Add vectorized mod2pi

[eford]

No description provided.

[eford]

I understand the Julia automated testing system.
If you want a test, you could add

@test_approx_eq maximum(abs(mod2pi(testCases[:,1]).-testCases[:,2])) 0.0

to the end of

julia/test/mod2pi.jl

[stevengj]

Now that map is fast, hoping to move away from vectorizing all the functions, and just have a general syntax like .mod2pi(x) for broadcasting. See #8450

[eford]

I definitely like making map faster, but that proposed syntax for
broadcasting functions strikes me as ugly and prone to creating bugs due
to confusing syntax.
At one point Julia prided itself in allowing syntax like 2func(x).
Which makes 2.func(x) ambiguous or at least confusing to humans.

On Fri, Apr 8, 2016 at 4:49 PM, Steven G. Johnson [email protected]
wrote:

Now that map is fast, hoping to move away from vectorizing all the
functions, and just have a general syntax like .mod2pi(x) for
broadcasting. See #8450 #8450

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#15797 (comment)

Eric Ford
Professor of Astronomy & Astrophysics
Center for Exoplanets & Habitable Worlds
Center for Astrostatistics
Institute for CyberScience
Penn State Astrobiology Research Center
Pennsylvania State University

[nalimilan]

The proposed syntax is actually func.(x), not not .func(x).

Anyway, I think the question is: are other similar functions vectorized or not? Arbitrarily vectorizing some functions and not others isn't a very good solution.

[ViralBShah]

Yes, at this point, we want to have some syntax for vectorization and discourage addition of more vectorized operations.

[eford]

Personally, I'm not convinced that the choice of syntax is wise. I was
happy with having a simple macro to vectorize functions that one wanted
vectorized. If we need a naming convention, I think I'd rather have
something like func_vec, rather than an obscure "." syntax that's easy to
overlook. But if that's the worst thing about julia, then I'll get over
it. I definitely appreciate the work of the julia developers.

On Mon, Apr 11, 2016 at 6:37 AM, Viral B. Shah [email protected]
wrote:

Yes, at this point, we want to have some syntax for vectorization and
discourage addition of more vectorized operations.

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
#15797 (comment)

Eric Ford
Professor of Astronomy & Astrophysics
Center for Exoplanets & Habitable Worlds
Center for Astrostatistics
Institute for CyberScience
Penn State Astrobiology Research Center
Pennsylvania State University

[nalimilan]

@eford That issue has been discussed a lot already. See the other issue for arguments.

[tkelman]

closing, given #15032