At JuliaCon, I described the idea of using Option types instead of NAtype to
make it easier to write type-stable code that interacts with NA’s. To help
facilitate a debate about the utility of this approach, I just wrote up a
minimal package that implements Option types:
https://github.com/john
For example, I used to have a base/userimg.jl file that read as follows:
Base.require("Distributions.jl")
Base.require("Optim.jl")
Base.require("DataFrames.jl")
Base.require("Gadfly.jl”)
— John
On Jul 30, 2014, at 9:13 PM, Ivar Nesje wrote:
> If you compile Julia yourself, you can use the und
enough about line search methods to have an
> intuition about that, but if it is the case, maybe the line search routine
> should throw a more informative error?
>
> -Thom
>
> On Wednesday, July 30, 2014 3:44:51 PM UTC-5, John Myles White wrote:
> Would be useful to understand
Wed, Jul 30, 2014 at 5:12 PM, wrote:
> I'm a bit stuck on this one. Could I get one more hint about a way I could
> get the same thing done without using the illegal indexing?
>
> On Wednesday, July 30, 2014 5:46:58 PM UTC-4, John Myles White wrote:
> Yeah, it’s the combina
utputarray[i+1] >= 0
> count += 1
> end
>
> Factoring in that Julia begins indexing from 1.
>
> On Wednesday, July 30, 2014 4:41:43 PM UTC-4, John Myles White wrote:
> This pseudocode almost works. Just replace Int64[1:len(outputarray)] with
> 1:lengt
I think you should assume you have to perform branching at run-time whenever
you want to condition behavior on values (like -1 vs. +1) rather than on types
(like Float64 vs. Int64).
So I think it’s going to be very difficult to achieve the goal you have in mind.
— John
On Jul 30, 2014, at 3:0
Would be useful to understand exactly what goes wrong if we want to fix this
problem. I’m mostly used to errors caused by inaccurate gradients, so I don’t
have an intuition for the cause of this problem.
— John
On Jul 30, 2014, at 10:45 AM, Thomas Covert wrote:
> No, I haven't tried that yet
This pseudocode almost works. Just replace Int64[1:len(outputarray)] with
1:length(outputarray).
There’s also a bug in your core logic, but I’ll leave fixing that as an
exercise to the reader.
— John
On Jul 30, 2014, at 1:03 PM, yaoismyh...@gmail.com wrote:
> Hi guys,
>
> I asked this in a
espace after the name of the
> function. I may find it hard to love a language where the addition of
> whitespace produces ambiguity. (Not to mention, I always put whitespace
> between
> name of function and args)
>
> John Myles White wrote:
>
>> Hi Neal,
>>
>
One typo in my code:
> FIR{S, T}(in::Vector{S}, coef::Vector{T}) = FIR{S, T}(zeros(S, size(in)),
> coef)
should be
> FIR{S, T}(in::Vector{S}, coef::Vector{T}) = FIR{S, T}(in, coef)
— John
On Jul 30, 2014, at 8:11 AM, John Myles White wrote:
> Hi Neal,
>
> I think rewr
Hi Neal,
I think rewriting your code to follow a little bit more conventional Julia
style will help you to see where you’ve gone wrong:
type FIR{S, T}
in::Vector{S}
coef::Vector{T}
function FIR(in::Vector{S}, coef::Vector{T})
new(zeros(S, size(in)), coef)
end
end
FIR{S, T}(i
I think you're assuming people's beliefs about programming languages are
simpler to modify than they've proven to be in our experience.
Here's an experiment you could try to see if I'm right: over the course of six
months, reply to every e-mail you see on julia-users that derives from a
misunde
On Jul 29, 2014, at 3:45 PM, Júlio Hoffimann wrote:
> 2014-07-29 19:32 GMT-03:00 John Myles White :
> I think the confusion is that Julio assumes Int is used for brevity, when it
> is actually used for cross-platform compability.
>
> -- John
>
> Yes, I assumed the In
I think the confusion is that Julio assumes Int is used for brevity, when it is
actually used for cross-platform compability.
-- John
On Jul 29, 2014, at 3:29 PM, Stefan Karpinski
wrote:
> Float would always be Float64.
>
>> On Jul 29, 2014, at 6:25 PM, Júlio Hoffimann
>> wrote:
>>
>> I'
Might we see the changes to the type system that are required to get
higher-order functions to specialize on input functions?
