I would also argue that it holds true for data as well: if some data is collected using public funds, it should be released as open data (as long as it does not contain sensitive and personal information of the population).
This is such an obvious thing to me, I have a hard time understanding how this can possibly not be the case already. Governments have a lot of spending power and they should use it wisely.
It’s a very obvious thing for us, but unfortunately there’s a lot of people in the top of the hierarchy chain that has no idea what the hell open source is, because 1) never heard of it or 2) think it is some thing that only hobbyists do, therefore not for serious people.
In my field (biology, public universities), until some years ago, talking about open sourcing your software would be met with a question “but aren’t you afraid that someone will steal your code?” There are still cases like that, of people that don’t provide their code for some analysis (like the case of one professor I knew that didn’t want to provide his R script that he used to determine why a certain area should be a conservation area), which is very weird to me. How is the third party going to know exactly how they came to that conclusion? ¯_(ツ)_/¯
Big, proprietary vendors be spending big on politicians. Copyright and patent laws remained strongly in their favor. Product purchases by the government for a long time, too. That corruption probably goes a long way toward maintaining lock-in and monopolistic/oligopolistic practices. They also benefit from university, tech transfer for a decent chunk of research getting federal funding.
Then, the publishers make lots of money on selling articles. They don’t want them being open. So, they’re mostly fighting against that trend. I don’t know if they’re doing campaign contributions to back it up. I did read they demand copyright of most stuff submitted to those journals. So, they can resist under copyright law.
Plenty of incumbents with lots of reasons to like the current system.
I used to be very much of this mindset, but having worked on government software projects, some of which have been open sourced, I’m now … less enthusiastic about it.
Open-sourcing code is not as simple as just pushing code to a public repo. Public code needs to be vetted more stringently and often there is an expectation (as discussed here a lot recently) of support, reliability, ability to accept contributions etc.
Open-sourcing is likely to improve the quality of the software. But, quality is a variable that the organisation should be able to choose. There are many internal government systems that serve their purpose quite well without being well architected (or frankly, well implemented). Sometimes cheap and nasty meets the business need, and that’s OK.
A bespoke system dealing with a unique set of business rules targeted towards a narrow and fixed set of hardware (which is what a lot of government-produced software is) is unlikely to be of use to anyone else, so the extra cost incurred comes with no real benefit. You can argue that it shouldn’t be like that, but that’s a far deeper issue than software.
That said, there’s certainly a place for open sourcing some systems, particularly public facing ones. But it should be done mindfully rather than by default.
often there is an expectation (as discussed here a lot recently) of support, reliability, ability to accept contributions etc.
That part is easy to solve. You just tell them the software is free to use. The owner is not responsible for or interested in anything from that point on. If they use it, they’re on their own. Some might also mention being willing to work on it for a price. The consistent message will be anyone giving away something for free should not be expected to do any extra, free work on it. We can ask but they can decline.
I think this is even less likely to fly for government produced OSS than for the regular variety, simply because there is already a (valid, mostly) expectation of government to be a service provider.
I could see them trying it. They already provided the service, though, by creating the OSS using the money they had. If they want more service, it costs more money. We could compromise a bit by making sure they distinguish between unsupported one-offs and OSS that will steadily be supported. From there, how much support still must be determined. A small-budget project can only do so much. A large-budget program should be expected to deliver a lot more.
When I was a doctoral student financed form public sources, I was very happy that all the code we produced was released under a liberal open source license (MIT). It really felt like the right thing to do.
When you buy open licence prices go up significantly. Also, it’s often cheaper to buy a completely new software than modify existing one. It’s not a city council job/field of expertise to develop software.
There is a reasonable middle ground: open APIs and open data. That prevents vendor lock in.
Why would prices go up if the license is open? If it’s a custom job which is built from scratch it shouldn’t matter; you’d typically be paying by the hour for the development effort anyway.
The customer (government in this case) determines the requirements, so that would simply mean those companies who “don’t want to give their intellectual property away” wouldn’t get the job.
Of course, but that’s the point behind this whole discussion: governments get to dictate the rules, but they have to decide to actually do this. If legislation says the rules include requiring opening up the code, it really becomes that simple: as a supplier, either you abide by the rules or you’re out.
A much more interesting spin on this idea is to read it backwards: Open Source code should be funded from a public source. In my opinion something like a corporate tax is the only way to find it, otherwise there’s no incentive.
I would also argue that it holds true for data as well: if some data is collected using public funds, it should be released as open data (as long as it does not contain sensitive and personal information of the population).
