Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

JuliaCon 2015 Preview

julialang.org

Post reply on HN
Print view View original

11–20 of 20 posts

Re: JuliaCon 2015 Preview

#11

Post by idunning »

transfire wrote: post #9
idunning wrote: post #8

Earlier quoted context omitted.

No, I don't think so. I don't think there is much that is going to change with modules in the future, or at least I'm unaware of any open issues about it. What did you have in mind? The hackathon on Wednesday will be a particularly good time to bounce around your ideas, and is free to attend.

Because the current system ( http://julia.readthedocs.org/en/latest/manual/modules/ ) is just sad. It's the one thing that has kept me from fully embracing Julia. I would love to go if I were any where in the neighborhood. But alas I'm over a thousand miles away.

Ouch, "just sad"... would be interested to hear more!

Even if you can't attend, anything you could write up as either a blog post, a Github issue, anything, would be helpful for getting some discussions going around it at the conference I'm sure.

Re: JuliaCon 2015 Preview

#12

Post by vanderZwan »

idunning wrote: post #2

If anyone has any questions about this post or the conference, fire away! Last year's first conference was a big success in my opinion. Videos are available at http://juliacon.org/2014/ , although the speakers weren't mic'ed so the sound isn't so great - we'll be fixing that this year.

> although the speakers weren't mic'ed so the sound isn't so great - we'll be fixing that this year.

That's great, it was the only significant annoyance last time! Looking forward to see what people have been doing with this language last year.

Re: JuliaCon 2015 Preview

#15

Post by rrock »

elliott34 wrote: post #13

Is anyone on HN using Julia professionally?

Yes, I've used it for a number of academic research problems in biophysics. Our first manuscript that used it is in review right now.

To get "real work" done, I stuck with the 0.3 line. I've found it to be quite stable. Master is quite lively but should be settling down soon, at which point I'll migrate.

Re: JuliaCon 2015 Preview

#16

Post by sgt101 »

transfire wrote: post #9
idunning wrote: post #8

Earlier quoted context omitted.

No, I don't think so. I don't think there is much that is going to change with modules in the future, or at least I'm unaware of any open issues about it. What did you have in mind? The hackathon on Wednesday will be a particularly good time to bounce around your ideas, and is free to attend.

Because the current system ( http://julia.readthedocs.org/en/latest/manual/modules/ ) is just sad. It's the one thing that has kept me from fully embracing Julia. I would love to go if I were any where in the neighborhood. But alas I'm over a thousand miles away.

Some people I've talked to have complained that modules aren't objects, and that throwing away the coupling of types and reuse files is "abandoning 30 years of learning" and that "I don't trust myself to be disciplined"

I think that these arguments are not strong! But what would be useful would be some detailed critical thinking on the topic.

My take is that thinking about where you put multiple dispatch parameterized type matched code components is actually a real challenge. The module structure chosen for a library requires "artful" approaches. Perhaps a good workshop on the topic would yield a guide or an approach to making these decisions? Perhaps a labeling system could be evolved to provide metadata for modules better enabling reuse and promoting discover-ability ? Perhaps static analysis of a code base might be provided that gives recommendations (and assistance) for creating good decompositions and associations?

Re: JuliaCon 2015 Preview

#17

Post by porker »

sgt101 wrote: post #16
transfire wrote: post #9

Earlier quoted context omitted.

Because the current system ( http://julia.readthedocs.org/en/latest/manual/modules/ ) is just sad. It's the one thing that has kept me from fully embracing Julia. I would love to go if I were any where in the neighborhood. But alas I'm over a thousand miles away.

Some people I've talked to have complained that modules aren't objects, and that throwing away the coupling of types and reuse files is "abandoning 30 years of learning" and that "I don't trust myself to be disciplined" I think that these arguments are not strong! But what would be useful would be some detailed critical thinking on the topic. My take is that thinking about where you put multiple dispatch parameterize…

> But what would be useful would be some detailed critical thinking on the topic.

Detailed critical thinking on any topic would be wonderful. Perhaps I have attended the wrong conferences, but I rarely see presentations like this. We really should think and discuss more and preach our position less in conference settings.

Re: JuliaCon 2015 Preview

#18

Post by Argentum01 »

elliott34 wrote: post #13

Is anyone on HN using Julia professionally?

I used it professionally from 2012 - 2014, and I know of about a half a dozen companies either developing things in it or supporting it as part of saas offerings. Not that many, but each group that I know of made the switch after extensive reviews of other language options, and they've all had positive experiences so far.

Re: JuliaCon 2015 Preview

#19

Post by sgt101 »

sgt101 wrote: post #16
transfire wrote: post #9

Earlier quoted context omitted.

Because the current system ( http://julia.readthedocs.org/en/latest/manual/modules/ ) is just sad. It's the one thing that has kept me from fully embracing Julia. I would love to go if I were any where in the neighborhood. But alas I'm over a thousand miles away.

Some people I've talked to have complained that modules aren't objects, and that throwing away the coupling of types and reuse files is "abandoning 30 years of learning" and that "I don't trust myself to be disciplined" I think that these arguments are not strong! But what would be useful would be some detailed critical thinking on the topic. My take is that thinking about where you put multiple dispatch parameterize…

I can't reply to Porker! But.. in reply, yes, I agree, critical analysis is horribly lacking in science atm.

Maybe this is something a plenary could do well - what are the assumptions that underpin some of the things that we are doing? Can we challenge those? Can we pick out the complaints/problems that people outside have about what we are doing?

That's hard for any community to do...

BTW. for the record, I learned years ago that objects did not for a good modularity make...

Post reply on HN