Fork me on GitHub
#off-topic
<
2018-09-12
>
emccue02:09:28

the closest thing that exists right now is hylang afaik

emccue02:09:56

it isnt clojure, but it is a "lisp for python"

emccue02:09:04

that being said, it isnt very good

emccue02:09:06

I think a better version of hylang would work with persistent data structures

mpenet08:09:00

fennel-lang is a good example of a clojure-like language to lua transpiler

mpenet08:09:53

it could be a good source of inspiration for the same for python, it generally made the right choices imho

mpenet08:09:56

The whole thing is less than 2k lines of (commented) code, a single file

mpenet09:09:47

but python is a slightly more complicated language than lua

Denis G11:09:14

Does anybody know what continuation means? It is somehow related to trampoline thingy, but still don't quite get it

Denis G11:09:33

Thanks! is the book already out? with the standard lib of clojure?

dominicm11:09:47

You'd have to ask @U054W022G 🙂

reborg12:09:15

Really really close!

reborg12:09:35

last chapter, then reviews, then out… late October my guess

manutter5111:09:37

There’s also a pretty good discussion at https://en.wikipedia.org/wiki/Continuation (not Clojure-specific)

Denis G12:09:19

Was reading the Continuation Passing Style page on wiki and was confused

john12:09:59

It's basically structuring a computation as a chain of callbacks

Conor12:09:26

If you want to understand continuations, you could do worse than playing the game Recursed (https://store.steampowered.com/app/497780/Recursed/)

Conor12:09:53

It's really good! And it has a whole lot of mind-bending, vaguely computer-y puzzles

Denis G12:09:42

@dannyt but does this imply that Haskell, MLs or any other language which supports currying does this by default, or am I wrong

henrik13:09:25

So, the EU link tax was approved: <Pay $2 to unlock link now>

manutter5113:09:08

That's why we have the letters "W", "T" and "F" 😞

john13:09:14

@denisgrebennicov not by default, but you could build them in any environment that has first class functions, I believe

joelkuiper13:09:18

it’d be funny to see how it’s going to implemented though. It’ll probably be just as terrible as the cookie walls and the GDPR weirdness (still can’t access some sites from the EU because of it)

john13:09:11

If you already understand js callback hell, cps can be easier to understand in that context: https://medium.com/@b.essiambre/continuation-passing-style-patterns-for-javascript-5528449d3070

Denis G13:09:43

callback heaven 😄 Thanks @john

henrik13:09:28

And the mandatory upload filter stuff was approved as well. So the media companies we have today will be the only media companies we’ll ever have.

henrik13:09:09

Of course, it’s up to the rights holders to choose who to go after, and they’re only likely to do so if there’s sufficient money in it. But it’ll work well as a form of entrapment: if your media aggregator startup/social media startup/<anything that has user generated content> startup grows big enough, you’ll most likely already be guilty.

john13:09:07

"only media companies we'll ever have" That seems to be a thing, right? It's like gentrification through regulation, where small devs can't afford to pay tech debt rent.

henrik13:09:28

On the other hand Google, FB and the others might decide that they don’t need links to European sites particularly, just check destination IP, and auto-block them, leaving them to work out how to promote their content on their own. Meanwhile, I predict a staggering rise of VPN usage within the EU.

joelkuiper13:09:38

yeah it’d probably steer clear of start-ups that offer anything remotely similar to a public facing input/upload button. Technically an upload filter is also a bit of challenge, like who will own that technology and decide if it’s “fair” … probably the court that nobody has the money/time/energy to fight even if they are right

henrik13:09:55

Upload filters could easily be the most complex, expensive, and ML heavy part of the entire system.

john13:09:10

sounds like a fantastic censorship machine too

joelkuiper13:09:37

yup, would need some really complex perceptual hashing, and a massively distributed lookup system (probably something k-nn) … and who is going to own/maintain that system and the authoritative data in it …

henrik13:09:27

Well, that’s what it is. The only difference is what is censored. They’re throwing terrorist content into the mix as well—you’re now responsible for that not appearing.

henrik13:09:58

Arguably, censoring terrorist content is a good thing, but it doesn’t make the technology any easier to deal with.

john13:09:00

what about anonymizing networks like Tor? Will the be banned?

henrik13:09:22

If the EU provided the filter in question, and took responsibility to maintain its correctness, then yeah, maybe.

joelkuiper13:09:36

at this scale … even being wrong 0,001% of the time potentially means millions of entries being unjustly filtered. So they’ll probably opt for recall instead of precision to minimize false positives, which gives all sorts of interesting adversarial scenarios to consider.

henrik13:09:59

I don’t think they’re going after Tor unless they see a massive shift to its usage.

henrik13:09:32

They’ll go where the money is. That’s the difference to China.

john13:09:36

I suppose if your a massive company making money on tor, you're no longer anonymous anyway

henrik13:09:56

Yeah, pretty much.

john13:09:04

But this'll just put further pressure on P2P networks that give us back control

john13:09:21

pressure to make them the for-real internet 2 thing

joelkuiper13:09:27

they’ll just own the exit nodes or put honey pots like they already do with high profile criminal cases

john13:09:09

If somebody makes a Youtube2.0 on ethereum, what regulator is going to come in and shut it down?

