Live data from Hacker News

Startups are building with the Julia Programming Language

juliazoid.com

31–40 of 42 posts

Re: Startups are building with the Julia Programming Language

#31

Earlier quoted context omitted.

I've been hearing about improvements for the last 8 years. It has been a constant complaint and a constant broken promise.

Since 1.0 it has gotten way better. There's still work to do (obviously) but there has been continual improvement.

My concern for TTFX is the case when X is something infrequent but sensitive to latency? This a ridiculous example but when it is time for the anti-collision feature on my car to kick in, I want it to run, not compile. I think the right answer is to run a bunch of tests to get all the paths compiled and then build a new system image. But I haven't found a ton of guidance on how to do that. Seems like a perfect pipeline task.

Re: Startups are building with the Julia Programming Language

#32

Earlier quoted context omitted.

Since 1.0 it has gotten way better. There's still work to do (obviously) but there has been continual improvement.

My concern for TTFX is the case when X is something infrequent but sensitive to latency? This a ridiculous example but when it is time for the anti-collision feature on my car to kick in, I want it to run, not compile. I think the right answer is to run a bunch of tests to get all the paths compiled and then build a new system image. But I haven't found a ton of guidance on how to do that. Seems like a perfect pipeli…

That’s the idea. Here’s an example: https://discourse.julialang.org/t/a-julia-dataanalysis-sysim...

Re: Startups are building with the Julia Programming Language

#33

Earlier quoted context omitted.

Jumping the gun slightly, but TTFX likely about to get a lot better (5x+ improvement in times) in the next version (1.9). Via cachine of the native compiled code. https://github.com/JuliaLang/julia/pull/47184#issuecomment-1...

I've been hearing about improvements for the last 8 years. It has been a constant complaint and a constant broken promise.

I'm not sure that there's been any "promises" to break. Folks have been saying that it's possible but it'll just take time and effort.

Re: Startups are building with the Julia Programming Language

#34

Earlier quoted context omitted.

Jumping the gun slightly, but TTFX likely about to get a lot better (5x+ improvement in times) in the next version (1.9). Via cachine of the native compiled code. https://github.com/JuliaLang/julia/pull/47184#issuecomment-1...

I've been hearing about improvements for the last 8 years. It has been a constant complaint and a constant broken promise.

Indeed, every couple of years I try julia again, and every time it's still slow as molasses, all the while being told it's massively improved. One of the most annoying aspects of slow TTFX is when X is your language server. It takes like 10 minutes or something ridiculous for completions to start working after opening a moderately complex file. I had the same experience in vscode and emacs.

For others who feel they've been gaslit by the julia community, this blog post [1] resonated with me, especially the 2022 update.

[1] http://danluu.com/julialang/

Re: Startups are building with the Julia Programming Language

#35

Earlier quoted context omitted.

I've been hearing about improvements for the last 8 years. It has been a constant complaint and a constant broken promise.

Since 1.0 it has gotten way better. There's still work to do (obviously) but there has been continual improvement.

I think Jevon's paradox is also hitting to some extent, coupled with the ecosystem getting bigger, so a typical project may have more transitive dependencies behind it. The functionality is going up, and the speed of compilation is going up, but the total latency is staying similar or increasing, in my experience.

Re: Startups are building with the Julia Programming Language

#36

Earlier quoted context omitted.

Since 1.0 it has gotten way better. There's still work to do (obviously) but there has been continual improvement.

I think Jevon's paradox is also hitting to some extent, coupled with the ecosystem getting bigger, so a typical project may have more transitive dependencies behind it. The functionality is going up, and the speed of compilation is going up, but the total latency is staying similar or increasing, in my experience.

that's definitely possible. I'm somewhat hopeful that some changes in 1.9 will let us cut back on transitive dependencies a lot, but it will take a few months to know for sure.

Re: Startups are building with the Julia Programming Language

#37

Earlier quoted context omitted.

I've been hearing about improvements for the last 8 years. It has been a constant complaint and a constant broken promise.

Indeed, every couple of years I try julia again, and every time it's still slow as molasses, all the while being told it's massively improved. One of the most annoying aspects of slow TTFX is when X is your language server. It takes like 10 minutes or something ridiculous for completions to start working after opening a moderately complex file. I had the same experience in vscode and emacs. For others who feel they'v…

Would it not be better at this point to move on?

Re: Startups are building with the Julia Programming Language

#38

I'm a huge julia fan, but man is this ever an insubstantial and unconvincing listicle.

Yeah, it doesn't really seem intended for a HN type audience. More of a high level overview of the language to show to a non-technical co-founder or senior executive.

Re: Startups are building with the Julia Programming Language

#39
post #10

I heard a talk from a large cybersecurity company that was using Julia for research and is apparently in the process of dropping Julia due to: - difficulty in hiring, apparently it was too expensive/difficult to hire Julia programmers, - poorly performing garbage collector. If I recall correctly, they had issues when trying to process huge amounts of data. Seems like small teams will not run into these issues at leas…

Please put them in touch with me if possible. RelationalAI has similar GC issues which they are the core Julia developers are working hard to resolve. I have also seen zero postings for Julia roles in cyber security companies so it is likely that they aren't tapping into the community which is why they can't find people.

It was a talk by folks from this team: https://gradient.avast.io/people-behind/

Re: Startups are building with the Julia Programming Language

#40
post #10

I heard a talk from a large cybersecurity company that was using Julia for research and is apparently in the process of dropping Julia due to: - difficulty in hiring, apparently it was too expensive/difficult to hire Julia programmers, - poorly performing garbage collector. If I recall correctly, they had issues when trying to process huge amounts of data. Seems like small teams will not run into these issues at leas…

My (more or less extensive) experience with Julia has been rather good. Step 1. Prototyping. Small data, don't care about performance. Julia is as easy as python, but faster. GC doesn't matter. Step 2. Production. Very large datasets. Performance is very important so I make sure that hot loops don't allocate, and so GC doesn't matter. Really, I struggle to see how someone can not like this. Regarding hiring. My exper…

I am not exactly sure what they were working on, but I am not convinced that any code can be made non-allocating easily.

For example, when you are building large trees with complex structures in the nodes and you need to both add/remove the nodes, the effort to make it non-allocating might be too much compared to just using C++ and writing things more naturally?

Post reply on HN