LWN: Comments on "The Open Source Pledge: peer pressure to pay maintainers"

Gold, Silver, Bronze

sammythesnake — Fri, 01 Nov 2024 07:19:14 +0000

Though at least with M6J6 you have to go to some other junction before going back around the same bit of overused road :-P

German state pilot program to hire open-source maintainers

pabs — Fri, 18 Oct 2024 23:54:52 +0000

It is available to folks outside Germany too, but you have to choose the Freelance Contracting option instead of the Employment option.

German state pilot program to hire open-source maintainers

pabs — Fri, 18 Oct 2024 23:50:29 +0000

That isn't about individual contributors, but about project maintainers.

Incentive to reduce employed developers?

pabs — Fri, 18 Oct 2024 23:45:17 +0000

Isn't that the whole point of Open Source? To reduce development costs by sharing them with other companies.

Sentry is unfortunately not Free Software/Open Source

pabs — Fri, 18 Oct 2024 23:42:20 +0000

Free Software sponsorship is essentially cheap marketing to these companies, they get to spread their logos around, and as a bonus some of the libre rubs off on the non-free companies in the minds of visitors. Lots of people strongly dislike it when FOSS takes money from the likes of Google for example, although that is declining as they do it more though.

Individual contributors?

pabs — Fri, 18 Oct 2024 23:38:21 +0000

I wonder if any of these "lets fund FOSS" projects are thinking about individual contributors instead of or in addition to maintainers.

Wrong solution to the wrong problem

smurf — Mon, 14 Oct 2024 18:19:37 +0000

*whose* itch, if you please.

Wrong solution to the wrong problem

raven667 — Mon, 14 Oct 2024 13:51:34 +0000

> I agree that free software is a good way to produce software for lots of reasons.

I do too, I'm pragmatic about it but my preference is toward FOSS whenever it meets requirements.

> I don't think funding it is necessarily a good way to encourage that, in practice. Most good free software comes from people wanting to scratch itches

Everyone doing FOSS work is funded in some way, developers aren't homeless monks living in communes, they have jobs, they eat food, sleep under a roof, which costs money. So the question is _who's_ itch is being scratched, often the case is that its an _employer_, Linux kernel is a perfect example of that, the Linux Foundation is a vendor consortium where large private corporations coordinate their efforts.

> I don't think it's right to try to create social pressure when they're not doing anything wrong. There is nothing wrong with "taking advantage" of a public good without contributing back. That's the whole point of free software!

Public good doesn't come about magically, it's intentionally created, subsidized and maintained or it atrophies and goes away, at some level it's self-destructive for large corporations to extract large monetary value from freely-provided software without maintaining the conditions that allow that software to be created in the first place. Putting social pressure on takers (or legal pressure like the GPL) to encourage them to contribute to the well-being of the environment that allows them to profit is the mildest form of pressure there is, a stronger form would tax profits at a level which pays for UBI so that developers really could work on whatever they wanted, scratch whatever itch, take risks, without worrying about their standard of living.

No software is free to *make* as that devalues human labor which is not provided for free, its always paid for in some way, we can be explicit and intentional about how that happens.

Wrong solution to the wrong problem

roc — Mon, 14 Oct 2024 04:33:46 +0000

> Most good free software comes from people wanting to scratch itches, and monetising it runs the risk of encouraging people to do it for the wrong reasons which I doubt leads to much good.

For most people, if they can't monetise their free software project then it has to remain a part-time occupation at most while they work a paid full-time job to support themselves. That is problematic, because a lot of valuable projects demand more effort than that. Where would we be if Linus had to hold down a full-time not-Linux job and could only work on Linux during evenings and weekends when he wasn't too tired?

I've been around a long time and I've seen very few cases of talented but venal software developers messing up projects they joined "for the wrong reasons". You get problematic contributors but they're much more often motivated by pride than money.

How does this obligation attach?

Wol — Sat, 12 Oct 2024 19:49:54 +0000

After I posted that, I thought "make it moderately easy to get brone/silver with the engineering budget" and then you use the advertising budget to transition to gold. We just need to make sure it's seen as a "good thing" to help provide the public good ... ime companies want to pay for support, we just need to make sure that support goes to the right places.

Cheers,
Wol

How does this obligation attach?

rcampos — Sat, 12 Oct 2024 14:30:56 +0000

Interesting. But for that to work, then, there would need to be a good social acceptance that influences people making the $choice$ to prefer companies that fund open-source, I guess?

