5 ms·

Having never used it, I wonder, could Julia replace Python as the de-facto standard in scientific computing and data analysis? What does it do well?

ComputerGuru 8y ago
I thought that was what R was meant to do? But, yes, Julia was designed with something like that in mind from what I recall.
- ellisv 8y ago
  R is actually only 3 years younger than Python (25yrs and 28yrs respectively).
  - kgwgk 8y ago
    And the array data type in Python is just 23 years old.
pjmlp 8y ago
It compiles to native code via JIT on the standard implementation, no need to re-write code in C. The object system is close to CLOS with support for multi-dispatch.
- cultus 8y ago
  The object system is IMO the best feature of Julia. All functions are multiple dispatch, so the expression problem is more or less solved. This makes it easy to do things like allowing users to define custom sparse matrix types and use them seamlessly with other matrices and vectors. It makes it easy to abstract away implementation details in a much greater way than Python or R.
collinf 8y ago
My only fleeting experience with Julia is in a Numerical Analysis Course three years ago... but it has 1-indexed arrays so its good for scaring away all people who are programmers before analysts/scientists! Really though, I think it has a better type system and a syntax that translates easier to mathematical expressions. Other than that, Python's breadth of packages will be hard to overcome.
- cultus 8y ago
  You get used to the 1-based indexing pretty quick. It's a non-issue. 1-based indexing is the standard in math, and also Matlab and Fortran. For the primary use-case of Julia, it makes sense. Julia REALLY tries to be easy for Matlab users, who still dwarf Python users in science.
  - kgwgk 8y ago
    > Julia REALLY tries to be easy for Matlab users Not so much currently: https://discourse.julialang.org/t/why-eye-has-been-deprecated/12824 https://discourse.julialang.org/t/why-eye-has-been-deprecate...
- chriselrod 8y ago
  I know it is not the default, but this package has no overhead over the default arrays according to recent benchmarks, and provides 0-indexed arrays: https://github.com/JuliaArrays/OffsetArrays.jl https://github.com/JuliaArrays/OffsetArrays.jl I confess, not being the default is a big thing. I've definitely had times where I thought "this would be easier with 0 indexed arrays", but it can then be harder to commit to adding a dependency and making that change vs just adding awkward looking "+1"s to all the indexes. Coming from math/science, there's lots of times 1-indexing makes more sense / is more familiar. It's normal there to start counting from 1, so it can be easier to translate.
  - phkahler 8y ago
    It's non uncommon in mathematics to index from zero as well.
- geoalchimista 8y ago
  Even among scientists there are people who don't like 1-indexed arrays (me, one data point)! However, Julia 0.7 and 1.0 have the keyword "begin" so that you can basically program arrays in an index-agnostic way.
- pjmlp 8y ago
  When I started programming, all my languages with exception of Assembly used 1-indexing, so no, it doesn't scare all programmers.
montalbano 8y ago
Having migrated my Python+Fortran code to Julia in October last year I would say yes. For reasons why I would read the documentation basics, check the benchmarks and play with a bit of code. I was sceptical when my advisor originally suggested it but now I dread the thought of scientific software in any other language.
- nerdponx 8y ago
  IMO industry will start switching once a Scikit-learn or Caret equivalent for Julia shows up. PyCall.jl and RCall.jl should take care of everything else until the rest of the ecosystem catches up.
  - FridgeSeal 8y ago
    Last time I tried to move all my stuff to Julia I ran into issues with the data frame equivalent package. It seemed the devs and community were in debate as to whether Julia and the dataframe should be more pure numerical focused a la matlab or python/pandas and appropriate for data analysis. As such null values were handled weird or not supported, which made it practically useless for all of my use cases. I’d love to know if that was still the case though, because I’d so much rather use Julia than Python for my analysis and stuff.
    - Certhas 8y ago
      Actually 0.7 made a big step in resolving these issues: https://julialang.org/blog/2018/06/missing https://julialang.org/blog/2018/06/missing
      - FridgeSeal 8y ago
        Like music to my ears!
  - montalbano 8y ago
    Not necessarily, I recommend the talk given today at JuliaCon by the head of an actuary research group at Aviva - a very large insurance company: https://www.youtube.com/watch?v=_jx1VmWxgVY https://www.youtube.com/watch?v=_jx1VmWxgVY They've been using it sucessfully for quite some time.
crwalker 8y ago
I hope so - Julia has been one of the most refreshing software experiences I've tried, and that was only Julia 0.4.
sgillen 8y ago
Python has so much inertia at this point that it seems unlikely to happen in the near future. Probably not quite to the scale as C/C++ vs rust for systems programming, but a similar idea. Rust has all these great features but most people doing systems know C, all their code is already in C, and so the cost of switching is very high. Not that a switch will never happen, it's just that no matter how good Julia is any transition is going to take a long time. (I do think Julia is a good language though)
tmalsburg2 8y ago
Python is not the de-facto standard to my knowledge. It really depends on the field and task. R is huge in my research area, MatLab is king in some other areas, and Fortran is still being used for large-scale number crunching. Will Julia replace all these? That's the ambition but I doubt it's going to happen. For instance, R has a unique and vast ecosystem of extension packages. It would take decades to build something comparable in Julia.
- FridgeSeal 8y ago
  If there were pandas, sklearm and Keras equivalents (don’t even need feature compleye, just mostly there) I would abandon python literally today.
  - mindB 8y ago
    - pandas :: your choice of DataFrames or JuliaDB - scikitlearn :: no single package since skl is a meta-package of sorts, but most of the stuff is there spread across the eco-system - Keras :: checkout Flux or Mocha Welcome to Julia! Also take a look here [1] for more package that might be to your interest. [1] https://github.com/svaksha/Julia.jl/blob/e305195ab60e6859e782f625677c08ad36f5d918/AI.md https://github.com/svaksha/Julia.jl/blob/e305195ab60e6859e78...
    - oxinabox 8y ago
      Pandas actually has many alternatives. Including JuliaDB, DataFrames, and Pandas.jl Keras also many. TensorFlow.jl, Flux, Mocha, KNet, MxNet. Sklearn is all there but bring it together will take a bit. Some important parts are in JuliaML org. Also Clustering.jl, and MultiariateStats.jl (For DR) the classifiers are really scattered. I'ld love to fix that if I had time.
    - FridgeSeal 8y ago
      You're really making my day here! :D
- Certhas 8y ago

