8 ms·

One nice thing about free OSS is that it's not paid for. Hence, less influence by external forces, fewer obligations, no strict deadlines, etc. More often than

One nice thing about free OSS is that it's not paid for. Hence, less influence by external forces, fewer obligations, no strict deadlines, etc. More often than not, this leads to better quality of such software (given that maintainers are experienced and smart). I'm not sure how establishing a direct relationship between money and work (as opposed to donations and grants) would affect the rather high standards of good OSS..

chrisseaton 8y ago
I would imagine that the vast majority of free/OSS is done as paid employment.
- jlarocco 8y ago
  Do you have anything to back that up, because it directly contradicts my experience. I've only seen people paid to work on the biggest projects, often only when it's become entrenched in everybody's workflow (Linux, some GCC developers, some Python devs), had the source code released after closed development (OpenSolaris, Firefox), or got taken over by corporation (Clang and Apple, khtml/webkit/Chromium with Google and Apple). The vast majority of OSS projects IME are tiny, have few developers, and not many users outside of the developer.
  - chrisseaton 8y ago
    > Do you have anything to back that up Have you read Berdou's PhD research on the topic? http://www.lse.ac.uk/media@lse/study/pdf/PhD_Berdou_2007.pdf http://www.lse.ac.uk/media@lse/study/pdf/PhD_Berdou_2007.pdf Paid developers are more likely to maintain critical parts of open source code bases.
    - jlarocco 8y ago
      I had not, but it seems to back up my point. From the Overview section: "F/OS software is developed by online communities of globally distributed contributors, the majority of whom are volunteers who rarely meet face to face." And it uses KDE and GNOME as case studies, which are both very large, entrenched projects that are subsidized by companies like Red Hat and Canonical. What I'm getting at, though, is that if you pick 1000 arbitrary repos on GitHub (or BitBucket or GitLab), 999 of them are not going to have any paid contributors. Once you start filtering by user and/or contributor counts you're removing a large amount of projects that aren't popular, but never the less count as open source projects.
      - chrisseaton 8y ago
        That's the majority of people, not the majority of the work. I said majority of the work. The PhD supports my assertion that the majority of the actual work is done paid. I don't believe you for your 999 out of 1000 repos, but I'm not sure how we'd check that. When I browse people's GitHub profiles it's almost all little forks of things they're clearly using in their work. I don't know what people think OSS is - it's not hobbies. The majority of our infrastructure is OSS - operating systems, compilers, browsers, virtual machines, tools, it's almost all OSS and all majority paid contributors.
        jlarocco 8y ago
        > That's the majority of people, not the majority of the work. I said majority of the work. The PhD supports my assertion that the majority of the actual work is done paid. The PhD supports that the majority of work on KDE and Gnome are done paid. Those are not typical examples, though, and I don't believe the conclusions apply to OSS in general. Projects like those make up a tiny percentage of the OSS ecosystem, whether measuring by people or by work.
- atomashpolskiy 8y ago
  Great if so. Writing OSS as part of the job is actually a good thing, because (1) there's a relatively stable entity that is responsible for this software at the end of the day (and not a single proud and starving artist, who may get ill, get married, get another project, etc.) and (2) someone actually has a _real job_
zapita 8y ago
That is a naive thing to say. Open-source work doesn’t become purer when underpaid. Making open-source pay better for everyone, not just a privileged few, will allow more people to join the open-source movement, raising the standard of software quality for everyone. Helping open-source maintainers get paid is better for open-source. I think what the Devflight people are trying to do is admirable. I hope they succeed.
- atomashpolskiy 8y ago
  Everyone who? It's highly unlikely that companies will pay an awful lot of money to these maintainers anyways, so only the less "priviled" (i.e. second/third world citizens mostly) will benefit. Moreover, while some individual maintainers will gain in short term, the developers as a workforce will definitely lose (because certain tasks are taken from freelancers, contract workers and employees and distributed to, well, public). Business does not need more developers (let alone better paid ones), it needs more work done faster and for less money. Also, I think that most people, who maintain decent OSS projects, have well-paying jobs anyways, so they don't need more money and more work to worry about. What they really need is pull requests and bug reports. Which the business can provide by _hiring_ people, who need money and work, to do the job and send the patches. Double benefit, everyone's happy.
  - chuhnk 8y ago
    Can I ask. Do you write or maintain OSS code? Do you have large number of users? Do you work full time on OSS? Because your opinions are incredibly naive.
    - atomashpolskiy 8y ago
      That's completely irrelevant. Working full time on OSS means being a businessman in disguise (and, arguably, perverts the very ideas of free OSS), so we're on the opposite sides of the fence, I presume. Or do you live off something else and work on your projects completely out of sheer altruism?
vbordo 8y ago
This is a fascinating point. I absolutely agree that injecting money into software can have unintended consequences. After speaking with a lot of maintainers, one common theme is that the widespread commercial use of open-source software has already created this direct link between money and work. Maintainers experience pressure to continue developing their project to keep up with users, issues and feature requests. The more popular the project, the more value is being created on top of their software, and the higher the pressure. Because maintainers are often required to work on their projects in their spare time, working around a full-time job, it’s easy to burn out and stop working on the project altogether. We’re using this contract route as a way to make open-source sustainable for maintainers and establish clear expectations for both maintainers and companies. Setting those expectations is key to giving maintainers their ideal work scenario where they won’t have to sacrifice quality for speed.
- atomashpolskiy 8y ago
  Thank you for the reply. Let me ask you, why premium is fixed at 10%? As I see it, this would make sense only for large quantities of small payments. I.e. it seems like you're expecting exactly what you say you're hoping to eliminate: small feature requests, bugfixes, maybe some integration tasks, docs, etc. How about yearly support contracts?
  - Semaphor 8y ago
    One of us misunderstood what they do. The way I read it was that they establish a contract between dev and corp with regular payments (of which they take 10%), in exchange for this the corp gets the things listed above like prioritized bug reports and so on. But they don't pay for the single bug fixes or anything.
- alexellisuk 8y ago
  Exactly this: > Maintainers experience pressure to continue developing their project to keep up with users, issues and feature requests. The more popular the project, the more value is being created on top of their software, and the higher the pressure. Because maintainers are often required to work on their projects in their spare time, working around a full-time job, it’s easy to burn out and stop working on the project altogether.
tmastro 8y ago
Your point about donations and grants is a good one. Working with us does not mean maintainers have to exclusively use contracts to make money. Making open-source more sustainable will definitely require an aggregation of resources. We’ve learned from maintainers that donations can also have unintended consequences and create certain levels of expectations from the individuals and companies that make them. However, for some projects the donation model can work well, so we definitely encourage maintainers to also accept donations in these situations. Vue.js is a perfect example (https://vuejs.org/ https://vuejs.org/).

