|
00:04
halloy7773 left
|
|||
| Voldenet | `random-variate(dist => 'bernoulli', :p(0.3), size => 1000)` looks nice but it'd be needed to be built into the library, which would be tedious | 00:38 | |
| I was going on about having a namespace, for which namespace would have precedence over other usable symbols | 00:42 | ||
| `class Thing::One is Thing { }; class Other::One is Other { }; sub foo(Thing x) { }; sub bar(Other x) { }; bar(One.new); foo(One.new)` | |||
| it's really weird to read in the context, because I used colliding symbols and meaningless names | 00:43 | ||
| hm, I could use Dist::Uniform and Noise::Uniform perhaps | |||
| but that'd be property of Noise using uniform dist, no ideas come to mind | 00:44 | ||
| either way, that does simplify the code | |||
|
00:47
alati joined
|
|||
| alati | hi all... I hope things are well with you | 00:49 | |
| Voldenet | same, good day to you | 00:50 | |
| I mean, it's 3am where I sit, but it's still day | |||
| alati | :) good day | ||
| same but... you know, reminiscing and such | |||
| I remember Util was pretty active before. Perhaps it's not the time of day or week day to be on irc | 00:51 | ||
| it's nice that the community still has a prescence here though... | |||
| anyway, I shared on #-dev I'm working on using raku for daily tasks. definitely I feel like using it for a custom DSL and bdd work would be fun so | 00:53 | ||
| honestly I came because I was thinking of Util | |||
| but I do think raku would be extremely useful in qa work... | 00:56 | ||
| Voldenet | yeah, there are features that severely reduce complexity, people don't mention it often but ability of not using () for function call really makes code a lot nicer | 01:05 | |
| alati | kinda funny you mention the lack of () because honestly that sort of boggles my mind still | 01:06 | |
| [Coke] | Util still works with raku. I assume he'll be giving another talk at the next raku/perl con for NA in CAN in 2027 | ||
| alati | I suppose it's only been this week I got back in to raku | ||
| [Coke] | he's not often on irc though | ||
| alati | :D | ||
| well, good news overall then | |||
| thank you both for being available | 01:07 | ||
| Voldenet | well, dropping () is non-mandatory, but it feels right in a lot of places | 01:08 | |
| though you can write java in raku as well | |||
| or rather, code that structurally resembles java | 01:09 | ||
| alati | that's true, I admit I got fairly comfortable with leaving () out in a lot of python after some years but it still getting used to it | ||
| originally I was on ##php in the old node | |||
| and knowing raku can be like java in the regard.... could be interesting for my work. not personally but in the industry | |||
| Voldenet | welp, I used some python before and () wasn't too bad | ||
| alati | I can't say I work in java at all these days, I'm just helpdesk | 01:10 | |
| automation and ticketing | |||
| Voldenet | I'm glad that new compilers enable embedded raku scripting | ||
| alati | oO | ||
| Voldenet | (rakupp and mutsu, both) | 01:11 | |
| alati | ah yeah mutsu | ||
| err... I could search it but just to speak it to the ether here, raku + wasm for browser stuff? | 01:12 | ||
| I've got a friend doing some webgl and I guess vulkan things in browser | |||
| trying to argue that raku could be helpful in the project ;) but that's just me being obsessive likely | 01:14 | ||
| [Coke] | anton: Data::Generators | ||
| died in a blin run with: | |||
| Voldenet | Well, rakudo on moarvm was compiled to wasm and working, but it was a bit slow | ||
| [Coke] | [Data::Generators] # Failed test at t/05-random-variate.rakutest line 72^M | ||
| [Data::Generators] # expected approximately: 0.4^M | |||
| [Data::Generators] # got: 0.3884^M | |||
