IWriteDaCode

joined 1 year ago
[–] IWriteDaCode@programming.dev 5 points 1 year ago (1 children)

Ah I think it's Twitter's new thing where you can't see replies of your not logged in.

[–] IWriteDaCode@programming.dev 12 points 1 year ago* (last edited 1 year ago) (3 children)

The tweet you link didn't indicate that. It said that an engine failure likely caused the overrun, running for 127 seconds instead of the planned 84. Why would something have a 2^7 int size check?

Edit: Quoted

The head of Roscosmos Yuri Borisov said that the main cause of the #Luna25 crash was an engine failure. Instead of the planned 84 seconds, he worked 127 seconds.

Am I missing something?

Unfortunately, trends are trends. I wonder if we can get people to use the open source live captions application that futo sponsored recently. At very least we can get reasonably high quality captions, as well as a full transcript automatically generated with each video. Live captions is done with a locally running AI so you don't have to reach out to any third parties or share data to use it.

JIRA is just an issue tracker.

Nope, I mean at it's core, yes it is, but it's used for sooooooo much more than that. It enables management from a far fistance, and that disencentivices managers from doing their job.

I get the premise, that tools just exist and it's us that put our own biases in them. But that looses a lot of nuance when a tool is specifically built for a purpose, such as oversight, tracking, and data collection. These design decisions take an "issue tracker" far away from what Trello, or a whiteboard with stickies on it for that matter, does.

It is a grave mistake to think that it's just an issue tracker, and that's all it can be. I've been in this industry long enough not to fall for that con. And it is a con, when someone manipulates you using a tool that is designed to make manipulation easier (I'M not telling you to point every story even if it doesn't make sense. But you know... Jira wants it, it's just... Outa my hands...).

Nah, Jira is for managers, not developers, and is far more than a simple issue tracker.

[–] IWriteDaCode@programming.dev 4 points 1 year ago (1 children)

I think I'll take my IDE that has the choice in which AI assistant I get... Why would anyone lock themselves in like that?

[–] IWriteDaCode@programming.dev 72 points 1 year ago (3 children)

Nah, I also hate Jira. It's slow, bloated, complicated, and has 1000x features I, as a developer, don't need.

But then again, I also hate the manager that makes me use it in ways that frustrate me.

But then again, the reason my manager loves Jira and wants me to use it that way, is that they can run a bunch of automated reports like "We did X work this week, consuming Y hours (Or points or whatever) and we predict that we will be done in Z timeframe".

Buuuuuut, that's all bullshit. Garbage data in, garbage reports out. Jira gives managers the CONFIDENCE that they know what's going on, instead of just talking to developers, having conversations, etc. As it turns out, programming is hard, and doesn't have clear A->B->C predictability. So those tasks that are left? non-exhaustive. Those hours we did? Didn't take into account the thousand little things that didn't go into the backlog (And would take longer to add it than to just do the work and ignore the extra time spent on the task). That burndown chart? Completely useless.

Jira is used to skirt around the complexity of software development. It enables bad management to exist much easier, because it allows said managers to not engage with the team or product in any meaningful way, then to push up the chain "progress reports" that are meaningless, then, when deadlines are passed, managers get to blame it on the developers for not tracking enough work in Jira.

Jira enables bad management.

On the other hand, bad managers abuse every tool they are given, and bad managers existed before Jira, just instead of automated reports, they had email reports and hand tracked hours. So whatever, the tool was built to service a broken industry anyway.

[–] IWriteDaCode@programming.dev 1 points 1 year ago (1 children)

Well, there is webassembly, this will enable static programming languages like c, c++, and rust. Rust is high enough level that it's a pretty good sell for web apps. Even so, once webassembly calms down, we can build alternative interpreters that run on it, and run Python/Go/Java/etc. in the web. This will not be quite as efficent, but as c is about 50% slower in webassembly than on native hardware, I think a Go/Java would run well, Python might be a bit behind, but Lua is simple enough it might work.

TLDR: You don't have to kill JS to get those language functionalities, just wait for webassembly and all the cool stuff that comes about from that.

Caveat: Though I'm a developer, I'm not a webassembly developer. I've heard of these things as theoretical possibilities, but don't know the specific limitations. Sounds promising but who knows how long it will take to get there.

Damn, 26 people, let's assume they get paid reasonably well, though they probably aren't all developers. I'm going to assume 100K on average, just spit balling.

That's two and a half million dollars per year to build a terminal (very conservative estimate). And, like, does it reeeeealy do more then other terminals? Especially when you include different shells with plugins? AI, it's so hot right now, but it is better than zsh or fish autocomplete? I built the simplest AI shell script to ask GPT-4 questions, easy, many FOSS options already out there, is that not good enough for people?

Yeah, I'm just having trouble figuring out how this isn't a waste of time and will implode when seed funding dries up.

[–] IWriteDaCode@programming.dev 2 points 1 year ago (1 children)

I will never be able to get behind a subscription-based terminal, with so much competition in the FOSS space for terminals, there's just no reason to.

[–] IWriteDaCode@programming.dev 1 points 1 year ago (1 children)

In my personal experience I have gotten much more praise from Big Corp company when the work I did was more visible to management. The best work I've ever done was completely ignored because it was more technical and difficult for management to understand what the work was about.

And it wasn't just about explaining the work, it just wasn't that interesting to people who aren't technical.

It was after getting an award for doing some extremely easy work, that I realized that it's much more important that you communicate what you do, than actually doing useful work. And this sucks real bad, because if you do good work, it means you have to spend a bunch of time outside of that work just explaining it and acting like it's a big deal, and you can easily beat the system by overrepresenting easy work, because you have a lot more time to explain what you did.

Just my experience with my Big Corp, it may not be quite like that everywhere.