fijal

As mentioned in the blog post the individual donation buttons are not a resounding success. I'm happy to sign contracts with corporate donors (or even individuals) that we'll deliver. My mail should be public, if not #pypy on freenode or fijal at baroquesoftware.com

I wish pilots flew for free so I don't have to pay so much for something that's so critical to my career. What's wrong with you? Do you genuinely believe that stuff like open source (especially critical one) should be created and maintained for free? By who?

I think you (or the journalist) are off by a few zeros. According to wikipedia it's 32-39km^3 and I think km^3 = 1e9m^3

Hi

I would say, ballpark, $100k. Email me at fijal at baroquesoftware.com

EDIT: the actual figure depends precisely on what you want, which parts of the stack have to be exactly how fast etc.

What's wrong with that code? Runs acceptably fast under pypy

in short - funding. If we can find someone who wants fast numpy AND fast python under the same hood, we can combine the approaches of cpyext and numpypy and make it fast. The project is just too big to do on spare time. I've been trying to find some funding for that for quite a while, but I haven't been able to find any sizable backer just yet.

Cheers, Maciej Fijalkowski

numpy, pandas (mostly) and (scipy) mostly work through cpyext. We're considering how to make them fast now. It would never be faster on large_array + large_array, but it should be faster (by a lot) with element access etc., bit like numba

We're working on it :-) In fact there are two projects - one is to just remove the GIL, one is to add STM-capabilities.

numpy works, scipy & matplotlib are very close. Stop spreading false information please.

That said, the cpython+cython solution is not going away any time soon.

Hi

Yes, absolutely. This is why it's marked as "beta", so not absolutely all the features are implemented.

well, I'm very sorry you wished we stopped supporting our users, but this is the reality which we'll be stuck in for a while. python 3 adoption has been speeding up, but still the vast majority of old big projects are based in python 2 and those are potential uses for pypy.

no

Hi

I've tried optimizing sympy and it's hard - there is a chance that it uses cython (which will slow things down on pypy) but also the gains were small (up to 2x) and only noticable if you run it for a while (at least few seconds).

SymPy is a bit of an example that while python as a language is not necesarilly slow, since you can make a plan for most parts, it breeds culture of unnecessary meta-programming, tons of copies and dict lookups that makes it REALLY HARD to optimize. ORMs go in the same category

yes, it "mostly works". There is an issue about graphic backends (you can either save it or use jupyter), but otherwise yes.

Great conference :-) can't recommend it enough

Maybe you're the one who can write a check, but this is a difficult topic...

I do corporate accounting for baroque software and I can write an invoice and pushing people to pay for it is still quite a sales challenge - you need to be able to talk to high enough people and prove the value of what you're offering on top of the software you are already giving for free. It's not different to a standard issue sales exercise, except a little harder. This is not a skill that most open source project maintainers would consider themselves experts at.

Giving advice like that is very dangerous - "only if you did that, you would make money". It's not a given that it would lead to something more than couple hundred dollars.

Sounds like a description of many politicians. Or businesman. What is needed is a system that, game theoretically, shows those people that you can be a complete asshole as long as you are within certain rules of the law. Then you don't get locked up and that's the deal.

I think the whole point of the article is to showcase that maybe you don't NEED all this additional security. I purposefuly try to avoid going to the US because of the security and bureaucracy involved in crossing this border in the first place

I'm not really commenting on the relative merits of the technology or who should use what. It's about the lack of admission that there is other stuff and other requirements

I mean, it's obviously unfair. Even in the scientific community someone uses PyPy. As far as download statistics go, I think between 0.5-1% of downloads from PyPI are for PyPy. This is far from "majority" but far from "noone" as well. For what is worth enough big players use PyPy that the consulting company that I run can stay afloat, so I somehow doubt that's the case :-)

I must say I'm always a little annoyed by the split in Python community around the scientific stack vs the rest. There are two ecosystems, two packaging tools etc (which is fine), but the insistence that there are no other python users that are worth considering (both sides are guilty) is really frustrating.

Hi, I represent PyPy project. We don't have personal connection to mozilla nor we're successful enough to pay engineers full time. Can you explain some of your accusations?

No, but "I donated $10 3 years ago and you guys are SLOW at responding to my queries!" is an incredibly common attitude. Open Source has shown that it's a far superior development methodology, but there is definitely some negativity associated with "expect stuff for free"

TL;DR; it's a visualizer for cProfile (with all the cProfile problems)

DISCLAIMER: I'm an author of cProfile competitor, vmprof

I would like to complain about name similarity to vmprof (my project)

I think you're confusing cause and effect. In open source world, you can't charge for something that exists. It needs to be a need to do so. In order to push PyPy into putting significant effort from a core team - we would need exactly that. A big chunk of money dedicated to just that. We've spent all the money we got and we did get something out of it. But it needs to be more seriously supported in order to kickstart that. Obviously people are free on whatever they want in their free time, but unless there is enough to push the PyPy3 support to a significant postion, it's unlikely to move at a higher pace.

Did you actually look at the blog post? The novelty is the idea that you generate a JIT compiler from the interpreter using a few hints. Truffle does that, but Truffle is newer than PyPy.

There is a similarity, but there is also a difference. The general difference (among PyPy/Pyjion/Pyston) is "what to do with CPython C API". As far as I understand:

* Pyjion says "we gonna add a JIT API to CPython" - so no touching the C API at all (or the object model)

* Pyston says "we gonna implement our own object model supporting CPython C API", but they seem to have copied large portions of CPython. This is a semi-rigid approach compared to Pyjion, because it's not runtime-swappable

* PyPy says "we gonna implement a fake layer on top of our own object model" which means that while C API is feasible (and mostly implemented), it's gonna be always a compatibility layer with slower approach, but the underlaying data types/GC can change at will

Hope that helps

They think (from what I understood) that the limit is "above pypy performance"

The talk is from Dec 2015, has already been on hacker news I believe (but can't find the discussion)