Taras Novak 🇺🇦

@dataSamurai@vis.social
193 Followers
103 Following
384 Posts
Logic - Wu Tang Forever ft. Wu Tang Clan (Official Audio)

YouTube
might be somewhat Dexter for some of our followers here, but pretty much sums up our life journey 😔
WESTSIDE BOOGIE - Rainy Days ft. Eminem

YouTube
Eminem - No Love (Explicit Version) ft. Lil Wayne

YouTube
might be moving to Alaska next year, so don't really care about this handle, the works done for #VSCode IDE in the past 5 years, or the other online handles we created and burned through 🤗

📝 New blog post 📝

'Positron IDE - A new IDE for data science'

🧞 In May I wrote about my favorite IDE's. As a result of that post, I was asked to join Posits private beta-testers for their new IDE, Positron. Now Positron is out of private beta, so let's see if it can replace my current favorite IDE's. ---

👀 Read more at https://drmowinckels.io/blog/2024/positron

#rstats #data #IDE #positron #vscode

Creating post summary with AI from Hugging Face - Dr. Mowinckel’s

When you have lots of old blog posts, writing good summaries for SEO can be very tedious. Using AI, we can get summaries for our posts easily. Using httr2 we can access models through the Hugging Face API.

Eminem - Tobey feat. Big Sean & BabyTron (Official Music Video)

YouTube
Snoop Dogg - Vato (Official Music Video) ft. B-Real

YouTube

📣 Published new free and public #DuckDB #SQLTools Preview v1.6.0 for VS Code IDE with DuckDB v0.10.2 support.

📥 https://marketplace.visualstudio.com/items?itemName=RandomFractalsInc.duckdb-sql-tools

See our #ProDataTools repo for advanced #DuckDBPro schema, commands, and new features updates:

📰 https://github.com/RandomFractals/pro-data-tools/blob/main/duckdb-tools.md#pro-data-tools

#VSCode #DataTools 🛠️ 🧙‍♂️ ...

DuckDB Sql Tools - Visual Studio Marketplace

Extension for Visual Studio Code - DuckDB Sql Tools extension adds DuckDB connections, database schema, and SQL query user interfaces to VSCode IDE.

for all the rock duckdb superstars in dev tools out there. lol!

https://youtu.be/W4VpE-0zitU?si=HMSrx_ZkJPm0li4A

#devTunes 🎶 ...

Cypress Hill - (Rock) Superstar (Official Video)

YouTube
×

📝 New blog post 📝

'Positron IDE - A new IDE for data science'

🧞 In May I wrote about my favorite IDE's. As a result of that post, I was asked to join Posits private beta-testers for their new IDE, Positron. Now Positron is out of private beta, so let's see if it can replace my current favorite IDE's. ---

👀 Read more at https://drmowinckels.io/blog/2024/positron

#rstats #data #IDE #positron #vscode

@Drmowinckels You don't mention the license...
@geospacedman I don't mention the licences of any of the IDEs I've posted about before, so I didn't really feel the need
@Drmowinckels It should be the first thing to consider when choosing scientific tools.

@geospacedman tbh, the elastic license seems quite permissive, and I can't say it has any effect.

Given that much of science is mainly done with proprietary, closed source software, elastic licence seems very favourable.

@Drmowinckels That's alchemy, not science, that's done with closed source software. Although the actual science (the computation) is done with FOSS (R itself), if you tie your project to a non-FOSS IDE you'll hit reproducibility problems in the future. Or you already have, with the "here" package.
@geospacedman @Drmowinckels Can we compromise a bit :P at least the time I learned emacs (long way to go!)
@defuneste @Drmowinckels You can, but I think its important to be aware yourself and to make people aware of the license choices they make, and not just click "Accept". I think we have great truly FOSS-compliant tools and we should use them and encourage the use of them. All else is open-washing.
@geospacedman @defuneste encourage yes, let's do that. Fair. You made me look into the license, and I still find it quite permissive. But you then follow up basically saying that everyone who doesn't use exclusively open license software are not even worthy of being called scientists. That is not encouraging, that is rather pushing people away.
@Drmowinckels @defuneste I think the alchemical analogy is a valid one. There's a viewpoint that alchemists were mystic grifters and frauds, whereas they were mostly using poor scientific methods and fooling themselves.

@geospacedman @defuneste I don't agree. I refuse to say that my colleagues using SPSS are doing poor science just because it's proprietary. I see them being rigourous, and validating results in multiple ways they are able.

I don't think it's a productive stance.

@Drmowinckels @defuneste Its not because its proprietary, its because its not reproducible. SPSS inc folds and stops issuing licenses, that science can't be reproduced. Then what value does it have?

@geospacedman @defuneste I agree it's not ideal, and will happily advocate (and do) for more open practices. But that being said, the language you use to talk about this matters. And I don't think the way you argue helps, I think it rather deters.

Why listen to someone who blatantly disrespects how you do things? FOSS is not always the most accessible to people who dont code daily. That doesn't make their work less.