This is such an obvious thing to me, I have a hard time understanding how this can possibly not be the case already. Governments have a lot of spending power and they should use it wisely.
It’s a very obvious thing for us, but unfortunately there’s a lot of people in the top of the hierarchy chain that has no idea what the hell open source is, because 1) never heard of it or 2) think it is some thing that only hobbyists do, therefore not for serious people.
In my field (biology, public universities), until some years ago, talking about open sourcing your software would be met with a question “but aren’t you afraid that someone will steal your code?” There are still cases like that, of people that don’t provide their code for some analysis (like the case of one professor I knew that didn’t want to provide his R script that he used to determine why a certain area should be a conservation area), which is very weird to me. How is the third party going to know exactly how they came to that conclusion? ¯_(ツ)_/¯
Ever heard of corruption? :-)
Big, proprietary vendors be spending big on politicians. Copyright and patent laws remained strongly in their favor. Product purchases by the government for a long time, too. That corruption probably goes a long way toward maintaining lock-in and monopolistic/oligopolistic practices. They also benefit from university, tech transfer for a decent chunk of research getting federal funding.
Then, the publishers make lots of money on selling articles. They don’t want them being open. So, they’re mostly fighting against that trend. I don’t know if they’re doing campaign contributions to back it up. I did read they demand copyright of most stuff submitted to those journals. So, they can resist under copyright law.
Plenty of incumbents with lots of reasons to like the current system.
I used to be very much of this mindset, but having worked on government software projects, some of which have been open sourced, I’m now … less enthusiastic about it.
Open-sourcing code is not as simple as just pushing code to a public repo. Public code needs to be vetted more stringently and often there is an expectation (as discussed here a lot recently) of support, reliability, ability to accept contributions etc.
Open-sourcing is likely to improve the quality of the software. But, quality is a variable that the organisation should be able to choose. There are many internal government systems that serve their purpose quite well without being well architected (or frankly, well implemented). Sometimes cheap and nasty meets the business need, and that’s OK.
A bespoke system dealing with a unique set of business rules targeted towards a narrow and fixed set of hardware (which is what a lot of government-produced software is) is unlikely to be of use to anyone else, so the extra cost incurred comes with no real benefit. You can argue that it shouldn’t be like that, but that’s a far deeper issue than software.
That said, there’s certainly a place for open sourcing some systems, particularly public facing ones. But it should be done mindfully rather than by default.
That part is easy to solve. You just tell them the software is free to use. The owner is not responsible for or interested in anything from that point on. If they use it, they’re on their own. Some might also mention being willing to work on it for a price. The consistent message will be anyone giving away something for free should not be expected to do any extra, free work on it. We can ask but they can decline.
I think this is even less likely to fly for government produced OSS than for the regular variety, simply because there is already a (valid, mostly) expectation of government to be a service provider.
I could see them trying it. They already provided the service, though, by creating the OSS using the money they had. If they want more service, it costs more money. We could compromise a bit by making sure they distinguish between unsupported one-offs and OSS that will steadily be supported. From there, how much support still must be determined. A small-budget project can only do so much. A large-budget program should be expected to deliver a lot more.
When I was a doctoral student financed form public sources, I was very happy that all the code we produced was released under a liberal open source license (MIT). It really felt like the right thing to do.
Google was originally funded with a public grant ;-)
When you buy open licence prices go up significantly. Also, it’s often cheaper to buy a completely new software than modify existing one. It’s not a city council job/field of expertise to develop software. There is a reasonable middle ground: open APIs and open data. That prevents vendor lock in.
Why would prices go up if the license is open? If it’s a custom job which is built from scratch it shouldn’t matter; you’d typically be paying by the hour for the development effort anyway.
Not everyone is happy about giving out their intellectual property away.
The customer (government in this case) determines the requirements, so that would simply mean those companies who “don’t want to give their intellectual property away” wouldn’t get the job.
It’s not that simple. Government procurement has it’s own rules and procedures, you can’t set arbitrary requirements.
Of course, but that’s the point behind this whole discussion: governments get to dictate the rules, but they have to decide to actually do this. If legislation says the rules include requiring opening up the code, it really becomes that simple: as a supplier, either you abide by the rules or you’re out.
A much more interesting spin on this idea is to read it backwards: Open Source code should be funded from a public source. In my opinion something like a corporate tax is the only way to find it, otherwise there’s no incentive.
The person holding the purse strings controls the product.
I don’t want government setting software requirements for the rest of us.