Fork me on GitHub
#clojure-europe
<
2023-04-27
>
genRaiy06:04:30

Good morning. Here’s a footo.

👣 9
genRaiy16:04:29

:drum_with_drumsticks: one of your best Lee 🙂

simple_smile 2
borkdude07:04:16

morning. anyone else watching the conj stream today?

simongray07:04:26

unfortunately, I don’t really have the time :S

Mario Trost08:04:30

Yes, today and hopefully tomorrow too

imre08:04:52

it's a bit late for our timezone but I'll try my best

lread12:04:28

Yup! Looking forward to it!

Mario Trost08:04:15

Good morning!

thomas08:04:01

Morning and a happy Kingsday from the Netherlands.

🍊 6
mdiin09:04:15

Good morning 👋

schmalz10:04:21

Morning all.

otfrom13:04:23

if I needed to create some locally unique ids based off of existing ids that I didn't want to keep, do you think hash (https://clojuredocs.org/clojure.core/hash) would work or should I go to MD5 or SHA1?

otfrom13:04:17

My ids are all around 10 characters long, but I'd like the further pseudonymise them. (part of an overall defense in depth on the data, including, keeping it safe from leaking)

slipset21:04:32

@U050ECB92 mentioned in his talk that they use SHA256 for this IIRC

❤️ 2
ordnungswidrig13:04:56

I tend to use SHA-256 as MD5 is sooo broken nowadays.

❤️ 6
🔒 2
delaguardo13:04:07

I use https://cljdoc.org/d/io.github.nextjournal/valuehash/0.1.10/doc/readme as a generic hashing library. But if your ids are all strings then clojure.core/hash should work just fine.

❤️ 4
timo20:04:31

why is it not fine to use hash with non-strings?

delaguardo22:04:20

https://clojurians.slack.com/archives/C03S1KBA2/p1644083293795019 hash is not guaranteed to be stable across restarts

👍 2
borkdude14:04:28

@U4P4NREBY Did you mean "daaaaaatom"?

4
simongray14:04:49

Ah but apparently they didn't actually open source it.. Very confusing.

borkdude14:04:59

It's free but still black box

grav15:04:28

First time I've heard of binaries released under a OSS license ...

6
grav15:04:16

Are we allowed to reverse engineer? Or maybe patch 'em? Not sure ...

Ben Sless15:04:45

Is reverse engineered code a derived work?

simongray16:04:56

doesn't matter that much, we have lots of great open source alternatives anyway. It's so strange to me why they didn't just open source it.

ordnungswidrig20:04:35

It's not easy to open source something after the fact. (Disclaimer: not in the product team, so I'm making general claims here)

2
genRaiy20:04:15

It will open up some business models (eg third-party hosting) that were not possible before this announcement.

6
ordnungswidrig20:04:39

Interesting. I did not think of that.

genRaiy22:04:05

I started something a long time ago which fell over due to the licensing so it’s relatable although tbh a bit passé personally

agigao23:04:56

I was expecting this <3

simongray06:04:02

@U054UD60U How, though? I guess it could be an issue if it relies on a highly proprietary library or something, but what else?

ordnungswidrig07:04:03

I have not idea, but it could contain licensed interlectual properties or the source contains some company secrets you'd like to redact first etc. It's just more complicated than switching a github repo to "public" :)

grav07:04:39

need to grep for those swear words 😉

Ben Sless08:04:51

I'm interested in seeing the code from several perspectives One is writing a distributed system The other is writing a database (and datalog implementation) And I'm really curious to see a full on system developed by Rich, because it's probably the way he intended the language to be used. Could learn something 🙃

ordnungswidrig08:04:43

I'd expect a lot of transducers 😉

Ben Sless08:04:01

Datomic predates transducers and core.async

genRaiy14:04:18

My guess is that you will find some there - in fact, the motivating use cases may have come from Datomic.