Fork me on GitHub
#off-topic
<
2018-01-09
>
qqq10:01:22

how do I tell git : dump all versions of file ss_web.clj for me ? (it's in git history, but it's not in current tree)

chrisblom11:01:21

does anyone know a simple database for timeseries? I'm looking for something that is very stable, and can quickly perform queries like "what is the sum of events X with attribute Y over the last 2 hours", and requires a minimum of ops

qqq11:01:46

never used it myself, but finance people swear by it

mpenet12:01:08

it's not simple nor cheap tho

qqq12:01:22

"cheap" was not a requirement, and simple is a matter of whether you like APL šŸ™‚

mpenet12:01:24

depends on your ingest rate, number of entries you expect, volume etc I guess

chrisblom12:01:15

yeah, i'd like to learn kdb, seems really powerful

chrisblom12:01:29

but i don't think we have the budget for it

qqq12:01:33

the most important question is probably one of: do you wnat to lineraly scan all events from past 2 hours to get relevant ones, or do have some type of inverted index

chrisblom12:01:44

i'd like to store and an increasing sum indexed by (timestamp, id,attribute)

chrisblom12:01:44

i don't need all events

andre.stylianos12:01:11

depending on your needs, maybe https://www.timescale.com ? I heard of it a couple of months ago

chrisblom12:01:40

thats a postgres plugin right?

chrisblom12:01:37

cool, i'll check that out

qqq20:01:35

anyone else gettng weird slack "something with websocket; please restart your browser" msgs ?

noisesmith20:01:23

yeah, if twitter is any guide this happened to a lot of people for about half an hour

noisesmith20:01:36

(across slack, not just this instance)

jgh20:01:26

if only slack added more blockchain this would have been prevented

noisesmith20:01:44

wouldnā€™t federation plus tls suffice? how does proof of work help?

jgh20:01:57

you can have a blockchain without proof of work, other proofs can work, or no proof at all if you trust eachother - but at that point itā€™s just an impractical database.

noisesmith20:01:38

oh because it also has a history so itā€™s not just federated messages, itā€™s a federated archive

jgh20:01:40

a blockchain-based chat app probably wouldnt be too awful as long as it can keep the transaction rate fast..i wonder if there are blockchain algos that allow you to get a digest of the current state of the world and then backfill rather than having to download everything

jgh20:01:24

but i think the assertion that slack could ā€œeasily decentralizeā€ is what i found funny

sundarj20:01:18

could Slack not utilise Matrix?

noisesmith20:01:27

OK - poeā€™s law applies here - I hear many insane assertions about the utility of distributed services and block chains and such so I honestly canā€™t tell parody from earnest advocacy

noisesmith20:01:42

or, simpler, we could use Matrix instead of Slack šŸ˜„

sundarj20:01:53

good point

noisesmith20:01:39

in fact, there is a clojure community on Matrix

jgh20:01:49

im not familiar with matrix

jgh20:01:28

oh i see, does it use something like consul/gossip for consistency?

sundarj20:01:28

i tried their web client and it was ok. i still wanna write my own at some point though šŸ˜

jgh20:01:20

well seems like itā€™s open source so nothing stopping you šŸ˜‰

jgh20:01:10

what the world needs is a slack alternative built on blockchain that rewards people for generating blocks while chatting šŸ˜®

jgh20:01:28

proof of chat

sundarj20:01:56

spend 5 eth to add emoji

jgh20:01:57

(if anyone takes that idea they need to give me some ico coins or whatever)

jgh20:01:07

copyright (c) 2018 jgh

noisesmith20:01:24

OK Iā€™m gonna have to cite poeā€™s law again here

jgh20:01:01

it could work, just have to figure out how to stop people from spamming to improve their chances of minting coins šŸ˜‰

seancorfield21:01:57

Re: connection problems with Slack -- they do have a status page ya know...

