just the useful vectorization routines.
> The large difference comes from the fact that since they have a bunch of
> tools for working with images and the like, they have a lot of dependencies
> (including Tk). I think that makes the package much less useful, especially
> since if you want to set it to precompile you'll probably run into issues
> if you don't want to deal with some of the dependencies (they don't have
> precompile enabled?). However, I think that some of the functions they
> implement would do well moving over to VectorizedRoutines.jl, which they
> could then import to their larger project.

I see.  I never looked at the package but remembered that it was around.
If there is no response from them then you should just use their useful
bits in your package.  M

> On Friday, July 15, 2016 at 12:43:04 AM UTC-7, Mauro wrote:
>>
>> There is https://github.com/MatlabCompat/MatlabCompat.jl, which seems to
>> have a similar aim (although sans R/Python support).  Maybe join forces?
>>
>> On Fri, 2016-07-15 at 01:34, Chris Rackauckas > > wrote:
>> > Hey,
>> >   After some discussion (and letting the idea hang around for a long
>> time)
>> > I decided to create VectorizedRoutines.jl. The idea behind this package
>> is
>> > to include the useful and familiar vectorized routines that one knows
>> and
>> > loves from MATLAB/R/Python. An example of such functions include things
>> > like meshgrid or accumarray. The reason for the package is three-fold:
>> >
>> > 1) Such a package will help newcomers who are first trying Julia and
>> used
>> > to using these functions
>> > 2) It will make porting codes to Julia easier (even though it is already
>> > easy!)
>> > 3) Sometimes these functions are a really quick solution to a problem.
>> They
>> > can be nice to use!
>> >
>> >   The functions I am looking to store here are ones not included into
>> Base.
>> > There are good reasons for not wanting to bloat Base with all of these:
>> in
>> > many cases they are not necessary (nor the best way to things) in Julia
>> > because Julia does not require vectorization like other scripting
>> > languages. Thus I see a package dedicated to holding these functions
>> (while
>> > making sure they achieve optimal performance, have proper
>> > documentation/testing) as a viable alternative.
>> >
>> >   Please feel free to submit pull requests for your own implementations
>> of
>> > "popular/familiar" functions you are missing in Julia. Also, feel free
>> to
>> > submit pull requests for "non-standard" vectorized routines which follow
>> > the same vein and are generally useful. Please include a docstring to
>> > discuss its usage or have it link to proper documentation. Also, feel
>> free
>> > to donate some tests! If this begins to build into something really
>> useful,
>> > I will make sure that this gets proper documentation through
>> Documenter.jl
>> > and add it to METADATA.
>>

Re: [julia-users] New Package Looking for Contributors: VectorizedRoutines.jl

[Chris Rackauckas]

Thanks for pointing that out. Never heard about that one. It seems like its 
been dead for quite awhile and they haven't answered any issues in almost a 
year, and haven't committed in 8 months? If anyone is on that project, 
please join the thread.

I think there are some major differences though. While they are targeting 
MATLAB porting, I am looking for just the useful vectorization routines. 
The large difference comes from the fact that since they have a bunch of 
tools for working with images and the like, they have a lot of dependencies 
(including Tk). I think that makes the package much less useful, especially 
since if you want to set it to precompile you'll probably run into issues 
if you don't want to deal with some of the dependencies (they don't have 
precompile enabled?). However, I think that some of the functions they 
implement would do well moving over to VectorizedRoutines.jl, which they 
could then import to their larger project.

