Hacker News Reader

Comment by projektfu | Hacker News Reader

I hate to say it but free software has eaten everywhere many programmers want to work. If you want to work in operating systems, frameworks, databases, editors, tools, compilers, etc, you’d better be willing to do it on your own time. If you’re like top 10, you’ll get picked to work for the big leagues and make money doing it. Meanwhile the paying gigs are the things that either are services or can’t be open sourced.

I guess I’m just sad that all the tools conpanies are gone or shells of their former selves. Borland, Symantec, etc. New ones pop up but there’s no money in support.

Companies are making huge money on their services that are 90% open source and 10% special sauce that they don’t contribute back. We’ve seen how they’ll spend $300,000+ per programmer but won’t pay $10,000/year for a tool or framework. It’s like the economics are messed up now.

View on HN

Free software has definitely created some weird economics. Like I can't imagine paying money for a programming language on the other hand I don't blink about paying for Jira.

Maybe not for a programming language per se but Developers often do pay for IDEs and other niceties around developing software. Swift is another interesting example of a language that you have to pay for developing in (not directly but in order to participate in Apple's developer program).

Swift on Linux costs no money haha

Infrastructure software pays exceptionally well and companies will pay impressively large amounts of money for it. You just can’t deliver a commodity product that has a vague equivalent in open source. I always recommend using open source first, if it will work, and I literally sell closed source database technology. Open source doesn’t come close to competing with anything I do, so it isn’t a threat, and the market for what open source can’t do is astronomical.

There is a booming market for state-of-the-art infrastructure software but that market is no longer retail and it doesn’t publish much, because that business model no longer makes sense. An uncomfortable fact, but one that many enterprise customers acknowledge, is that open source database technology (the part of the business I know) is quite poor. If you want to get paid in databases, produce software that offers an order of magnitude or two higher efficiency, scalability, throughput, etc on the same hardware than anything in open source. Companies will happily pay for that because it saves them millions of dollars a year in OpEx alone.

The economics works just fine, better than fine actually. You just need to deliver a product that materially outperforms the not-that-high bar of open source. This market is huge, it just isn’t as visible anymore because they don’t do marketing on billboards in Silicon Valley.

The real issue is that there aren’t enough people that can design infrastructure software that is materially better than open source. There is a large skill gap.

> I guess I’m just sad that all the tools conpanies are gone or shells of their former selves.

And the tools they made are largely gone with them. The whole point of free software (well, one of the benefits anyway) is that it keeps being useful even after the original programmer stops making it or no one can make money from it anymore.

(And just for fun, since you're mostly talking about money and money wasn't mentioned in the article: https://www.gnu.org/philosophy/selling.html )

The software stays but without support it dies or remains in a stagnant place until it’s unusable by modern equipment.

That's not what happened. Pre-internet, we paid for software. Post internet, we paid for services.

We run our own Jira instance, but yet we pay for it.

I think we need to find a new paradigm for funding infrastructure work. You guys might not like it, but may be you could get something like grants from NSF or the like.

In the computational world, NSF, DOE, and DOD often fund the development of computational codes that are then used by researchers across the US and even the world. It isn't the most stable way to fund development, but it's better than nothing that is the status quo. For an example of this, see this[0]. The ecp is kind of an exception, but smaller grants fund all sorts of smaller codes.

[0] https://www.exascaleproject.org/

I don't see this at all. You can easily get a high paying job contributing to LLVM, Clang, Swift, Chromium, WebKit, GraphQL, React, K8, and 1000s of other open source projects. And, many companies do contribute one way or another.

> If you want to work in operating systems, frameworks, databases, editors, tools, compilers, etc, you’d better be willing to do it on your own time.

This true. I was very glad that this contest show up (https://repl.it/jam) just for see somebody try to give some incentives to early and experimental try. This help me in push for make a MVP for my own lang (https://tablam.org) that I dream could be something like FoxPro/Access/Excel.

I wish exist a way to access to micro-funding in this area. Even if is still for "hobby" I think it could make a dent because from my experience most believe this kind of tools are "too hard" for the average developer to work on, until somehow you find some time and just try.

This way the talent pool can be build...

Before, the best paid jobs were in developing operating systems or databases or compilers for companies that sold them. Now the best paying jobs are optimizing ads whether in regards to efficiently storing and processing people’s private data or making it more likely people will click an ad.

Reply on news.ycombinator.com