It seems like easy to get in the same issues as sustainability, like "we do X, certified by Y" and it doesn't really mean something useful. Maybe a slightly better problem to have, if we gained social acceptance... but unsure how to turn that into quality funding.

Wrong solution to the wrong problem

Wol — Sat, 12 Oct 2024 13:14:34 +0000

> Historically, most computer science researchers got paid in various ways --- in many cases, through government programs.

And, given that a lot of that money is/was funnelled through universities, how much of it is/was of practical use?

Ivory tower research is all well and good, but at the moment it seems to be fuelling gas-guzzling data centres producing AI artificial stupidity results and the like, and probably always has. My first experience with "intelligent" web search engines like Jeeves et al, was to scream at them "stop giving me more of the same, you've screwed up, that's not what I'm looking for !!!". And the more I tried to vary my search terms, it seemed like they tried ever harder to come up with the same results !!!

The end result of most government funded research seems to be a paper and a Ph.D. and then it just gets filed and forgotten. Not saying it wasn't money well spent - education is never wasted - but so much of this seems to be abandoned once it's served its purpose of making the researcher more employable.

And how much of this research is simply discovering the same thing over and over? Somebody pointed me at a paper on "cuckoo hashing" from the ?mid?-90s. Only for it to trigger the reaction in me "hey, that's what that early-80s paper was describing, except it didn't give it a name".

DARPA / seedcorn funding seems a far better bet. If somebody comes to you and says "hey this idea could change the world, I just need some help", this is where you provide sponsorship and say "it needs to be Open Source". And then let them divert funds to other people to help them ... (properly accounted for, of course).

Cheers,
Wol

Wrong solution to the wrong problem

smurf — Sat, 12 Oct 2024 09:07:42 +0000

> Most good free software comes from people wanting to scratch itches

Sure. The problem is that in some areas, the itches are too deep and/or require too much specialized knowledge and/or professional training / education.

For instance, consider why we don't have an open-source drafting program that an architect might want to use professionally. (No, FreeCAD doesn't count.)

Wrong solution to the wrong problem

milesrout — Sat, 12 Oct 2024 08:12:45 +0000

I agree that free software is a good way to produce software for lots of reasons. I don't think funding it is necessarily a good way to encourage that, in practice. Most good free software comes from people wanting to scratch itches, and monetising it runs the risk of encouraging people to do it for the wrong reasons which I doubt leads to much good. We already see this, for example, with security. People submit very low-quality fuzzer output because they want their names to be associated with Linux security. Or you see people posting nonsensical pull requests on GitHub projects in order to get free t-shirts for "contributing to open source software".

That said, funding people that have already scratched their itches to expand and maintain their projects probably isn't likely to be too harmful.

> Historically, most computer science researchers got paid in various ways --- in many cases, through government programs.

I don't think this is relevant at all. The point of my post is that the idea of programmes like the "Open Source Pledge" is to put social pressure on software businesses to give money in a certain way to a certain type of cause. I don't think that in this case they're solving the right problem. I also think the solution is a bit offensive. I don't think it's right to try to create social pressure when they're not doing anything wrong. There is nothing wrong with "taking advantage" of a public good without contributing back. That's the whole point of free software!

How does this obligation attach?

Wol — Sat, 12 Oct 2024 07:20:38 +0000

> Am I missing something?

Yes. Branding.

People pay higher prices for crappier products, because of the brand name.

This "open source pledge" needs to be paid for out of the *advertising* budget, and the brand managers need to be told to make good use of the money so spent ...

Whether that will ever happen ...

Cheers,
Wol

German state pilot program to hire open-source maintainers

rcampos — Sat, 12 Oct 2024 02:23:52 +0000

The German Sovereign Tech Fund has launched a new pilot program to provide compensation to individual contributors in 2025.

They offer lot of flexibility in the way to hire (so whatever is best for you), it's for 1 year and people living in Germany is eligible only.

The deadline to apply is October 20.

https://www.sovereigntechfund.de/programs/fellowship

How does this obligation attach?

rcampos — Sat, 12 Oct 2024 02:13:21 +0000

But under capitalism, why a company choose to include in their prices or what ever the "give to open source" part?

I mean, then other companies will choose not to, if OSS is funded all will be fine, and they would have the advantage in margins.