-- John
On Jul 29, 2014, at 9:49 AM, Stefan Karpinski wrote:
> There is no official roadmap for 0.4 yet – just a collection of ideas in
> various people's heads and s
gt; repeat(A, inner = Int[], outer = Int[])
> i.e isn't the comma in the wrong position?
>
> On Tuesday, July 29, 2014 4:06:22 PM UTC+2, John Myles White wrote:
> repeat([1, 2, 3], inner = [2])
>
> On Jul 29, 2014, at 7:03 AM, Andrei Berceanu wrote:
>
repeat([1, 2, 3], inner = [2])
On Jul 29, 2014, at 7:03 AM, Andrei Berceanu wrote:
> Whats the easiest way of going from an array
> [1,2,3] -> [1,1,2,2,3,3]
> i.e. repeating all elements 2 (or more generally, n) times?
>
> A
The DataStream functionality was moved out of the main DataFrames code because
it's not sufficiently robust. We need to update the docs.
Have you tried append! on DataFrame objects? It should resolve some of the
problems you're facing when using vcat/rbind caused by having to do more and
more c
I'm not aware of any such function. Are you talking about the ecdf() function
that comes with base R?
-- John
On Jul 28, 2014, at 2:05 PM, Darf Ferrara wrote:
> I'm looking for something like ecdf() in ggplot2.
I think a correlation matrix can have arbitrary rank, but might be wrong.
-- John
On Jul 28, 2014, at 9:36 AM, Stefan Karpinski wrote:
> Does this computation not always return a rank-1 matrix?
>
>
> On Mon, Jul 28, 2014 at 12:33 PM, John Myles White
> wrote:
> But how
But how would you know the rank of the correlation matrix in advance?
-- John
On Jul 28, 2014, at 9:25 AM, Stefan Karpinski wrote:
> This is the sort of thing that just begs for a custom representation of a
> rank-1 matrix, which fortunately, isn't terribly hard to implement in Julia.
>
>
>
Art wrapped FLANN, which should make things much easier.
— John
On Jul 28, 2014, at 8:19 AM, Dahua Lin wrote:
> We should probably incorporate (approximate) nearest neighbor search into
> Clustering.jl at some point.
>
> Dahua
>
>
> On Monday, July 28, 2014 10:09:59
FWIW, there’s a KD-tree implementation in NearestNeighbors.jl
— John
On Jul 28, 2014, at 7:27 AM, Jacob Quinn wrote:
> This probably isn't very helpful currently, but I've been meaning to try to
> do a `kd-tree` implementation that allows for fast clustering for up to 7-10
> dimensions. (the
type
> return el_type
> end
>
> # Example use case
> mat = {1.0 2; 3 4.0}
> el_type = infer_eltype(mat)
> tight_mat = convert(Array{el_type}, mat)
> @test eltype(tight_mat) == Real
> @test size(tight_mat) == size(mat)
> @test isequal(tight_mat, mat)
> ```
>
> Hop
Do we have a function that will take in an Array{Any}, compute the typejoin of
its elements and convert the Array to the tightest possible type?
— John
Is CUTEst.jl easier to get working these days? The issue I opened in March
seems to still be open.
— John
On Jul 27, 2014, at 6:40 AM, Tim Holy wrote:
> A package of test functions sounds worthwhile. There's also CUTEst.jl:
> https://github.com/lpoo/CUTEst.jl
>
> --Tim
>
> On Sunday, July 2
Re. (3), I’d like to move the optimization problems we use for testing out of
Optim into a separate package. Having a nice test suite would be a big gain for
JuliaOpt.
— John
On Jul 27, 2014, at 6:25 AM, Hans W Borchers wrote:
> Ken:
>
> (1) Thanks for pointing out this approach and for imp
You don’t push to Dict’s because they’re not ordered.
You can add a entry using indexing:
dc = Dict{UTF8String, Int64}()
dc["A."] = 2
— John
On Jul 26, 2014, at 5:09 PM, Ben Ward wrote:
> Is it possible to extend a dict with a key value pair?
>
> I don't see a method to do it in the standar
aily builds and
> will be incorporated into the next "Current Release"?
>
> Thank you agian for the quick response and the help.