chrisseaton 8y ago

I would imagine that the vast majority of free/OSS is done as paid employment.

jlarocco 8y ago
Do you have anything to back that up, because it directly contradicts my experience. I've only seen people paid to work on the biggest projects, often only when it's become entrenched in everybody's workflow (Linux, some GCC developers, some Python devs), had the source code released after closed development (OpenSolaris, Firefox), or got taken over by corporation (Clang and Apple, khtml/webkit/Chromium with Google and Apple). The vast majority of OSS projects IME are tiny, have few developers, and not many users outside of the developer.
- chrisseaton 8y ago
  > Do you have anything to back that up Have you read Berdou's PhD research on the topic? http://www.lse.ac.uk/media@lse/study/pdf/PhD_Berdou_2007.pdf http://www.lse.ac.uk/media@lse/study/pdf/PhD_Berdou_2007.pdf Paid developers are more likely to maintain critical parts of open source code bases.
  - jlarocco 8y ago
    I had not, but it seems to back up my point. From the Overview section: "F/OS software is developed by online communities of globally distributed contributors, the majority of whom are volunteers who rarely meet face to face." And it uses KDE and GNOME as case studies, which are both very large, entrenched projects that are subsidized by companies like Red Hat and Canonical. What I'm getting at, though, is that if you pick 1000 arbitrary repos on GitHub (or BitBucket or GitLab), 999 of them are not going to have any paid contributors. Once you start filtering by user and/or contributor counts you're removing a large amount of projects that aren't popular, but never the less count as open source projects.
    - chrisseaton 8y ago
      That's the majority of people, not the majority of the work. I said majority of the work. The PhD supports my assertion that the majority of the actual work is done paid. I don't believe you for your 999 out of 1000 repos, but I'm not sure how we'd check that. When I browse people's GitHub profiles it's almost all little forks of things they're clearly using in their work. I don't know what people think OSS is - it's not hobbies. The majority of our infrastructure is OSS - operating systems, compilers, browsers, virtual machines, tools, it's almost all OSS and all majority paid contributors.
      - jlarocco 8y ago
        > That's the majority of people, not the majority of the work. I said majority of the work. The PhD supports my assertion that the majority of the actual work is done paid. The PhD supports that the majority of work on KDE and Gnome are done paid. Those are not typical examples, though, and I don't believe the conclusions apply to OSS in general. Projects like those make up a tiny percentage of the OSS ecosystem, whether measuring by people or by work.
atomashpolskiy 8y ago
Great if so. Writing OSS as part of the job is actually a good thing, because (1) there's a relatively stable entity that is responsible for this software at the end of the day (and not a single proud and starving artist, who may get ill, get married, get another project, etc.) and (2) someone actually has a _real job_

