@jni@fosstodon.org
@jni@fosstodon.org avatar

jni

@jni@fosstodon.org

Core dev and co-founder napari.org
Core dev scikit-image.org
Co-author elegant-scipy.org
Research Fellow https://www.monash.edu/discovery-institute

Blog https://ilovesymposia.com

This profile is from a federated server and may be incomplete. Browse more on the original instance.

jonny, to random
@jonny@neuromatch.social avatar

I think in the future if I am ever writing a code paper I am just going to take the list of contributors and copy paste that into the authors list with links to a git blame (with consent). If we're going to have a credit assignment system as broken as authorship, we can at least err on the "include everyone" side of the brokenness - I want the person who submits a PR to fix a typo in the docs to get credit for helping. People being incentivized to make lots of little contributions is good, actually.

It should be the same way with regular papers too - put your lab techs and undergrads on the paper! Put on the grad student/postdoc who isnt explicitly assigned to this project but ends up helping out anyway. Its literally free! Authorship inflation is a made up problem thats not even a problem!

jni,
@jni@fosstodon.org avatar

@jonny @herrsaalfeld There is some minimum marginal utility that needs to be exceeded or PRs become actively harmful (through sucking up maintainer time). See e.g. the issue with Hacktoberfest spam from a couple of years ago:

https://blog.domenic.me/hacktoberfest/

https://dev.to/devteam/an-update-on-hacktoberfest-37a

So, although I agree that we should err on the side of inclusivity and generosity, I also agree with Stephan that there should be some human judgement involved to avoid creating perverse incentives.

glyph, to random
@glyph@mastodon.social avatar

Is it possible to configure visual studio code to run a custom test runner in Python? I want ⌘R to run the unit tests for the current file, but that often involves running Twisted's trial, and vscode's built-in test runner doesn't seem to even work for pytest half the time

jni,
@jni@fosstodon.org avatar

@glyph @greg I’m a weirdo with you @glyph. I 110% agree about the magic action at a distance and the name matching. But what alternative do you use?

jni, to random
@jni@fosstodon.org avatar

Every now and then a kind soul upvotes an old answer of mine on SO, and it turns out to be a fun read years later. 😃

https://stackoverflow.com/q/62135639/224254

"For some reason, my mcp object didn't have an offsets attribute […] but searching the source code shows me that offsets is defined with the skimage.graph._mcp.make_offsets function. So I did a bad thing and imported from that private module, so I could claim what was rightfully mine — the offsets list […]"

Can we please start writing papers like this?

jni,
@jni@fosstodon.org avatar

@albertcardona three things:

  1. I read way more papers than I write, so I want others to do it, not just me.

  2. Most papers are collaborative and it’s hard to convince co-authors to adopt informality on purpose. And,

  3. Worst of all, even if I am trying, I revert to academese all too easily in the paper-writing context. The maze is in the mouse. It’s like a horrible virus. When I try to break free I get a mish-mash of styles. 😕

But I want to!!!

mcnees, to random
@mcnees@mastodon.social avatar

Every story the Universe tells is a history.

Danish astronomer Ole Rømer was born in 1644. His observations of Jovian moons confirmed light propagates at finite speed. This profound conceptual shift opened the door for work by Huygens, Fresnel, and Maxwell.

Portrait: Jacob Conring / Wikimedia Commons

jni,
@jni@fosstodon.org avatar

@bertwells @mcnees the historical approach is my favourite way to learn. Science is a detective story, and humans love stories, not facts. Being told just the ending is boring! Speed-running the story is exciting! It’s also a better way to really understand the relationship between experiment and theory.

A good example (sans maths) is Big Bang, by Simon Singh. That book was almost spiritual for me.

jni, to random
@jni@fosstodon.org avatar

Reminder to turn on full text search for your posts in Preferences > Public Profile > Privacy and Reach > (Search section) > Include public posts in search results! This goes a long way towards making Mastodon interactions less ephemeral.

Screenshot of mastodon preferences UI showing "Public Profile" and "Privacy and reach" selected.

albertcardona, to android
@albertcardona@mathstodon.xyz avatar

Probing the internet hivemind, trying to help a colleague: their android phone country setting says UK, but their google account was originally created in the US, and hence the Google Play store is still set to US. There are no buttons – despite what google support online says – offered to change the country in the Google Play account settings. So UK-only apps cannot be installed. Updating the address in the google account didn't help. What else can be done?

#android #google

jni,
@jni@fosstodon.org avatar

@albertcardona I think they are doomed. I changed my Google account country a decade ago. But recently I gave in to Google’s protection racket and got a family YouTube premium account — only to find that I can’t share it with my family because they allegedly are in a different country.