>
> Regards,
> Don
>
> On Saturday, July 26, 2014 12:10:20 PM UTC-4, John Myles White wrote:
> If you look carefully at the issue
If you look carefully at the issue where the “merged” icon occurs, you’ll see
that a patch was merged that solved the issue. This is a new way that GitHub
has recently started visualizing events inside of an issue comment thread.
— John
On Jul 26, 2014, at 9:07 AM, Donald Lacombe wrote:
> De
You have to construct the contained objects explicitly. An Array without
initialization just contains junk memory for all types.
bs = Array(B, npop)
for i in 1:npop
bs[i] = B()
end
-- John
On Jul 25, 2014, at 5:36 PM, Ross Boylan wrote:
> I think I've narrowed down my previous problem. I
This is an interesting case. You need semicolons to get this to work on a
single line:
function bounds(p::Vector); [min(p), max(p)]; end
The preferred way to do this, by the way, is with the one-liner special syntax:
bounds(p::Vector) = [min(p), max(p)]
-- John
On Jul 25, 2014, at 11:21 AM,
I'd be in favor of having both assert and enforce.
-- John
On Jul 25, 2014, at 7:08 AM, Magnus Lie Hetland wrote:
> Ah, right. Yeah, I did look for those in the Julia code, but couldn't find
> them. I see now that they're in repl.c, where they affect jl_compileropts.
>
> Well, there are two
I'd be in favor of adding the feature to Julia.
-- John
On Jul 25, 2014, at 6:38 AM, Iain Dunning wrote:
> Someone could add the feature to Julia itself, and I guess its in there sorta
> in that it needs that knowledge to resolve dependencies, but I think making
> "nice" output probably is b
Arrays of strings aren’t part of a vector space. There’s no reason you should
be able to multiply them by a scalar.
— John
On Jul 24, 2014, at 8:45 PM, cnbiz850 wrote:
> Well, * works with numerics
>
> julia> A = [1:5]
> 5-element Array{Int64,1}:
> 1
> 2
> 3
> 4
> 5
> julia> A *3
> 5-element
Vector{String} is semantically equivalent to Array{String, 1}. It's just more
readable.
Also worth noting that you basically never want to work with Vector{String}.
You almost certainly want Vector{UTF8String}.
-- John
On Jul 24, 2014, at 7:47 PM, cnbiz850 wrote:
> Thanks. Didn't know abou
I'd love to have this functionality, but it's worth noting that some code (e.g.
Optim) has assert being used for error handling right now. So there'd need to
be a period where people transition away from using assertions for
error-handling rather than for testing.
-- John
On Jul 24, 2014, at
4,1},Float64,Float64,Float64,Float64)
>
> I am using 0.2.1.
>
> On Thursday, July 24, 2014 2:13:58 PM UTC-4, John Myles White wrote:
> You must splat if you use push!
>
> You must _not_ splat if you use append!
>
> -- John
>
> On Jul 24, 2014, at 11:10 AM, y
4)
> Do I need to use a different sort of array?
>
> On Thursday, July 24, 2014 12:48:15 PM UTC-4, Jameson wrote:
> Yes
>
> Or use append! instead of push!
>
>
> On Thursday, July 24, 2014, wrote:
> Ah, is ... equivalent to the * in Python?
>
> On Thursday
I believe the main problem with the global scope isn't that it's called the
global scope (rather than the scope of a let block), but that the scope is
never terminated during a REPL session, preventing end-to-end analysis of types.
-- John
On Jul 24, 2014, at 7:14 AM, Arnaud Amiel wrote:
> I
You need to splat the tuple first:
push!(tuplestorage,newtuple...)
-- John
On Jul 24, 2014, at 9:38 AM, yaoismyh...@gmail.com wrote:
> Hi,
> Translating some of my code from Python, I am trying to find a means of
> concatenating and appending tuples to an array.
>
> In Python, I was able to
I think the difficulty here is that the people who are worst affected by these
kinds of performance changes may be people who also might not know that they
should opt in to using Lint/TypeCheck. To get the proper effect from those
tools, you probably need to impose them from above by default rat
Why not view this as a learning opportunity to discover that changes you
perceive to be minor are, in fact, not minor -- but are instead substantive
changes to the semantics of your program?