@Drmowinckels @defuneste I think we need a diversity of approaches to open science. Its not like I go knocking on random lab doors screaming at anyone with SAS open on their desktop.

@geospacedman @defuneste I should hope not 🤣

But the way you approached here was basically telling me I am a poor scientist because my entire tool chain is not FOSS. You won't win me over with an approach like that, and I dare say you won't win anyone over with an argument like that.

@geospacedman @defuneste and circling back to the OP, positron as an IDE is not introducing anything that I can see that stops reproducibility.

The thing with the here package is just as easily done through terminal and is default R behaviour when called from shell. In a sense, Positron is a step better than RStudio for reproducibility because of this. It's a literal interface, not something that alters behaviour.

@Drmowinckels @defuneste I admit its a bit extreme, its really only food for thought to encourage self-reflection on one's own journey in the scientific process. I've never rejected a paper for using STATA, I'm not *that* much of a "Reviewer 2".

@geospacedman @defuneste good to hear, we all need less reviewer 2 in our lives.

I'm glad you had me look at the lisence, I do agree it's something to pay attention too. But I daresay I have a more lenient interpretation of what is good enough than you.

@geospacedman Just back off. You've said what you've come to say and your message was received, albeit with disagreement. At this point, you are just harassing Dr. Mownickel in a semi-public forum.
@zkamvar Please feel free to mute/block/ignore.
@geospacedman Look, I see you harassing a scientist whose work actually brings people into FOSS and R. Me muting/ignoring you is not going to stop your behavior toward her. But maybe if I suggest that you go touch grass, that might work. Online arguments are not worth your energy.
@zkamvar Dr M is also free to mute/block/ignore, or engage. Are you white knighting? Also, I had a new rose bloom in the garden this morning and it smells lovely. Not sure what my horticultural situation has to do with anything but there you go.
@geospacedman @zkamvar If you say something inflammatory and kinda mean directed toward an individual, and then when they push back and give you insight as to how your communication was perceived and you just say "just block me if you don't like it!"—that's the definition of a troll. I hope for your sake that you have better interactions with people IRL because that approach has got to make life difficult for you.
@geospacedman @Drmowinckels @defuneste I think your premise is flawed. Good peer reviews ed science does not lose its value because it can’t be reproduced 20 years later because a software vendor went out of business. Also what happens when the FOSS maintainer stops maintaining the software and no one picks it up. Is the scientist expected to build the tool they want to use too? All software can EOL, even FOSS. Does that make the science less valid?
@danwwilson Not being able to run someone's methods for your own research seems like a clear "loss of value" to me. Take two papers with identical methods developed, one in a defunct proprietary system, the other in a free open environment. Which has more value? Plus FOSS s/w never really EOLs, they get dropped but they'll always be legal to run, even if you need a VM or container. And you can take over the dev or pay someone to do it. Commercial Open Source is a thing.

@Drmowinckels @geospacedman

team compromise here, I think it is a road to pursue but you can be at multiple stage of it (I am terrible at translating analogy in english).

(btw license are freaking important in coporate env.)

@geospacedman @Drmowinckels @defuneste this is a bizzare stance. You realize this probably applies to *most* science in the last couple off decades, right? We should throw it all in the trash because MATLAB or SAS were used??
@LeafyEricScott The problem is we can't tell what is deserving of trash and what isn't. And obviously there's a difference between the philosophy of scientific practice and the practice of it... But going forward, less of the proprietary and more of the reproducible is, I hope, a point of agreement?
@geospacedman Sounds like it's almost like you can't tell because "didn't use FOSS" is maybe relatively meaningless when it comes to determining the quality of science. Alchemy is pseudoscience, not science done in a less reproducible way.
@geospacedman I'll also say that it's weird that you'd rather people block you than say "sorry I implied that you are a garbage scientist if you're not using 100% FOSS, that was a bad approach to communication". Like, how do you live like that?
@geospacedman if you are going to call all science done with proprietary software for alchemy, I'm afraid we are never going to see eye to eye. I don't think a stance like that will ever sway people your way.
@geospacedman the problem with the here package is actually from R it self and how things are executed. I'll have the same issue if I try to run a here script from in subdir from shell. This is not new to me, and somewhat expected behaviour,.that RStudio solves because they were made for eachother.
@geospacedman @Drmowinckels Forget the license, I still cannot unsee the stick figure

@Drmowinckels I can't really use Positron for work because it doesn't have the Remote SSH feature of VS Code (yet).

For home projects, I found that I didn't need a variable/data explorer as much as I thought - since I tend to write code in VS Code then import it into jupyter to visualize / use.

I really like the list of extensions. I installed a few I hadn't tried!

@marcoshuerta that makes sense! I'm looking forward to SSH features too, but I'm not that encumbered by it lacking.

I've never enjoyed using jupyter, but I can see how the.festures added by positron is not as necessary when you have that type of workflow.

@Drmowinckels Using separate profiles for teaching vs. one's own work is a gooooooood idea. 👀
@juliasilge I asked for something similar in RStudio some years ago but it never got traction. I was soooo pleased to see it in Positron. Teaching profile was my absolute first idea 😂