Sharing with @perlman worked no problem though 😂😭

So there is a long-lost bit in these accounts that is impossible to edit as a user.

Only silver lining is my wife Gen has the same bit. 🤷‍♂️

kisharrington, to random
@kisharrington@mastodon.social avatar

this is a fun video for oat selection.

slime mold growing in a petri dish with oats. this video shows the slime mold growing with a more venous structure as opposed to being blobby.

jni,
@jni@fosstodon.org avatar

@kisharrington amazing! Please tell me this is going into a napari sample data plugin!

jonny, to random
@jonny@neuromatch.social avatar

Alright my dogs what is the best way to serialize multidimensional array data for cross-language use? Is there any kinda standard or am I naïve.

I want to be able to translate between arrays stored in hdf5, relational databases, and binary blobs in content hash graphs. Is that too much to ask?

jni,
@jni@fosstodon.org avatar

@jonny @migurski @brainwane npy is not cross language nor does it intend to be. @zarr is and does and there’s a large community behind it across many fields and there are working or almost-working implementations in Python, js, C, and Rust (that I know of). tldr use zarr!

Only downside afaic is there is not yet a one-file implementation in the standard — it’s always at a minimum one chunk and one metadata file.

jni, to random
@jni@fosstodon.org avatar

Whoo! CZI just announced round 6 of the Essential Open-source Software for Science (EOSS)! (Now jointly with the Kavli Foundation and the Wellcome Trust.) My favourite thing CZI does. If you write open source software used in science, check it out!

If you don't know, this is not a "propose a cool new software package" call — this gets at the heart of software sustainability and supports the boring stuff for established projects — maintenance, docs, onboarding, sprints…

https://chanzuckerberg.com/rfa/essential-open-source-software-for-science/

hywan, to random
@hywan@fosstodon.org avatar

The case for Nushell, https://www.jntrnr.com/case-for-nushell/.

Relevant article about shells, and how Nushell pushes the boundaries further. I highly recommend reading it.

jni,
@jni@fosstodon.org avatar

@isomeme @cstross @Keltounet @hywan thanks for an eloquent post. On the one hand I had found these things low-level infuriating over time without deeply understanding why. (And indeed feeling like I’m just a grumpy old man who doesn’t like change.) On the other, as someone who works with UIs in open source software, I hadn’t considered this side of the coin. I will aim to be more mindful in the future. 🙏

nedbat, to random
@nedbat@hachyderm.io avatar

Python proposal (inspired by Lua!):

args = {'a':1, 0:98, 1:99, 'b':2}
f(**args)

is the same as:

f(98, 99, a=1, b=2)

I've long wished to have one thing that could serve as both *args and **kwargs. Int keys are already invalid, so nothing would break.

jni,
@jni@fosstodon.org avatar

@nedbat @comex @christmastree I think it’s quite cute as a simple extension of existing syntax, but you definitely lose me at triple-star…

jni, to scipy2023
@jni@fosstodon.org avatar

Cool, my @SciPyConf @scipy2023 talk on @napari is up! 🥳 With great thanks to the organisers. 🙏❤️ Check it out here:

https://youtu.be/s362FTRzdg0

As a reminder, the repo and slides can be found here:

https://github.com/jni/napari-talk-scipy-2023

though you'll need to adapt it to use your own data to run much of it.

Finally, check out all the other amazing talks from the conf! Really looking forward to catching up with the tracks I missed (bc conflicts or frantically preparing for my own talk 😂)

https://www.youtube.com/@SciPy-Conf/videos

jni,
@jni@fosstodon.org avatar

@joshmoore @SciPyConf @scipy2023 @napari @zarr yes I am sure I will be way better prepared next time! Way better! 😂

albertcardona, to Blog
@albertcardona@mathstodon.xyz avatar

I wish there was an htaccess rule to force a prompt for a question upon first access to my website, like "what color is the white horse of Santiago" that required a written reply and a press on an "OK" button. If someone wants to read my entirely static, ad-free blog, surely they can spare 5 seconds to type it in.

Adding more and more IPs and bot names under Deny rules seems futile. At this point I'd be better off blocking the whole website with basic auth and posting the password in social media.

jni,
@jni@fosstodon.org avatar

@albertcardona wait why do you care about bots? Is it a bandwidth thing?

jni, to random
@jni@fosstodon.org avatar

Looking at data in @napari is such a great way to quickly debug imaging code. Here I am learning that reading a dicom series is not as simple as stacking together the output of pydicom.dcmread().pixel_array. 😅 red: the above approach; grayscale: the napari-nibabel plugin.