-- John
On Jul 23, 2014, at 9:38 AM, Arnaud Amiel wrote:
> As suggested by a few of you, I put everyt
FWIW, the mean of the min and max is called the midrange.
— John
On Jul 22, 2014, at 8:17 AM, Spencer Russell wrote:
> Hi Elliot,
>
> The median isn't the mean of the min and max (though that was fun to say).
>
> Wikipedia says it well:
>
> "the median is the numerical value separating the
There's one complication here, which is that a single ? already changes the
mode of the REPL into help mode. So this would have to be a mode only
triggerable from inside of help mode.
-- John
On Jul 21, 2014, at 6:47 PM, Ethan Anderes wrote:
> +1. For some reason my fingers always get tied u
t;>
>> These have been working quite successfully for some packages, such as
>> Distributions, DataStructures, and Graphs, etc.
>>
>> Dahua
>>
>>
>>
>> On Saturday, July 19, 2014 5:26:36 PM UTC-5, John Myles White wrote:
>> I su
I suppose one very effective solution to creating a coherent ML organization is
to do all the work yourself. :)
-- John
On Jul 19, 2014, at 1:40 PM, Dahua Lin wrote:
> Linear Least Square and Ridge Regression are also included now. (see
> http://multivariatestatsjl.readthedocs.org/en/latest/
FWIW, shared memory parallelism is going to improve in Julia in the future.
-- John
On Jul 19, 2014, at 9:40 AM, Tracy Wadleigh wrote:
> A more robust path to shared memory parallelism. SharedArray is still not a
> first-class Julia citizen. And, even if it were, memory safety would still be
; On Friday, July 18, 2014 9:11:08 PM UTC-5, John Myles White wrote:
> This is really great, Dahua. We should probably deprecate MultivariateStats,
> which was always pretty unreliable.
>
> -- John
>
> On Jul 18, 2014, at 9:13 PM, Dahua Lin wrote:
>
>>
This is really great, Dahua. We should probably deprecate MultivariateStats,
which was always pretty unreliable.
-- John
On Jul 18, 2014, at 9:13 PM, Dahua Lin wrote:
> Recently, I developed a new package for Julia (under JuliaStats):
> MultivariateStats, for multivariate statistical analysi
This might seem like just a reframing of the situation, rather than an answer,
but I personally don't think there's any background knowledge you need to get
started using Julia. What you need is a willingness to figure out what's going
on by doing some digging through the manual, the general web
To follow up on Dahua's comment:
If you're interested in using the fastest code, the code that Dahua posted is
the fastest since it uses several clever tricks to get the absolute best speed.
If you're interested in writing a faster implementation than the one you wrote,
you need to avoid alloca
I think you want something more like a >= 0 && b >= 0 && c >= 0
Also, you probably don't want to use global variables for code where
performance matters. You should pass a, b and c to gen() as parameters gen(a,
b, c).
-- John
On Jul 15, 2014, at 6:31 PM, yaoismyh...@gmail.com wrote:
> Hi,
>
Basic linear algebra is done by BLAS and LAPACK, so I don't think these issues
cause performance problems.
-- John
On Jul 14, 2014, at 6:55 PM, Andrei Zh wrote:
> Linear transformation is super-frequent operation in scientific computations,
> so I'd like to know how to perform it in a most e
Why not just use straight indexing like below?
df[Bool[ismatch(r"a", x) for x in df[:Column]], :]
-- John
On Jul 14, 2014, at 9:32 AM, Andrew G wrote:
> vec = map(x -> ismatch(r"regex_expression", x), df[:column])
> vec = convert(DataArray{Bool}, vec)
> subset(df, results)
>
parse isn’t a cure for anything. parse interprets inputs under the assumption
that they are valid Julia code. Your data isn’t valid Julia code, so you should
not use parse.
— John
On Jul 13, 2014, at 12:01 AM, Paul Analyst wrote:
> At first I did and it was a problem with the "Inf". I realiz
I loved this two-part essay. Highly recommend it to anyone who hasn’t read it.
— John
On Jul 12, 2014, at 7:55 PM, Stefan Karpinski wrote:
> Graydon Hoare (original author of Rust) wrote a truly lovely essay in two
> parts about the history of programming languages, the predominance of
> two
On Jul 12, 2014, at 1:04 PM, J Luis wrote:
> That is also true but a much more rare case, typemax(Int32) is still a quite
> high number for an array size and before an Int64 is needed changes are non
> negligible that a memory requested failed because a big contigous chunk of
> memory was not
I’m really confused. Do you want the indices to be 32-bit integers instead of
64-bit integers? Isn’t that equivalent to asking for your code to be broken
anytime your vector has more than typemax(Int32) elements?