ComputerGuru 8y ago

I thought that was what R was meant to do? But, yes, Julia was designed with something like that in mind from what I recall.

ellisv 8y ago
R is actually only 3 years younger than Python (25yrs and 28yrs respectively).
- kgwgk 8y ago
  And the array data type in Python is just 23 years old.

pjmlp 8y ago

It compiles to native code via JIT on the standard implementation, no need to re-write code in C. The object system is close to CLOS with support for multi-dispatch.

cultus 8y ago
The object system is IMO the best feature of Julia. All functions are multiple dispatch, so the expression problem is more or less solved. This makes it easy to do things like allowing users to define custom sparse matrix types and use them seamlessly with other matrices and vectors. It makes it easy to abstract away implementation details in a much greater way than Python or R.

collinf 8y ago

My only fleeting experience with Julia is in a Numerical Analysis Course three years ago... but it has 1-indexed arrays so its good for scaring away all people who are programmers before analysts/scientists! Really though, I think it has a better type system and a syntax that translates easier to mathematical expressions. Other than that, Python's breadth of packages will be hard to overcome.

cultus 8y ago
You get used to the 1-based indexing pretty quick. It's a non-issue. 1-based indexing is the standard in math, and also Matlab and Fortran. For the primary use-case of Julia, it makes sense. Julia REALLY tries to be easy for Matlab users, who still dwarf Python users in science.
- kgwgk 8y ago
  > Julia REALLY tries to be easy for Matlab users Not so much currently: https://discourse.julialang.org/t/why-eye-has-been-deprecated/12824 https://discourse.julialang.org/t/why-eye-has-been-deprecate...
chriselrod 8y ago
I know it is not the default, but this package has no overhead over the default arrays according to recent benchmarks, and provides 0-indexed arrays: https://github.com/JuliaArrays/OffsetArrays.jl https://github.com/JuliaArrays/OffsetArrays.jl I confess, not being the default is a big thing. I've definitely had times where I thought "this would be easier with 0 indexed arrays", but it can then be harder to commit to adding a dependency and making that change vs just adding awkward looking "+1"s to all the indexes. Coming from math/science, there's lots of times 1-indexing makes more sense / is more familiar. It's normal there to start counting from 1, so it can be easier to translate.
- phkahler 8y ago
  It's non uncommon in mathematics to index from zero as well.
geoalchimista 8y ago
Even among scientists there are people who don't like 1-indexed arrays (me, one data point)! However, Julia 0.7 and 1.0 have the keyword "begin" so that you can basically program arrays in an index-agnostic way.
pjmlp 8y ago
When I started programming, all my languages with exception of Assembly used 1-indexing, so no, it doesn't scare all programmers.