Screenshot of a napari viewer with two layers in 2D slice view: a red one and a grayscale one. The objects in the slices are obviously misaligned.

jni, to random
@jni@fosstodon.org avatar

The Wobbly Table Theorem might be the most useful I know in everyday life. Works every time! (I guess that's the thing about theorems! 😂)

https://arxiv.org/abs/math/0511490

tldr: if your table is wobbly, rotate it around its central axis. At some point, as if by magic, all four feet will be firmly planted on the ground.

Amazingly, this theorem does not seem to have a Wikipedia page.

jni, to random
@jni@fosstodon.org avatar

It’s 2023 and Apple still hasn’t surpassed the 2002 iMac G4 design (the desk lamp). (And the commercial is still one of their best too!) https://www.theverge.com/23808948/imac-twenty-five-visual-history-apple-design-models

jni,
@jni@fosstodon.org avatar

@albertcardona I can only agree with you, although the M1 Pros are at least a step back in the right direction after more than half a decade of garbage. Sadly I am a little addicted to a bunch of macOS functionality, but for Linux these days it would be Framework all the way for me — Apple is far down the list. Which is very sad.

Related, great article:

https://solar.lowtechmagazine.com/2020/12/how-and-why-i-stopped-buying-new-laptops/

jni,
@jni@fosstodon.org avatar

@albertcardona Do you have the 1.0 hinge or the newer ones? Apparently they did go through a round of significant improvement. And because it's Framework, you can just get the new hinges from https://frame.work/es/en/marketplace/parts and install them easily.

I thought most people liked the keyboard well enough... But maybe you should get a Nuphy 60 to carry around with it? 😊

https://www.reddit.com/r/framework/comments/117pqgm/nuphy_60_mechanical_keyboard_on_framework_fit/

kisharrington, to random
@kisharrington@mastodon.social avatar

Fun with Penrose tilings in . Tooted from with .

jni,
@jni@fosstodon.org avatar

@kisharrington what! That looks like way more shapes than I would expect @napari to handle…!

gvwilson, to random
@gvwilson@mastodon.social avatar

Just canceled my Zoom subscription: they might backtrack on their decision to start training AI models on users' data with no opt-out, but even if they do, it'll be a temporary victory.

jni,
@jni@fosstodon.org avatar

@gvwilson @melissawm not Jitsi?

jni, to random
@jni@fosstodon.org avatar

Wanna work on open source scientific software and get paid for it? This is a great opportunity to do that for the Astropy team! Remote anywhere on Earth ok! 😍

https://jobs.numfocus.org/job/2023-07-31_astropy

DrAnneCarpenter, to random

Let me know if you're interested in an invite to Bluesky. Seems to be a good SciTwitter community there and several folks are collecting invite codes.

jni,
@jni@fosstodon.org avatar

@DrAnneCarpenter @penguinoops @florianjug @albertcardona Are you looking for more sciencey goodness (in which case you guys are the missing ingredient, please cross-post here 🙏), or more general stuff to stay entertained (in which case you should just follow @lisamelton, who is the best algorithm)?

Anyway, re Bluesky, I'm 110% with Albert — given the people behind it and my disgust at what's happened to Twitter (and Reddit, and Google Reader, and…), I am not touching it with a 10ft pole.

jni,
@jni@fosstodon.org avatar

@DrAnneCarpenter @penguinoops @florianjug @albertcardona

Here's an interesting algorithmic timeline open-source app: https://chaos.social/@pkreissel/110099714893948163

and a freemium online service:

https://news.feedseer.com/about

And don't forget to use FollowGraph to find interesting accounts to follow:

https://followgraph.vercel.app/

I'll admit I'm surprised that there's not more algorithmic feed apps, but clearly they're coming, and anyone can create and run one! Anyone! That's pretty badass.

And please cross-post! 🙏

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • megavids
  • thenastyranch
  • magazineikmin
  • tester
  • Leos
  • InstantRegret
  • rosin
  • Youngstown
  • mdbf
  • slotface
  • everett
  • osvaldo12
  • kavyap
  • DreamBathrooms
  • JUstTest
  • ethstaker
  • khanakhh
  • anitta
  • GTA5RPClips
  • tacticalgear
  • Durango
  • cubers
  • ngwrru68w68
  • cisconetworking
  • provamag3
  • normalnudes
  • modclub
  • lostlight
  • All magazines
    •