— John
On Jul 12, 2014, at 12:43 PM, J Luis wrote:
>
> julia> find(x->x>5,a)
>
Maybe we should have Learning and Teaching separately? Or maybe Teaching should
be under a section about people using Julia?
— John
On Jul 12, 2014, at 9:35 AM, Stefan Karpinski wrote:
> Well, originally, it was actually teaching in the sense of places where
> people have used Julia in teach
I'd use a package like Distributions.jl as your example rather than SGD.jl,
which is really, really rough around the edges.
-- John
On Jul 10, 2014, at 8:32 AM, Eric Chiang wrote:
> Yeah I have.
>
> Beyond just the actual technical requirements, I was wondering more generally
> about best p
This is a really great tutorial.
-- John
On Jul 10, 2014, at 5:54 AM, David P. Sanders wrote:
> Many thanks to all those who suffered my inane questions and helped this to
> take shape, in particular:
>
> - Jeff, who persuaded me to try out Julia
> - Alonso & Luis, who convinced me that it w
I think you're hitting this issue: https://github.com/JuliaLang/julia/issues/265
-- John
On Jul 10, 2014, at 1:31 AM, Johannes Degn wrote:
>
>
> Im not sure if this is a bug or if I am a bug. Maybe someone can explain this
> to me:
>
> I am running julia Version 0.3.0-prerelease+3649 on li
to Mamba.jl -
> http://mambajl.readthedocs.org/en/latest/intro.html#the-mamba-package ).
>
> Regards,
> Rob J. Goedman
> goed...@icloud.com
>
>
>
>
> On Jul 9, 2014, at 8:45 AM, John Myles White wrote:
>
>> I’m not sure I understand how a C++ FFI af
gt; Julia's v0.3-prerelease (as it uses the new DataFrames).
>
> Finally, a significant part of the value of CmdStan/Stan is the Stan team &
> community. Even with a mature C++ FFI, I would like to remain as
> 'interchangeable' with both Stan and CmdStan as possib
What you're doing isn't a workaround: it's the correct way to do this in the
current version of Julia. There may be shorthand in the future, but this is the
right approach today.
-- John
On Jul 8, 2014, at 2:01 PM, Andrei Zh wrote:
> Here's another question about code style. Let's say I want
no method optimize(Function, Int64)
>
> I am running Version 0.3.0-prerelease+3884 (2014-06-25 10:41 UTC) and have
> updated all packages.
>
> Thanks for your help
> Mathieu
>
> On Tuesday, July 8, 2014 2:48:24 PM UTC-4, John Myles White wrote:
> This should work, exce
This should work, except that you're not providing any initial values for x. Is
your problem one-dimensional or multi-dimensional?
-- John
On Jul 8, 2014, at 11:44 AM, Mathieu Taschereau-Dumouchel
wrote:
> I am trying to optimize a simple function that takes a parameter as an
> argument. So
This is a pretty impressive range of different talks for a single event. Very
cool.
— John
On Jul 6, 2014, at 5:35 PM, Sorami Hisamoto wrote:
> Last week's JuliaTokyo meetup in Japan ended in great success!
>
> Around 40 people attended, 6 main talks and 4 lightning talks.
>
> Here's the li
This is really nice to have. I'll definitely be using it for some of my work.
Now that Keno's been working on a C++ FFI, we might even be able to avoid using
CmdStan.
-- John
On Jul 7, 2014, at 4:40 PM, Viral Shah wrote:
> It is great to have a Stan interface. I personally do not do MCMC stu
Thanks for looking into this, Abe. That’s too bad that the CSV parser is much
slower than the hand-crafted one. PEG seems like a great tool for tasks where
maximum performance isn’t as important.
— John
On Jul 4, 2014, at 11:09 AM, Abe Schneider wrote:
> I got sidetracked by a couple of othe
I think anonymous function usage and list comprehensions are separate issues.
Anonymous functions are not type-specialized, so they’re basically always slow.
I avoid them in any code that’s not one-off.