seancorfield21:01:44

(and it was showing red on messaging and connections when the outage first started -- and then red on everything for a while)

seancorfield21:01:21

Re: blockchain -- :face_with_rolling_eyes:

seancorfield21:01:38

(the crypto people are so amusing)

naomarik21:01:09

@seancorfield whatā€™s your full opinion? šŸ™‚

seancorfield21:01:26

Hahaha... not sharable within the Code of Conduct here šŸ™‚

naomarik21:01:00

hehehe, iā€™ve been putting it down for 2 years now. iā€™m a bit surprised it has spread to the extent it has now

naomarik21:01:24

enterprise software sucks? oh look a new shiny thing called blockchain!

naomarik21:01:42

wondering how long this will take to blow over

naomarik21:01:56

iā€™ve been in the fray in UAE with all this. itā€™s crazy how many people think that blockchain is the miracle solution that will automate their all their processes, and no alternatives ever existed up till now.

jgh21:01:45

thereā€™s definitely a weird mysticism built up around it right now

jgh21:01:24

which, i think, is being propagated by cyrptocurrency enthusiasts/investors to try to prop up their investment. If people believe blockchain is the magic elixir then it drives the cryptocurrency narrative.

jgh21:01:03

it helps too to drive home this mysticism that the cryptocurrency community has developed dense and occasionally nonsensical jargon

jgh21:01:52

blockchain is simply a public/distributed backwards linked list. To even call it a database is to overstate its functionality and purpose.

jgh21:01:43

in my mind itā€™s definitely a valuable addition to computer science, but I donā€™t think itā€™s the elixir people are selling.

noisesmith21:01:51

in before ā€œquantum blockchains to fix branch prediction securityā€

jgh21:01:27

the unfortunate side-effect that seems to come with the mysticism around blockchain/cryptocurrency is the assumption that only geniuses can understand it. Us mere mortals are eternally bound to ignorance. And that leads to absurd statements like the one above. The guy posting that, if he was serious, doesnā€™t really understand what slackā€™s scale problems are (and Iā€™m sure none of us here truly do either) and is just prescribing a cure out of ignorance.

jgh21:01:28

Disregard the bugs that lead to hundreds of millions of dollars in theft, the Ethereum team are infallible.

noisesmith21:01:58

itā€™s not theft, the blockchain says itā€™s a legitimate transfer of wealth

noisesmith21:01:22

itā€™s just that the people sending the wealth did it by mistake šŸ˜›

jgh21:01:23

right, well the people on the other end of that transaction may disagree with that assessment but the blockchain itself doesnt know the difference!

noisesmith21:01:36

donā€™t forget the part money plays - thereā€™s no proposition so stupid that you canā€™t convince someone to agree for money

sveri22:01:28

The nice thing about blockchain is, is that it is something where a lot of people talk about it and like 99% talking about it, really dont know what it is. All they have is an imagination in their head about what they think it can do. So blockchain is kind of a personalised IT reality dream show or something like that. And of course it also has a lot of positive connotations, like being super secure + some more bullshit bingo.

sveri22:01:50

Really an interesting topic from a psychology point of view.

jgh22:01:00

yeah thats where i was going with the mysticismā€¦itā€™s just sort of something people project their tech fantasies onto, regardless of how realistic that is (which is, often, nonsensical).

jgh22:01:24

the super-secure thing is a red herring too, itā€™s just a conflation of cryptographically-secure digests vs encrypted data

jgh22:01:21

yeah thatā€™s why iā€™m careful to be specific about cryptocurrency vs blockchain vs cryptography because theyā€™re all different things.

jgh22:01:29

to be fair though hashing and digests come from the cryptography field, and theyā€™re fundamental to the functioning of a blockchain, so i get the naming.

jgh22:01:16

but the ā€œcryptoā€ part in cryptocurrency doesnt mean your data or identity is secure, just that the integrity of the blockchain is secure (caveat: as long as someone doesnt own 51% of the block generation capability)