| [Data::Generators] # maximum absolute tolerance: 0.01^M | |||
| [Data::Generators] # actual absolute difference: 0.0116^M | |||
| Voldenet | otoh rakupp and mutsu both give you some js wrappers, I'm betting the libs they provide are using wasm | 01:15 | |
| [Coke] | assuming that's a flapper | ||
| Voldenet | the raku.online/ one and tokuhirom.github.io/mutsu/playground.html one | ||
| alati | I like it | ||
| oO | 01:16 | ||
| Voldenet | yeah, they both are using wasm, just checked | ||
| alati | bit of an exciting feeling | ||
| can't believe it's actually 417 now... sleep schedule gone awry | 01:17 | ||
| i blame andor | |||
| Voldenet | and I've planned to do this on weekend :> | 01:18 | |
| alati | :D | ||
| Voldenet | that's what weekends are for | 01:19 | |
| alati | I suppose they are | ||
| I feel soon every day will be a weekend for me however | |||
| office closure | |||
| nothing official on my position which is interesting | |||
| not to detract, weekends are indeed where the mind wanders | 01:20 | ||
| it's nice | |||
| and I'm happy to hear Util is active in general | 01:25 | ||
| I have yet another silly idea: selinux management tools but in raku | 01:28 | ||
| I swear I'll stop it's just... obsession | |||
| 20 years and I'm still behaving pretty much the same way | |||
| Voldenet | well I've used iptables for a long while | 01:29 | |
| alati | and here I'm still learning iptables | ||
| Voldenet | and the syntax is far from user friendly | ||
| alati | yea but now? nftables | ||
| I think iptables is nice enough to translate but it's still 2 different things right? | |||
| Voldenet | yeah, nftables are so much better | ||
| for a lot of features it's mostly just new syntax | 01:30 | ||
| alati | I do keep hearing that... at work we still have deb9 soon to be decommissioned but | ||
| you know | |||
| Voldenet | turns out that "just new syntax" is a lot | ||
| alati | oh yeah, speak better work things out | ||
| Voldenet | people have been hating on genAI, but I like it | ||
| alati | same same | 01:31 | |
| Voldenet | I can now, instead of printing things to console and having my head hurt, just make a nice UI for it, with d3 graphs and so on | ||
| alati | I tried local models but you still have to build a rapport on with them | ||
| I admit using chatgpt | |||
| and mistral | |||
| Voldenet | local models work, but hardware I have is very limited compared to what claude, chatgpt, copilot or deepseek have | 01:32 | |
| eh, I also use built-in google one sometimes, but it's very limited | |||
| I think it's gemini flash | |||
| actually worse than local models | 01:33 | ||
| alati | :D that's a new one for me, cool though | ||
| worse xD | |||
| sorry to hear | |||
| Voldenet | But it's good if you only need answer to a simple question | ||
| alati | hm sort of a simpler rag I suppose | ||
| Voldenet | like "i don't remember name of the tool that did X" | ||
| alati | I did try with chatgpt making a simpler nn/nlp/rag modeling | ||
| yes! | 01:34 | ||
| even worse, the name of a song | |||
| Voldenet | I remember being shocked how well grok did with media | ||
| alati | I have recovered certain classics that I vaguely remember and it just happens to help me find them | ||
| indeed a fun tool | |||
| hm | 01:38 | ||
| the world we live in I suppose :> | |||
| ...just one more maybe? | 01:41 | ||
| proxmox has some sources online and it relies on some perl modules | |||
| could upgrade it to use raku instead :))) | |||
| pmx9 hah... | 01:42 | ||
| but seriously I have rediscovered raku language and I think it is inspirational | |||
| for me at least | |||
| especially with DSL, I'd rather sandbox some code I don't trust and have a system to audit | 01:43 | ||
