Live data from Hacker News

Open Source does not win by being cheaper

github.com

231–240 of 287 posts

Re: Open Source does not win by being cheaper

#231

Earlier quoted context omitted.

You are still getting sued.

It's a theoretical benefit, not a real one, and the scenario is so rare that most people don't think of it critically. It's like buying insurance against "damage by asteroids". E.g.: "We can just sue Microsoft" is a common one. a) Their SLA literally states that they will compensate you only for the cost of the individual service they provided. The $1/month Azure DNS Zone outage caused your $100M/month business to va…

Most customers won't sue you. They'll terminate their business with you and go elsewhere.

You have to decide, as a business, if trustworthy vendors are available and select the right one. You can also decide that it's important enough to your value proposition that you want to "own" it.

The real reason companies hire vendors is that you don't need some mythical "FTE" that can do all the things. For the "price of a single FTE", a vendor can provide you access to a range of FTEs that are knowledgeable and capable in different domains, and you can flex between them. Thinking about RedHat, as an example, if you pay RedHat the equivalent of a single FTE, let's say $300k/yr, is that just one FTE working on maintaining Apache? A kernel contributor? GCC?

This is about fractionalization and labor specialization, not some direct 1:1 tradeoff between hiring some mythical person with commit bits in hundreds or thousands of OSS trees.

Re: Open Source does not win by being cheaper

#232
Posthog isn't truly open source as the author claims.

It's "too complex" to run natively and is only recommended for "hobby" projects. They basically say, yeah our source code is there but if you want to run it yourself, forget about it.

I was thinking of some clients I have that would benefit from this in their secure environments but was quickly chastened by their docs.

Fair enough, they have to make money but seems like the ole bait-n-switch method of customer acquisition.

Re: Open Source does not win by being cheaper

#234
post #150

Earlier quoted context omitted.

> That tension seems to inevitably lead to obvious features missing or functionality/knowledge that's needed to operate at any scale coistered away as closed source so the sponsoring entity can make some money. This is not inevitable at all. You're right that there's a tension there, but there are companies that manage to do this right, and please both OSS and commercial users. The fact many companies don't is not an…

One of the main reasons to donate a project to a foundation, such as the Cloud Native Computing Foundation (CNCF), is to protect the community against this type of rugpull. A requirement for them to even consider adopting a project is that it isn’t under majority control of one single commercial entity. Do people honestly believe none of these projects compete with commercial solutions, or are they mostly oblivious t…

A better measurement would be how many of those projects are still relevant in 10 years from now.

Re: Open Source does not win by being cheaper

#235

Posthog isn't truly open source as the author claims. It's "too complex" to run natively and is only recommended for "hobby" projects. They basically say, yeah our source code is there but if you want to run it yourself, forget about it. I was thinking of some clients I have that would benefit from this in their secure environments but was quickly chastened by their docs. Fair enough, they have to make money but seem…

Posthog is a bit of a farce sadly.

Don't hate the player, mind.

Re: Open Source does not win by being cheaper

#236

The only reason to keep something closed source is if you're trying to hide questionable behaviors (e.g. NVIDIA drivers, M$ Windows) or your code's embarassing (most enterprise software).

Or you know, if you're trying to make a profit on sales of the software itself. All the working open source business models work on selling service, not the software. Which is fine, but not everyone wants that business model.

Re: Open Source does not win by being cheaper

#237
post #34
post #3

I've always been annoyed that companies will pay 6 figures for a support contract, but won't consider hiring a full time engineer for the same or less pay to support an open source product that does the same thing. The support contracts are often disappointing too, like you are paying $200,000+ a year for the stupid thing and they won't even fix your problems? "That product is not our current area of focus". So frust…

- Vendor contracts are hard - So we avoid doing them - So we never get good at them - So every contract is a painful lesson in trying Open source wins because I don't have to involve the entire company to use it. I can literally just raise my voice to see if anyone in the group cares and then use it. And even when it's free for small companies, not approving a license is holding up something that we think is already…

Not in the projects I am involved, there are clear guidelines, CI/CD pipelines are connected to internal repos only, and there is always a legal and security assement required before anything gets added to those repos.

Re: Open Source does not win by being cheaper

#238

Earlier quoted context omitted.

Profit has many meanings. The IRS considers my wages to not be profit to my company, but they are profit to me. I could setup a non-profit that explicitly doesn’t aim to profit, but it does so by paying my very comfortably. Profit‽ This disconnect is what causes many businesses that are successful (they earn enough for expenses and rainy days) to explode and die when sold, because they don’t earn enough for all that…

Your wages are not profit by any definition I've seen. You arrived at an employment contract with your employer. The contract is an exchange of value: your labor for some of their money. If you are being paid more than the price your employer could otherwise obtain your labor for, then I suppose you could call some of that profit. But that is rather unlikely. Your employer is buying your labor, and is almost certainl…

If the owners overlap with the employees, then what part of the revenue is divided between salary, benefits and profit is often a tax-optimization question, which I think is the point in the GP. A single-person company paying themselves 200k in yearly wages or 100k in wages and 100k in dividends from 200k in revenue can look like a “non-profit” or a high-margin company with the same effect for the owner.

> I've been running an open source, libre project for closing in on 24 years now, and generating revenue from it for about 17 years. There's never been any "profit", but there has been revenue. I regard profit as what's left of the revenue after you pay the people who work on the project and any expenses, which is the way most corporations and accountants would view it.

Re: Open Source does not win by being cheaper

#239

The only reason to keep something closed source is if you're trying to hide questionable behaviors (e.g. NVIDIA drivers, M$ Windows) or your code's embarassing (most enterprise software).

Or you know, if you're trying to make a profit on sales of the software itself. All the working open source business models work on selling service, not the software. Which is fine, but not everyone wants that business model.

Why would a client pay for a closed source wrapper over battle-tested open source libs?

Excl. the rare case where the problem space requires a novel solution.

Re: Open Source does not win by being cheaper

#240

"Open Source" doesn't win, period, because it doesn't compete with anything. Open Source is not a business model or a product. It's an engineering widget. You can use it to build a product, or you can use it to make a jig to shore up an uneven table leg. Nobody cares about the license or whether they can read your code. They care if you build a product that solves their problem. You know why companies use Open Source…

I guess I'm lucky that the open source things I deal with are far and away better than the proprietary options. I'd choose Clang or GCC any day of the week yet another 7 figure proprietary toolchain with licensing terms with obvious bugs, for one recent example.

Meanwhile many HPC Linux clusters end up using ICC or xl compilers, for the special sauce optimizing to all those NUMA and MPI connected CPUs.
Post reply on HN