Then, others follow and we come back again to the same situation.

Am I missing something?

Wrong solution to the wrong problem

roc — Fri, 11 Oct 2024 05:46:22 +0000

Open-source is an extremely efficient way to produce software --- for lots of different reasons. If we can find good ways to fund it, people can afford to put a lot more effort into it, which means more useful projects and (perhaps more importantly) better-quality projects.

> Analogously, you don't have to go and pay a whole lot of money to some computer science researcher because you use an algorithm he devised and published decades ago in your product.

Historically, most computer science researchers got paid in various ways --- in many cases, through government programs.

Sentry is unfortunately not Free Software/Open Source

roc — Fri, 11 Oct 2024 05:40:11 +0000

Ok, I'm sorry.

I remain genuinely puzzled by the issue here. What's wrong with non-free software companies supporting free software?

Wrong solution to the wrong problem

wittenberg — Thu, 10 Oct 2024 16:34:51 +0000

I agree that the point of free software was that people freely donate the fruits of their labor because they enjoy their labor. When it was just a bunch of hobbyists building things for the joy of it that worked very well. But now we're in a world where almost everyone depends (directly or indirectly) on these projects. They are essentially public goods (in the economists' sense of the word). Public goods are traditionally (in capitalist societies) paid for out of taxes.

This leads to two problems: Who should collect the taxes (and The Open Source Pledge is one proposal), and who should get paid. My (very few, very small) contributions were made when I was working for a company that occasionally used open source programs, and I contributed fixes which made my job easier. That was the ethos of the time. If that code is still in use, who, if anyone should be paid? Me, my employer, or nobody, on the grounds that I got far more benefit than I contributed?

Getting paid normally means accepting certain responsibilities. What responsibilities does an open-source coder or maintainer accept if he/she gets paid? To make changes that would benefit the group paying? To accept liability if the code causes damage?

We are trying to take a method of work which was set up by hobbyists for hobbyists and change it to support a system that most of the world runs on. Can this be done incrementally? If not, is there the will to install a major re-design?

--David

Wrong solution to the wrong problem

milesrout — Thu, 10 Oct 2024 12:06:18 +0000

The assumed problem in all of this for many of the people involved is that some firms are freeriders on the work of others. But that's the whole point of free software! It is meant to be freely redistributable. It is meant to be a gift to the world. That is why people do it, and also why people don't: because some people don't want to give away their work.

So I really don't see what the issue is that requires so much handwringing. There is no moral obligation to "pay back" free resources that have been given to humanity just because you've started making money. Analogously, you don't have to go and pay a whole lot of money to some computer science researcher because you use an algorithm he devised and published decades ago in your product.

People publish free software for many reasons but it seems a lot of people now seem to think it is a good way to get drive-by contributions and some street cred, but balk immediately when they find out the real consequences of what they've done: give away the results of their work. I don't have much sympathy tbh.

Load-Bearing Internet People / Loadsharers

jnareb — Thu, 10 Oct 2024 12:00:13 +0000

There were many attempts to solve the problem of OSS maintainers doing unpaid work on critical systems. One of those I remember was ESR (Eric S. Raymond) "Load-Bearing Internet People" project, later renamed to "Loadsharers"... though I don't see that it got any significant uptake:

Announcement: http://esr.ibiblio.org/?p=8383
Project page: https://esr.gitlab.io/loadsharers/

Gold, Silver, Bronze

milesrout — Thu, 10 Oct 2024 11:33:24 +0000

I don't care at all what is funded. The software I use was created by people that do it for fun. Yes people being paid to write free software have contributed a lot since then, but thry arent even covered by this pledge!

Is the cost negotiable?

kleptog — Thu, 10 Oct 2024 09:09:17 +0000

If you look at the original blog it estimates more like 1-3% of the total employee compensation. So for the US that $2,000 might seem reasonable, but for just about anywhere else that's way too high. I think 2% of whatever the local compensation is is reasonable though.

Gold, Silver, Bronze

taladar — Thu, 10 Oct 2024 08:27:49 +0000

The problem is that "software that benefits you" is probably going to end up only benefiting direct dependencies while you would really want the entire dependency tree funded at the very least. Even that might be insufficient if you consider how much open source software benefits you indirectly, e.g. because some product you use that isn't software at all can be cheaper because that company has to pay less to get their website developed or because some data you need is available only because it can be hosted as an open API based on open source software.