List comprehensions are type-specialized, but the inference for them isn’t
super robust, esp
Not really sure. Can you try printing out each of the columns to see where the
error is? Maybe the problem is with our new code for printing shorthand rather
than eps() when p is very small. (That’s just total speculation, though.)
— John
On Jul 4, 2014, at 1:04 PM, Florian Oswald wrote:
> H
Real is the abstract type that all real numbers derive from. Int is a subtype
of Real, for example.
Int is a platform-specific alias for either Int32 or Int64, whichever is native
to your machine/OS combination.
So a type like
immutable Foo
x::Int
end
is quite efficient, whereas a typ
One big problem is the use of “Real”. You almost certainly should be using
Float64 instead.
— John
On Jul 5, 2014, at 11:46 AM, Michael Louwrens
wrote:
> https://github.com/Mike43110/GA-Spherical-Jl/tree/master
>
> Here is my basic implementation of a GA in Julia.
>
> Using similar code in
This seems like a temporary bug. My most recent build doesn’t behave like this.
Hopefully that means it’s been fixed already.
— John
On Jul 4, 2014, at 1:29 PM, james.dillon.dela...@gmail.com wrote:
> Hi All,
>
> Is there a reason why the irrational constant e seems to be rounded to three
>
This is really cool. I’ve wanted to do this for a while, but never had time.
— John
On Jul 4, 2014, at 11:39 AM, Iain Dunning wrote:
> Oh and I forgot, link to the repository:
> https://github.com/IainNZ/GraphLayout.jl
>
> On Friday, July 4, 2014 2:39:32 PM UTC-4, Iain Dunning wrote:
> I wan
Hi Paulo,
Rather than implement k-means from scratch, I'd encourage you to use the
implementation in the Clustering.jl package.
-- John
On Jul 3, 2014, at 2:51 PM, Paulo Castro wrote:
> Hi guys,
>
> I'm trying to implement the K-Means Clustering Algorithm, but I'm having some
> problems. T
not operations on a ring.
>
>
> On Thu, Jul 3, 2014 at 11:13 PM, John Myles White
> wrote:
> That is true, but the convention exists nonetheless. Consider, for example,
> this branch of the combinatorics literature:
> http://en.wikipedia.org/wiki/Square-free_word
> --
, a monoid is
> enough. It is true that in a ring, "+" is a commutative operator, and "*" no
> commutative required, however, a ring is far beyond what needed to do a
> string concatenation job.
>
>
> On Thu, Jul 3, 2014 at 10:59 PM, John Myles White
> wrot
This is just the standard convention in mathematics. See just about any
textbook on abstract algebra for a discussion of the reasons.
I think it's safe to say that `+` will never be string concatenation, so it's
better to focus energy on other proposals.
-- John
On Jul 3, 2014, at 7:57 AM, yi
Sadly, R does not. 1:0 expands to c(1, 0). This is something that has burned a
lot of people in my experience. I imagine this is inherited from S.
— John
On Jul 3, 2014, at 2:05 AM, Tobias Knopp wrote:
> Just as a side note (I entirely agree with Stefan), Matlab behaves the same
> as Julia:
performance you should use IOBuffer for
> repeated string concatenation, example..."
>
> Also a more instructive error on "hello"+"world" would make since.
>
> On Wednesday, 2 July 2014 20:48:26 UTC+1, John Myles White wrote:
> String concatenation is not comm
String concatenation is not commutative. Addition is generally used for
commutative operations. So if you're a mathematician, using addition for string
concatentation seems very wrong.
-- John
On Jul 2, 2014, at 12:45 PM, Ivar Nesje wrote:
> Not everybody is convinced there is a good reason.
What would that macro do? Right now, I’d guess that it would just take its
input out of quotes.
— John
On Jul 2, 2014, at 7:41 AM, Bryan A. Knowles wrote:
> What of we had a macro version of parse? Eg,
> ```julia
> function foo()
> foo2(x) = println(x)
> foo2(1)
>
of a
> string or symbol to result in a call to a nested function?
