Live data from Hacker News

“Open source” is not broken

nadh.in

181–190 of 259 posts

Re: “Open source” is not broken

#181
post #72

>If FOSS was broken, the internet as we know it today wouldn’t exist; the countless marvels of technology that we take for granted and techno-economies that thrive on them wouldn’t exist; I guess I just vehemently disagree. Nearly all of the early open-source software that made the internet possible was produced in universities. The only reason it was sustainable was because it was professors being paid by the univer…

Having worked with people in industry who understand the point and value of giving back this is a little naive I would argue.

A fraction of some talented persons time from say HP is probably worth 100x first year developers who aren't paid to understand the tools the company is using.

To turn your argument on its head how much would every company have to invest to build a modern website from complete scratch in isolation? Then think why do that when you can effectively spread the cost?

Both approaches have ups and downs but I'm not sure the "someone always picks up the cost" isn't anything other than a statement of realism. It is a good reason to explain why nobody just works on a project in their basement for free and do nothing else, but doesn't role out being able to do this if responsible companies pick up a fraction of the tab they should be paying via donations.

As others have said a huge amount of the value comes from support, community and the contributions from many people, be they working on the same tools for a product they sell, to make a product or service they plan to sell or to scratch that itch on that project in their spare time they're playing with.

Re: “Open source” is not broken

#182

Earlier quoted context omitted.

For some types of software, we really do not want students doing it, for free or otherwise. There are whole classes of software, like database engines, that are non-obvious and require many years of real-world domain experience before it is plausible that someone will design a competent, scalable architecture and implementation. If open source is going to run critical infrastructure, we don't want naive and inefficie…

> For some types of software, we really do not want students doing it, for free or otherwise. There are whole classes of software, like database engines , that are non-obvious and require many years of real-world domain experience before it is plausible that someone will design a competent, scalable architecture and implementation. Of course as with all things the situation is more nuanced than this. Since you mentio…

Database development has changed quite a lot since then. When I first started working on databases in the 1990s, two things were true that made it much easier for relatively inexperienced developers (like myself at the time) to produce a reasonably good result. First, the state-of-the-art implementations at the time were incredibly well-documented in an accessible way and the academic literature also reflected those designs. Second, the implementations were relatively simple and straightforward; you did not need esoteric systems knowledge and design theory to write code that was competitive with other implementations. The most complicated thing you had to worry about was lock structures and concurrency control. Not only were there relatively simple examples to copy and study, the gap between those examples and the state-of-the-art was pretty small.

Neither of these is true today. The design of state-of-the-art implementations are often poorly documented; the computer science has evolved radically since the 1990s with significant gaps in the literature; competitive implementations require real expertise in silicon architecture and Linux kernel internals, you can't just write something obvious in C and expect a good result. And that's without even getting into difficult topics like distributed execution, parallel orchestration, scheduler design, etc that we didn't have to worry about back then.

I'm not sure I'd be able to bootstrap the necessary expertise to build database engines today like I did back then.

Re: “Open source” is not broken

#185
Blaming "society" is a nothingburger. It doesn't point towards any solution. Ditto "capitalism" (although the OP doesn't go there) --- people pursue desires and respond to incentives; you can call those effects "capitalism" if you like but you'll struggle with them under any kind of economic arrangements.

My self-diagnosis as an unpaid open source maintainer (rr) is that we like to share, the marginal cost of sharing with potentially everyone is close to zero (at least early in a project), so throwing code on Github with a liberal license feels good. It also benefits the project to some extent because some improvements may come back. But then we see rr creating huge value for people, some of whom are very well paid for their work, and none of that is coming back to us, and that seems unfair even though we did technically agree to it --- sure, you don't have to give back, but it would be the nice thing to do. But in software it's very easy to extract a ton of value from dependencies without really noticing, and very hard to give back systematically. So it's the same old story --- perfectly good human instincts aren't a 100% fit for our modern environment.

What if we made it easier to identify the value we're all extracting and contribute back systematically? If we did, maybe we could build social norms around that. E.g. imagine we had tools that monitor your software development workflow, identify the tools and libraries you use, and quantify your usage via some heuristics. Then imagine you integrate something like Github Sponsors so you can allocate $X to support all your dependencies and make that happen at the press of a button. Then imagine we advocate for professional software developers to allocate 1% of their income that way, and agitate for Big Tech companies to make that a policy.

Re: “Open source” is not broken

#186
I learned from the actix-web debacle that I don't want to ever do libre or open source software. If people want to complain to me about my code and call me a lousy programmer, then they have to pay me for the right.

A sad day for Rust

https://steveklabnik.com/writing/a-sad-day-for-rust

Re: “Open source” is not broken

#187
post #72

>If FOSS was broken, the internet as we know it today wouldn’t exist; the countless marvels of technology that we take for granted and techno-economies that thrive on them wouldn’t exist; I guess I just vehemently disagree. Nearly all of the early open-source software that made the internet possible was produced in universities. The only reason it was sustainable was because it was professors being paid by the univer…

I dont understand how exactly is open source supposed to be broken. It is accepted and respected these days. There is tons of it too.

If does not produces flawless miracles, but commercial software is not flawless either. The log4j bug has impact it has literally because open source was successful.

Re: “Open source” is not broken

#188

Earlier quoted context omitted.

There are a lot of people who are very good at coding who don't care for high salaries. Despite mainstream pop being the most profitable, many musicians pursue niche genres.

This still doesn't make UBI relevant to this discussion. It's already fairly trivial for a talented SWE to make UBI-level income working 10 hours a week by picking up software contracts here and there. Eng already have access to the levels of income that UBI would provide, with plenty of time left over to dedicate to open source, and yet this path is relatively untrodden. Plenty of engineers (myself included) already…

> already fairly trivial for a talented SWE

This level of subjective qualification makes everything that comes after it essentially meaningless.

Re: “Open source” is not broken

#189
post #85
post #24

Earlier quoted context omitted.

Except they're mostly not. While it's not hard to find exceptions most of the maintainers and other coder associated with major open source projects are being paid by companies to do so.

This really depends on the project. Many people just do it in their spare time and have a normal software engineer or similar role working on proprietary software using open source. Some are still at university and studying. Some are contracting and probably also get some new customers thorough their open source work. I do not have real data, but I assume that more than 50% of the effort invested in open source is no…

Majority of people responding to FOSS studies is paid by companies. It is biased toward Linux related open source projects a bit.

The thing is, while there are maintenaners of important projects working for free, the idea that most of major open source development is for free is mostly mythology. Large projects are backed by companies.

Re: “Open source” is not broken

#190
post #32

Open source is most certainly broken, and not just due to the various financial, freedom and security issues these two articles focus on. My biggest peeve: documentation is often minimal (e.g. API docs only) or filled with useless toy examples that are effectively just rephrasing of API docs. The entire underpinning of free and open source software is silly: software in this context isn't an academic pursuit producin…

Wait till you read documentation commercial projects have. Our one is super quick to read, as we don't document at all.
Post reply on HN