[jpfr]

I'm surpised. At JuliaCon ~6 months ago the message was that no 2.0 is in the works [1]. I.e. no backwards-incompatible changes to the language.

I checked the usual places and did not find any information on 2.0 and an interface mechanism. Do you have a pointer?

[1] https://youtu.be/N4h46_TCmGc?t=1656

[adgjlsfhk1]

No, he's just making stuff up.

[adenozine]

Oh, well here's a post by Jeff Bezanson, since you've decided to just talk out of your ass...

https://discourse.julialang.org/t/what-dont-you-like-about-j...

Hope the uh... co-creator warrants enough merit as a source.

[jpfr]

So the 2 year old post by Jeff on discourse says that breaking changes would have to go into an eventual 2.0.

Whereas in the 6 month old keynote the same Jeff says there is no plan for breaking changes and hence no 2.0 is planned.

So Julia will remain stable. And maybe we get interfaces (or rather traits) at some point in the 1.x release family.

[adenozine]

Sure, if it happens, then I'll change my tune pretty quick.

I don't stay plugged in beyond reading the changelog whenever a new version comes out, so I didn't know that bit about the keynote. I've noticed as well that the Julia community produces an absolutely extraordinary amount of conference talks/video content, I don't have the time for the finer details.

I've also noticed a distinct and crucial lack of long-term vision for Julia from the co-founders. I've also read some dramatic (unverified) claims that I won't repeat here about one of them. Frankly, I think a better group of people could be assembled to steward the language and its ecosystem, but I don't see much chance of that happening. Julia exists in a weird little place where being terrible to read and maintain doesn't matter as long as it lives up to its promises of being very, very fast, in most of the applications it is used for. That's all well and good, but you don't build a solid foundation for a community to really depend on that tool for bedrock tasks.

I've been unfortunate enough to read some .jl code in '22, and it was dreadful. I truly don't understand how multiple dispatch makes anybody's life easier, it's an absolute nightmare of unmaintainable code that calls any of dozens or hundreds of methods, the performance of the entire application essentially dependent on whether or not that type is stable and of course, there's no way to know that without reading the definition of the unknown method that gets called.

Personally, I have my eyes peeled for github.com/exaloop/codon for higher performance stuff with Python. It's already an order of magnitude faster than pypy for most cases, and equally more usable for practical work than Julia, imo.

Anyhow, when it's all said and done, there's a lot of computing being done and a lot of money changing hands and so forth. All's well that ends well, despite never really being done well. /shrug

[DNF2]

> I don't stay plugged in beyond reading the changelog whenever a new version comes out ... > I've also noticed a distinct and crucial lack of long-term vision for Julia from the co-founders.

Don't you find those two statements contradictory? There's a pretty striking contrast between your claim to ignorance and your confidently sweeping generalization (sadly, those two often do come in pairs.)

> I've also read some dramatic (unverified) claims that I won't repeat here about one of them.

If this is what I think it is, the claims were not substantiated in any way (even though it would be easy), and seemed quite outlandish, frankly.

> I've been unfortunate enough to read some .jl code in '22, and it was dreadful. I truly don't understand how multiple dispatch makes anybody's life easier

It's really hard for me to understand this opinion, given that Julia code, to me, is far more appealing than all the most common alternatives. In particular, multiple dispatch is such an obviously natural paradigm, that it's just hard to fathom why everyone don't just 'get it' right away. I mean, not taking all input argument types into account now seems to me like a completely artificial, even perverse, restriction. Why?

I guess this is why people argue on in the internet.

[adenozine]

Ah, well I should’ve clarified. I was very, very excited about the language from about 0.4 to 1.2ish.

In that time, I was nearly obsessive in reading all the news, though I wasn’t writing any jl code at all during that time. After 1.0, I used it for a few little things here and there in my business, some basic csv munging and a few other one-off tasks. It did fine, ofc.

The claims, yeah I mean, I don’t know the parties involved personally but because we are not far apart in the social graph and I don’t need the headache this close to retirement. I love my opinions very much, but not when they involve personal controversy.

Multiple dispatch is a write-only benefit. In my line of work, I might come across different machines, different memory setups, different instruction sets, different c compilers, different latency profiles, different spoken language as documentation, etc.

I’m not a mathematician, but I clean up and leave businesses with pragmatic, clean, maintainable code after the mathematicians do their thing and move on. For me, good code is about how specifically and how clearly I can communicate an idea so that the next person (who won’t know shit about shit) can add a feature without breaking something, change a deploy, fix a small bug, update documentation, etc, without breaking things. In that regard, Julia code reads like someone’s manuscript about the issue and not a program written by a programmer.

To you, that’s desired. To me, that’s my actual worst nightmare because it means the longest mental checklist between getting onsite and leaving a finished product and getting paid.

You mention not taking all argument types, it’s just so stereotypically mathematician of an assumption that the code will just work with whatever garbage you send it.

Because maybe that function needs to run on a 32bit system somewhere, and it will fail?

Maybe some of the other methods have a bug, and you won’t notice until you use the original method with a certain type, that you hadn’t considered? Julia can’t prove this sort of thing, to my knowledge.

I’m not saying python can, necessarily, but I know that there are very strong confidence levels and that if something borks, I can arrive at the scene of the crime within a few minutes of any given stacktrace and docker image, or whatever.

I don’t mean to argue. It’s clear we disagree. That’s fine. I just can’t take Julia seriously because they treat data like math on a whiteboard, and it’s just a ridiculous way to program a computer. Like I said, it gets a lot of computing done, and a lot of people are going home paid well. It’s all good, to that end.

Edit: I likely won’t reply further, it’s really annoying to scroll back in HN to see replies without linked notifications. Be well!

[adgjlsfhk1]

You misread (although it wasn't especially clearly phrased.

"Interfaces are a really common topic of discussion and I think at this point we’re determined to do something about it in julia 2.0 (if it requires breaking changes)."

This means we really want a solution for interfaces and if we had a good enough design for interfaces that would require 2.0, they are important enough that it could be worth breaking existing code (and releasing a 2.0 with interfaces). However there still isn't a plan for interfaces (breaking or non-breaking).

[adenozine]

If we are to continue this conversation, I need it to be addressed that you falsely characterized my top level comment with “he’s making stuff up.”

Otherwise, further correction would be wasting my time.