>
> On Wednesday, July 2, 2014 1:17:07 AM UTC-4, John Myles White wrote:
> eval always happens in the global scope
>
> — John
>
> On Jul 1, 2014, at 10:15 PM, Cameron Smith wrote:
>
>> Can anyon
eval always happens in the global scope
— John
On Jul 1, 2014, at 10:15 PM, Cameron Smith wrote:
> Can anyone explain what I am misunderstanding about scope such that
>
> julia> function foo()
>foo2(x) = println(x)
>foo2(1); eval(parse("foo2(2)"));
>end
> foo (
Hey Spencer,
A function is defined to be "type stable" if the return types depend only on
the input types.
-- John
On Jun 30, 2014, at 9:51 AM, Tim Holy wrote:
> In your definition I think you meant "type stability," not "type
> instability."
> To be completely explicit, a function is type
Please don’t post code whose copyright status is not 100% certain. If you are
even vaguely unsure, you are putting the community in needless danger.
As a minimal example, it may be illegal for you to have written the e-mail you
just wrote.
— John
On Jun 30, 2014, at 7:07 AM, 'Stéphane Laurent
(1) Are you including compilation time in your timings?
(2) Your MyCounter type isn't very similar to your other examples, since the
i::Integer field has type-uncertainty, whereas the other examples don't.
-- John
On Jun 24, 2014, at 2:26 PM, JuliaFan wrote:
> Why do these three loops perfor
(1) The size of arrays isn’t part of their type signature, so you can’t use
multiple dispatch to do this. You’ll end up overwriting the existing Array
multiplication function.
(2) You’d want to do something like *{S <: Number, T <: Number}(a::Array{S,
2},b::Array{T,1}) to indicate that you’re a
The fast answer is that you want a copy of A, produced using copy(A). Somebody
else will probably fill in more details of the why.
-- John
On Jun 23, 2014, at 9:42 AM, JuliaLover wrote:
> Hi,
> I have a problem that when I assign value of one array to another they stick
> to each other whene
142
> drwxr-xr-x 24 root root 4096 jun 11 09:16 .
>
> shell> whoami
> tlycken
> `tlycken` is a sudoer, but if I start the REPL as `sudo julia`, `;whoami`
> returns `root`, so I don't think that's the problem.
>
> // T
>
> On Monday, June 23, 2014 4:07
Even with -l?
— John
On Jun 23, 2014, at 7:14 AM, Davide Lasagna wrote:
> The output is my linux root folder: /bin, /boot,
>
> On Monday, June 23, 2014 3:07:12 PM UTC+1, John Myles White wrote:
> This seems hard to know without also seeing the output of `ls -l /`
>
&g
This seems hard to know without also seeing the output of `ls -l /`
— John
On Jun 23, 2014, at 7:06 AM, Davide Lasagna wrote:
> Hi,
>
> Is this a bug?
>
> julia> iswritable("/")
> true
>
> (running from repl as normal user)
>
> Davide
Maybe. Did someone create a pull request?
— John
On Jun 22, 2014, at 5:22 PM, Thomas Covert wrote:
> did this code ever find its way into DualNumbers.jl? I do anticipate its
> going to be quite helpful.
>
> -Thom
>
>
>
>
> On Fri, Jun 6, 2014 at 10:32 AM, Thomas Covert wrote:
> Haven't
Multinomial() doesn't work because there's no reasonable default we could
assume. Did you want 3 dimensions? 100 dimensions?
Arguably none of the defaults should exist, but we haven't pushed on that yet.
Maybe we should start removing all the default distributions.
The other errors you noticed
I think the code that controls inlining is mostly here:
https://github.com/JuliaLang/julia/blob/master/base/inference.jl
But you'll want to get a response from someone who really understands it.
-- John
On Jun 18, 2014, at 1:30 PM, Sam L wrote:
> Apologies if this is in the manual. I must ha
Sounds like we need to rerun these benchmarks after the new GC branch gets
updated.
-- John
On Jun 17, 2014, at 9:31 AM, Stefan Karpinski wrote:
> That definitely smells like a GC issue. Python doesn't have this particular
> problem since it uses reference counting.
>
>
> On Tue, Jun 17, 2
repeat(A,inner=[1,1],outer=[1,1,2])
> 2x2x2 Array{Int64,3}:
> [:, :, 1] =
> 1 2
> 3 4
>
> [:, :, 2] =
> 1 2
> 3 4
>
>
> Is there a limit on how a docstring can be? Could we add more examples?
>
> On Thursday, June 12, 20
401 - 500 of 919 matches
Mail list logo