montalbano 8y ago

Having migrated my Python+Fortran code to Julia in October last year I would say yes. For reasons why I would read the documentation basics, check the benchmarks and play with a bit of code. I was sceptical when my advisor originally suggested it but now I dread the thought of scientific software in any other language.

nerdponx 8y ago
IMO industry will start switching once a Scikit-learn or Caret equivalent for Julia shows up. PyCall.jl and RCall.jl should take care of everything else until the rest of the ecosystem catches up.
- FridgeSeal 8y ago
  Last time I tried to move all my stuff to Julia I ran into issues with the data frame equivalent package. It seemed the devs and community were in debate as to whether Julia and the dataframe should be more pure numerical focused a la matlab or python/pandas and appropriate for data analysis. As such null values were handled weird or not supported, which made it practically useless for all of my use cases. I’d love to know if that was still the case though, because I’d so much rather use Julia than Python for my analysis and stuff.
  - Certhas 8y ago
    Actually 0.7 made a big step in resolving these issues: https://julialang.org/blog/2018/06/missing https://julialang.org/blog/2018/06/missing
    - FridgeSeal 8y ago
      Like music to my ears!
- montalbano 8y ago
  Not necessarily, I recommend the talk given today at JuliaCon by the head of an actuary research group at Aviva - a very large insurance company: https://www.youtube.com/watch?v=_jx1VmWxgVY https://www.youtube.com/watch?v=_jx1VmWxgVY They've been using it sucessfully for quite some time.

crwalker 8y ago

I hope so - Julia has been one of the most refreshing software experiences I've tried, and that was only Julia 0.4.

sgillen 8y ago

Python has so much inertia at this point that it seems unlikely to happen in the near future. Probably not quite to the scale as C/C++ vs rust for systems programming, but a similar idea. Rust has all these great features but most people doing systems know C, all their code is already in C, and so the cost of switching is very high. Not that a switch will never happen, it's just that no matter how good Julia is any transition is going to take a long time. (I do think Julia is a good language though)

tmalsburg2 8y ago

Python is not the de-facto standard to my knowledge. It really depends on the field and task. R is huge in my research area, MatLab is king in some other areas, and Fortran is still being used for large-scale number crunching. Will Julia replace all these? That's the ambition but I doubt it's going to happen. For instance, R has a unique and vast ecosystem of extension packages. It would take decades to build something comparable in Julia.

FridgeSeal 8y ago
If there were pandas, sklearm and Keras equivalents (don’t even need feature compleye, just mostly there) I would abandon python literally today.
- mindB 8y ago
  - pandas :: your choice of DataFrames or JuliaDB - scikitlearn :: no single package since skl is a meta-package of sorts, but most of the stuff is there spread across the eco-system - Keras :: checkout Flux or Mocha Welcome to Julia! Also take a look here [1] for more package that might be to your interest. [1] https://github.com/svaksha/Julia.jl/blob/e305195ab60e6859e782f625677c08ad36f5d918/AI.md https://github.com/svaksha/Julia.jl/blob/e305195ab60e6859e78...
  - oxinabox 8y ago
    Pandas actually has many alternatives. Including JuliaDB, DataFrames, and Pandas.jl Keras also many. TensorFlow.jl, Flux, Mocha, KNet, MxNet. Sklearn is all there but bring it together will take a bit. Some important parts are in JuliaML org. Also Clustering.jl, and MultiariateStats.jl (For DR) the classifiers are really scattered. I'ld love to fix that if I had time.
  - FridgeSeal 8y ago
    You're really making my day here! :D
Certhas 8y ago