mpenet13:09:53

there's already something like that

joelkuiper13:09:53

how would that even work, the files still need to live somewhere (can’t just put them on the block chain) and who is going to front that storage?

mpenet13:09:59

can't remember the name tho

john13:09:20

This law is somewhat incompatible with the nature of p2p networks (that are free to make arbitrary functions)

john13:09:08

Yeah, it sorta already exists. DTube. On IPFS

joelkuiper13:09:18

most “internet on blockchain” use it as a file table pointing to things like ipfs or torrent magnet links, or some other addressable p2p storage.

mpenet13:09:27

dtube I think

henrik13:09:29

It was passed in the parliament to standing ovations and booing of the people who opposed it. From where we’re sitting, this all seems very naive, of course. They seem to think they’re going to able to get their hands on tons of profit with no repercussions. Short term, we’ll probably see smaller media houses dying of atrophy, against their will. The law is written to forbid voluntarily opting out of the law. The small media houses have to go after the link tax from their users. Not that smaller media houses weren’t already languishing.

henrik13:09:22

It’s sort of understandable, I suppose. The Internet has effectively commoditized news and media, and commoditization is effectively a race to zero profit margin. This is traumatizing to hundreds of years old industries.

john13:09:28

Information is older than hundreds of year old industries. Maybe copyright law is traumatizing to older things that are important.

henrik13:09:24

Yeah, I’m not saying I agree, just that it’s a predictable reaction.

john13:09:10

Yeah, I'm a little bit fundamentalist wrt hacker ethic and copyright law

henrik13:09:29

Commoditization (or the opposite thereof) is one of the reasons I like Clojure, by the way. It’s rare to have such a thing as a technological advantage in this day and age, particularly one that is as readily available as just picking a different language.

john13:09:46

It's also not cool how the WebRTC (webp2p) thing is still hamstrung 8 years later by requiring dedicated ice/stun servers

john13:09:31

But fixing that and making web p2p a first class thing may hurt the bottom line of the google types

henrik13:09:32

P2P Internet will get there, particularly since there are more and more forces set in motion to steadily increase the value of having a truly decentralized web. If you go looking for a dichotomy, the universe is very happy to supply one.

wizard 4
orestis13:09:40

The European Union Parliament voted overwhelmingly to approve a copyright reform that includes two hotly debated measures critics have dubbed catastrophic for an open internet. But while the vote was a disappointment to opponents, it is also far from the final word on the matter.

The vote on Wednesday is what the EU calls a negotiating position. The measure now goes to a process called "trilogue," which is a negotiation between the EU Council and the Commission, where the details could change again. Once an agreement is hammered out, the EU Parliament will vote again, likely sometime early next year.
source: https://venturebeat.com/2018/09/12/eu-parliament-passes-copyright-reform-but-fate-of-controversial-link-tax-and-upload-filter-uncertain/

henrik13:09:59

Dude, you’re breaking the law.

henrik13:09:30

Just kidding, you’re fine until January.

joelkuiper14:09:09

just curious, what would that look like? Would @orestis have to pay a tax to VentureBeat for this?

henrik14:09:13

Slack would have to pay. The rights holders will be relieved of the responsibility of going after individuals. It’s kind of like you being able to sue whoever made the road if someone crashes their car into yours.

joelkuiper14:09:15

I mean, assuming “tax” as in a portion of revenue you’d be fine I think (it’s not a fee I don’t think).

joelkuiper14:09:28

So Slack would have to pay a portion of their revenue to the individual companies of whom a link was shared on their platform?

joelkuiper14:09:19

And as in, number of links, or clicks or how would that work 😛 I mean Slack is maybe a weird example since it’s not typically “public” but a news aggregator would definitely be in trouble. Just curious what it would look like

henrik14:09:26

I’m not sure how the fee is set, but yes, essentially. Their content is being reproduced on Slack, therefore Slack has to license that content.

joelkuiper14:09:26

it’s a tax right? as in a percentage of the profit, not a flat-fee? So if the platform makes 0 money, they can sue you for exactly 0 money

john14:09:06

Well, then I guess amazon is off the hook 😉

john14:09:06

Their accountants have very creative ideas of "profit and loss"

joelkuiper14:09:19

right that was what I was thinking, you can get extremely creative with the accounting of this 😕

joelkuiper14:09:09

I mean most start-ups under venture capital effectively operate under a net loss 😛

henrik14:09:34

Right, so they’ll have to pay YOU 😉

john14:09:06

Well there's currently 644 million pages on the internet, each with usually a few dozen or hundred outbound links. Would there be that many today if each link cost a penny?

john14:09:48

I guess we should start building an AI content creator and start building link farms and get our tort game on 🙂

joelkuiper14:09:14

most of the hyperlinks are effectively click traps trough ad proxies these days anyway. Which point to a javascript bootstrapped site without any real (as in a hypermedia payload) content. True hyperlinking was gone long before this 😛

henrik14:09:33

It has BOTH attributes of a good business plan! Soulless and completely viable.

john14:09:15

We have a responsibility to our shareholders

pablore18:09:01

what emulator de you use to develop android apps? I only know about Genymotion, which is too expensive for me and Bluestacks which I haven’t tested, but I dislike that is targeted at gamers