Sentry is unfortunately not Free Software/Open Source

smurf — Thu, 10 Oct 2024 08:17:12 +0000

Promoting and building infrastructure for the support of FOSS is not "morally grandstanding", it's acknowledging that we're all standing on the shoulders of all those guys+gals in Nebraska+elsewhere.

If you don't like people being able to take their code out of FOSS, well then don't contribute to MIT-licensed stuff and don't sign any CLA. Fine.

Re-licensing some is not morally or otherwise objectionable if the license allows it. The code (as of before the license change) is still free and can be forked easily. This is the real world, not some FSF utopia which considers non-libre source to be morally objectionable *in itself*.

If you want "objectionable", then let's talk about all the closed-source Linux device drivers and the locked-up devices and whatnot out there. These *do* prevent you from exercising your free-source rights.

How does this obligation attach?

fw — Thu, 10 Oct 2024 07:30:03 +0000

I think it's rare for the enterprise Linux to fund development through open-source foundations, donations, or direct contracting for specific tasks (as opposed to employment). The Linux Foundation is an exception that it has such funding and that it pays some developers. Other foundations mostly maintain project infrastructure (both legal and technical). Enterprise Linux companies seem to prefer to directly hire upstream developers, or grow developers into key upstream roles. For certain core components, this has historically worked quite well. Not everyone wants to work in a corporate environment, of course, so more options may make sense. On the other hand, a lot of the donation/micro-transaction approach are a clear attempt to turn open-source development into a gig economy, which I find rather troubling personally.

Gold, Silver, Bronze

Wol — Thu, 10 Oct 2024 07:24:41 +0000

So sponsoring that guy in Nebraska should be rated lower than paying someone to scratch your itch?

At the end of the day, I wanted my "different standards" to provoke thought, and I certainly (partially) agree with you. The trouble is, there are arguments every which way round Spaghetti Junction :-)

Cheers,
Wol

Sentry is unfortunately not Free Software/Open Source

milesrout — Thu, 10 Oct 2024 04:49:49 +0000

I agree. Sentry isn't free software. Yet they're morally grandstanding about the importance of supporting free software? It is like seeing Apple sponsoring an Android conference.

Gold, Silver, Bronze

milesrout — Thu, 10 Oct 2024 04:48:03 +0000

>Then, of course, you have the Gold Standard, which is paying people to work on Open Source No Strings Attached.

To me this isnt the gold standard at all. Why should unaccountable money being given to foundations to be awarded to people scratching their personal itches for riches be considered a higher ethical tier than giving money to someone for a particular service?

I would say the opposite is true. Paying someone to produce free software that directly benefits you is far better. It is more sustainable, as it is a real commercial transaction that won't disappear as soon as the books tighten.

Sentry is unfortunately not Free Software/Open Source

milesrout — Thu, 10 Oct 2024 04:41:48 +0000

What a nasty, bad-faith interpretation of florian's comment. That is obviously not what florian said AT ALL.

Sentry is unfortunately not Free Software/Open Source

roc — Wed, 09 Oct 2024 19:46:52 +0000

So profits from non-free software are tainted and should not be accepted by a charity? That's a pretty extreme position.

How does this obligation attach?

NYKevin — Wed, 09 Oct 2024 17:36:36 +0000

In any other field, if I pay a professional to provide me with some good or service, then any obligations related to producing that good or service are the other person's problem. If they need to buy carbon credits, make sure the raw materials are fair trade, or whatever, those things might cause my bill to go up, but I don't need to actually do those things myself. I may choose to look at the company's reputation and marketing copy to verify that they are in fact doing those things, but I might not, and it's really not anybody's business but my own.

So in the context of FOSS, I'm having a hard time believing that payments to Red Hat and the like are insufficient, unless your company is also using a lot of FOSS that Red Hat did not give you (or we're arguing that Red Hat is an unethical provider of FOSS, which is a more complicated discussion so let's not do that here - think of "Red Hat" as an arbitrary example of a company that *could* do FOSS-for-sale right, and not as Red Hat the company that really exists and really sells FOSS today). Red Hat could then give whatever donation is appropriate under the circumstances (which probably is not per-client-developer, but instead something like per-seat-license), and price it into the product. Of course, for companies that use FOSS and provide free-as-in-beer software to the general public, you can't really price it into the product directly, but many of those companies have scads of cash that they could part with, and most of them also have various dials they can turn on their monetization models to make their products "more expensive" in other (non-monetary) terms.

