News for hackers. Stuff that matters.
Re:Doing small network scientific machine learning in Julia ... by cycomanic on Thursday April 14, 2022 @02:56PM (#31031618)
Thank you for saying this. As an outsider I (I only tried Julia a few times) this is exactly the impression I get every time I read a post (and comments) about Julia on hn. Don't get me wrong, I believe there are lots of interesting things happening in Julia, but it is not a magic wand. Often it is characterised as python with the speed of c, based on some highly optimised benchmarks, while the reality is much more complex. I find this uncritical evangelism quite off-putting tbh.
Re:Doing small network scientific machine learning in Julia ... by ChrisRackauckas on Thursday April 14, 2022 @08:53PM (#31035466)
If all you read is the highlights of the accomplishments, then of course you're getting only the good stuff. Feel free to read all of the details: it's posted all over Github/Reddit/Slack/Discourse/etc. The same individuals share all of the nitty gritty performance issues of every aspect we've ever found.
Re:Doing small network scientific machine learning in Julia ... by ted_dunning on Thursday April 14, 2022 @08:04PM (#31035088)
Point taken.

I, speaking for myself, will keep this in mind next time I set myself to gushing about Julia.

Re:Doing small network scientific machine learning in Julia ... by cycomanic on Friday April 15, 2022 @03:34AM (#31037488)
Appreciated. I realise that part of this is just being human, if one likes something one wants to promote it. There is a real danger with overselling though. I got quite excited about Julia initially, but when I looked into it some more I found that it couldn't live up to the expectations that were set, and so I have become very sceptical towards all the news.