zapita 8y ago

That is a naive thing to say. Open-source work doesn’t become purer when underpaid. Making open-source pay better for everyone, not just a privileged few, will allow more people to join the open-source movement, raising the standard of software quality for everyone. Helping open-source maintainers get paid is better for open-source. I think what the Devflight people are trying to do is admirable. I hope they succeed.

atomashpolskiy 8y ago
Everyone who? It's highly unlikely that companies will pay an awful lot of money to these maintainers anyways, so only the less "priviled" (i.e. second/third world citizens mostly) will benefit. Moreover, while some individual maintainers will gain in short term, the developers as a workforce will definitely lose (because certain tasks are taken from freelancers, contract workers and employees and distributed to, well, public). Business does not need more developers (let alone better paid ones), it needs more work done faster and for less money. Also, I think that most people, who maintain decent OSS projects, have well-paying jobs anyways, so they don't need more money and more work to worry about. What they really need is pull requests and bug reports. Which the business can provide by _hiring_ people, who need money and work, to do the job and send the patches. Double benefit, everyone's happy.
- chuhnk 8y ago
  Can I ask. Do you write or maintain OSS code? Do you have large number of users? Do you work full time on OSS? Because your opinions are incredibly naive.
  - atomashpolskiy 8y ago
    That's completely irrelevant. Working full time on OSS means being a businessman in disguise (and, arguably, perverts the very ideas of free OSS), so we're on the opposite sides of the fence, I presume. Or do you live off something else and work on your projects completely out of sheer altruism?

vbordo 8y ago

This is a fascinating point. I absolutely agree that injecting money into software can have unintended consequences. After speaking with a lot of maintainers, one common theme is that the widespread commercial use of open-source software has already created this direct link between money and work. Maintainers experience pressure to continue developing their project to keep up with users, issues and feature requests. The more popular the project, the more value is being created on top of their software, and the higher the pressure. Because maintainers are often required to work on their projects in their spare time, working around a full-time job, it’s easy to burn out and stop working on the project altogether. We’re using this contract route as a way to make open-source sustainable for maintainers and establish clear expectations for both maintainers and companies. Setting those expectations is key to giving maintainers their ideal work scenario where they won’t have to sacrifice quality for speed.

atomashpolskiy 8y ago
Thank you for the reply. Let me ask you, why premium is fixed at 10%? As I see it, this would make sense only for large quantities of small payments. I.e. it seems like you're expecting exactly what you say you're hoping to eliminate: small feature requests, bugfixes, maybe some integration tasks, docs, etc. How about yearly support contracts?
- Semaphor 8y ago
  One of us misunderstood what they do. The way I read it was that they establish a contract between dev and corp with regular payments (of which they take 10%), in exchange for this the corp gets the things listed above like prioritized bug reports and so on. But they don't pay for the single bug fixes or anything.
alexellisuk 8y ago
Exactly this: > Maintainers experience pressure to continue developing their project to keep up with users, issues and feature requests. The more popular the project, the more value is being created on top of their software, and the higher the pressure. Because maintainers are often required to work on their projects in their spare time, working around a full-time job, it’s easy to burn out and stop working on the project altogether.

tmastro 8y ago

Your point about donations and grants is a good one. Working with us does not mean maintainers have to exclusively use contracts to make money. Making open-source more sustainable will definitely require an aggregation of resources. We’ve learned from maintainers that donations can also have unintended consequences and create certain levels of expectations from the individuals and companies that make them. However, for some projects the donation model can work well, so we definitely encourage maintainers to also accept donations in these situations. Vue.js is a perfect example (https://vuejs.org/ https://vuejs.org/).