When I read the "napkin math" carefully, it seems to tacitly assume that we're including all software shops whatsoever. Presumably, the rationale is that the industry as a whole benefits from FOSS, and so the industry as a whole should contribute to FOSS. While that is a nice ideal, we live in a capitalist system. It is far easier to persuade people to pay for something by saying "here is the bill" than by saying "please give money to [whatever cause]." So I think it makes more sense for these contributions to come primarily from for-profit packagers and distributors, who can then modify their business models in whatever way (pricing, profit margins, advertising, etc.) makes those contributions make sense.

I do not want this comment to be mistaken for an argument against the pledge as a whole. In general, I do agree that companies should contribute more to FOSS maintenance and development, and this pledge is generally a step in the right direction (assuming that it works). I'm just not entirely sure it makes sense for companies whose use of FOSS is more attenuated, and so the pledge may end up underestimating the necessary size of the contribution (i.e. you will have a lot of companies say something like "well, we don't *really* use that much FOSS" because the FOSS they use is mainly provided by third parties and not directly downloaded, they will not contribute anything, and then the contributions you do get will not collectively make up for that).

Incentive to reduce employed developers?

aragilar — Wed, 09 Oct 2024 11:53:24 +0000

$2000 USD (I'm presuming the value is static, as opposed being 2000 of the local currency or some kind of equivalent scaling based on the local costs) can be quite large depending on the location (as noted in https://lwn.net/Articles/993355/), and the original maths assumed Google-level salaries, so while I agree it doesn't seem likely to change much overall (given the targets seem to be profitable US-based companies), there is the small possibility of misaligned incentives.

Incentive to reduce employed developers?

taladar — Wed, 09 Oct 2024 08:54:00 +0000

Oh, I missed the bit where the donation figures where per year, not per month. That indeed makes it less likely to be an issue.

Sentry is unfortunately not Free Software/Open Source

smurf — Wed, 09 Oct 2024 08:23:28 +0000

So just treat them like any other commercial software shop that happens to support the Pledge with their wallet.

Incentive to reduce employed developers?

josh — Wed, 09 Oct 2024 08:07:23 +0000

The fully loaded cost of a software engineer is often at least a hundred times this figure, if not higher. Adding 1% is not going to suddenly create a new incentive to fire people where none existed before.

Incentive to reduce employed developers?

taladar — Wed, 09 Oct 2024 07:28:14 +0000

The problem I see with this per developer calculation is that it sets an incentive to reduce the number of employed developers. No developer is better off if the company boosts their per developer donation and mentions it everywhere for marketing purposes but does so by reducing the number of developers they employ.

Gold, Silver, Bronze

Wol — Wed, 09 Oct 2024 06:13:56 +0000

I get that they don't want to count general subscriptions like RedHat or SUSE "licence" fees towards this, but they are still payments towards Open Source.

I'd like to see something like general maintenance fees towards eg RH, SUSE, Canonical, count towards a Bronze (maybe you have to pay $5000 per developer to account for the reduced value).

If you sponsor things of direct benefit to you, that should still count so long as the result is FLOSS - after all, it's helping maintainers eat and the result is available to everyone (I'm thinking of payments towards eg Collabra for LibreOffice), and that would get you a Silver.

Then, of course, you have the Gold Standard, which is paying people to work on Open Source No Strings Attached.

But having those lower tiers would get companies hooked, and once they've pledged and seen the benefits, they'll want to go up the scale. Of course, we then have little problems like what happens if a small end-user company with nobody classed as a developer decides to pay a load of money towards Silver class subscriptions ... but maybe that's telling us that the focus on "developer" may be a false one ...

Cheers,
Wol

Sentry is unfortunately not Free Software/Open Source

floriansnow — Wed, 09 Oct 2024 05:47:29 +0000

This sounds like a great effort, but Sentry being the main entity behind this, weakens the whole endeavor unfortunately. Sentry is great, but it's no longer Free Software. I know it becomes Free Software after two years, and this is much better than simply being proprietary, but unfortunately this means that there's now a great effort to support Free Software lead by a company with a primarily non-free product. If this pledge helped it become free again, that would be great and if this decision came with this pledge, I would feel much better about it.