On Friday, July 15, 2016 at 12:43:04 AM UTC-7, Mauro wrote:
>
> There is https://github.com/MatlabCompat/MatlabCompat.jl, which seems to 
> have a similar aim (although sans R/Python support).  Maybe join forces? 
>
> On Fri, 2016-07-15 at 01:34, Chris Rackauckas  > wrote: 
> > Hey, 
> >   After some discussion (and letting the idea hang around for a long 
> time) 
> > I decided to create VectorizedRoutines.jl. The idea behind this package 
> is 
> > to include the useful and familiar vectorized routines that one knows 
> and 
> > loves from MATLAB/R/Python. An example of such functions include things 
> > like meshgrid or accumarray. The reason for the package is three-fold: 
> > 
> > 1) Such a package will help newcomers who are first trying Julia and 
> used 
> > to using these functions 
> > 2) It will make porting codes to Julia easier (even though it is already 
> > easy!) 
> > 3) Sometimes these functions are a really quick solution to a problem. 
> They 
> > can be nice to use! 
> > 
> >   The functions I am looking to store here are ones not included into 
> Base. 
> > There are good reasons for not wanting to bloat Base with all of these: 
> in 
> > many cases they are not necessary (nor the best way to things) in Julia 
> > because Julia does not require vectorization like other scripting 
> > languages. Thus I see a package dedicated to holding these functions 
> (while 
> > making sure they achieve optimal performance, have proper 
> > documentation/testing) as a viable alternative. 
> > 
> >   Please feel free to submit pull requests for your own implementations 
> of 
> > "popular/familiar" functions you are missing in Julia. Also, feel free 
> to 
> > submit pull requests for "non-standard" vectorized routines which follow 
> > the same vein and are generally useful. Please include a docstring to 
> > discuss its usage or have it link to proper documentation. Also, feel 
> free 
> > to donate some tests! If this begins to build into something really 
> useful, 
> > I will make sure that this gets proper documentation through 
> Documenter.jl 
> > and add it to METADATA. 
>

Re: [julia-users] New Package Looking for Contributors: VectorizedRoutines.jl

[Mauro]

There is https://github.com/MatlabCompat/MatlabCompat.jl, which seems to
have a similar aim (although sans R/Python support).  Maybe join forces?

On Fri, 2016-07-15 at 01:34, Chris Rackauckas  wrote:
> Hey,
>   After some discussion (and letting the idea hang around for a long time)
> I decided to create VectorizedRoutines.jl. The idea behind this package is
> to include the useful and familiar vectorized routines that one knows and
> loves from MATLAB/R/Python. An example of such functions include things
> like meshgrid or accumarray. The reason for the package is three-fold:
>
> 1) Such a package will help newcomers who are first trying Julia and used
> to using these functions
> 2) It will make porting codes to Julia easier (even though it is already
> easy!)
> 3) Sometimes these functions are a really quick solution to a problem. They
> can be nice to use!
>
>   The functions I am looking to store here are ones not included into Base.
> There are good reasons for not wanting to bloat Base with all of these: in
> many cases they are not necessary (nor the best way to things) in Julia
> because Julia does not require vectorization like other scripting
> languages. Thus I see a package dedicated to holding these functions (while
> making sure they achieve optimal performance, have proper
> documentation/testing) as a viable alternative.
>
>   Please feel free to submit pull requests for your own implementations of
> "popular/familiar" functions you are missing in Julia. Also, feel free to
> submit pull requests for "non-standard" vectorized routines which follow
> the same vein and are generally useful. Please include a docstring to
> discuss its usage or have it link to proper documentation. Also, feel free
> to donate some tests! If this begins to build into something really useful,
> I will make sure that this gets proper documentation through Documenter.jl
> and add it to METADATA.

[julia-users] New Package Looking for Contributors: VectorizedRoutines.jl

[Chris Rackauckas]

Hey,
  After some discussion (and letting the idea hang around for a long time) 
I decided to create VectorizedRoutines.jl. The idea behind this package is 
to include the useful and familiar vectorized routines that one knows and 
loves from MATLAB/R/Python. An example of such functions include things 
like meshgrid or accumarray. The reason for the package is three-fold: 

1) Such a package will help newcomers who are first trying Julia and used 
to using these functions
2) It will make porting codes to Julia easier (even though it is already 
easy!)
3) Sometimes these functions are a really quick solution to a problem. They 
can be nice to use!

  The functions I am looking to store here are ones not included into Base. 
There are good reasons for not wanting to bloat Base with all of these: in 
many cases they are not necessary (nor the best way to things) in Julia 
because Julia does not require vectorization like other scripting 
languages. Thus I see a package dedicated to holding these functions (while 
making sure they achieve optimal performance, have proper 
documentation/testing) as a viable alternative.

  Please feel free to submit pull requests for your own implementations of 
"popular/familiar" functions you are missing in Julia. Also, feel free to 
submit pull requests for "non-standard" vectorized routines which follow 
the same vein and are generally useful. Please include a docstring to 
discuss its usage or have it link to proper documentation. Also, feel free 
to donate some tests! If this begins to build into something really useful, 
I will make sure that this gets proper documentation through Documenter.jl 
and add it to METADATA.