| Voldenet | yes it feels nice to write in, feels a bit like ruby but more familiar syntax | 01:47 | |
| alati | please forgive me for an irrational disgust of ruby | 01:50 | |
| I tried it for a few weeks not out of work obligation | |||
| it's different | |||
| would rather a different management system | 01:51 | ||
| it's a little funny because I helped train an old colleague in coldfusion but he then went on to ruby for medical and... you know more power to him | 01:52 | ||
| I just couldn't | 01:53 | ||
| inflexibility is a problem | |||
| maybe | |||
| I need to be nice to myself | |||
|
02:00
hulk joined
02:01
kylese left
|
|||
| alati | is it ok to ask about the perlmonks site? | 02:02 | |
| whats with the ssl issue | |||
|
02:15
hulk left,
kylese joined
|
|||
| [Coke] | coldfusion? there aren't many of us around. :| | 02:28 | |
| alati | :') | 02:29 | |
| I can't actually admit to even doing modules, just basic cf stuff | 02:30 | ||
| I swear I'm only awake currently because coffe and pizza | 02:31 | ||
| 5am is the magic hour | |||
| sant'e | 02:37 | ||
| also because of a poor schedule | 02:42 | ||
| need to cron my life | |||
|
02:45
alati left
02:48
alati joined
|
|||
| alati | small reboot | 02:48 | |
| hm... I thought it was small | 02:58 | ||
|
05:07
alati left
05:13
abraxxa joined
05:17
alati joined
05:30
human-blip left
05:35
human-blip joined
06:01
rnddim joined
06:02
jcallen_ joined
06:03
avuserow_ joined
06:04
saltine joined
06:06
ShimmerFairy left,
rnddim is now known as ShimmerFairy
06:07
tinita_ joined
06:08
alati left,
jcallen left,
avuserow left,
jjatria left,
tinita left,
erry left
06:14
erry joined,
jjatria joined
06:15
kjp left
06:19
kjp joined
|
|||
| disbot | <jubilatious1_98524> @Voldenet , from the R-help pages: "dunif gives the density, punif is the cumulative distribution function, and qunif is the quantile function of the uniform distribution. runif generates random deviates." | 06:43 | |
| <jubilatious1_98524> Maybe something like this? unix.stackexchange.com/questions/7...681#790681 | 06:50 | ||
|
07:27
johnjay left
08:06
Sgeo left
09:08
abraxxa left
10:10
belluzj joined
11:33
belluzj left
|
|||
| disbot | <antononcube> @Coke Since Rakudo does not have reproducible random number generation the random data generation tests of "Data::Generators" occasionally fail. Usually, a second run makes those tests past. | 12:46 | |
| <antononcube> I can change the test files to exactly, that. | |||
| <antononcube> On the other hand, Rakupp does have reproducible random number generation -- with the same random seed the the same random number sequence is obtained. | 12:47 | ||
| <antononcube> @Volenet "[...] Dist::Uniform and Noise::Uniform [...]" -- these interesting designs, but they are not telling much. Think about a "standard" data scientist that has to read Raku code in order to understand certain analysis document (article or blog post). Names like "Dist::Uniform" and "Noise::Uniform" are too ambiguous. | 12:52 | ||
| <antononcube> This statement is pretty clear: random-variate(UniformDistribution.new(min => 10, max => 200), size => 1000) | 12:54 | ||
|
13:51
abraxxa joined
13:58
stanrifkin joined
|
|||
| Geth | docker: 72aa026f1f | (Anton Oks)++ (committed using GitHub Web editor) | .github/workflows/update_release.yml update_release.yml: Re-addint (organizational) token |
14:05 | |
|
14:23
tinita_ is now known as tinita
15:18
johnjay joined
16:37
thaewrap- left
17:25
Sgeo joined
17:37
stanrifkin left
19:06
undermine joined
19:33
scott2 is now known as fondue,
fondue left,
fondue joined
20:41
abraxxa left
|
|||
| Voldenet | of course, but I'm not sure if I'd design the api for "cumbersome to write, wordy and descriptive, but obvious" | 20:58 | |
| both use cases make sense | |||
|
22:09
sftp left
|
|||