One of the g0v movement strategies is that most of our products that are intending to be official at some day, are CC0. The government doesn’t really need to ask our copyright if they somehow make it the official version. That’s pretty much it.
It took about eight months to arrive at a national version, which is mostly the same as the initial community draft. It has a disclaimer, I think, but I don’t think there’s any substantial differences. Once that happens, now we have something that we can refer to when saying that all the government agencies must publish data under a open definition conforming license.
When they need to choose one, they are now faced to either invent their own bespoke license and then submitting to meet the open definition, or they can just take the one that, by definition, fits the open definition. Most choose the latter.
Now, even the National Palace Museum publishes under this license, so we don’t have a proliferation or ambiguity of licenses as a result of this.
Yeah, certainly. During the community discussion, there is actually a lot of issues raised in the Hackpad discussion.
For example, there was a use case where the providing agency, as they often did, requested a copy of any derived work from the data - be it an application or whatever, as a tracking mechanism, if you will.
That’s actually one of the very common clauses before open data, when we only had the freedom of information, Public Information Publishing Act, is that if you make derived works, let us know somehow, as part of the terms of use.
There was quite a few discussions around this clause, because there’s clearly a need by the providing agency, but it’s actually not so clear that it actually fits open definition. I may want to make something that really requires a very high barrier of entry, and I don’t want to provide a copy of that to the providing agency.
After some debate, it was deleted from the community version, because it’s not part of any of the three initial design goals. It’s more of a request, because it wouldn’t then fit the open definition definitions, so they took it out.
There were some other issues as well. There was about whether usage of this data or misuse can cause the damage in reputation for the providing agency, and whether the user of this data should provide the warranty or compensate for any damages.
It was then discussed after quite a few discussion that existing civil code actually covers this already. It really doesn’t matter whether it’s part of the license or not, so that was taken out. I think that’s the two substantial debates that we have.
There are a few things. The first is that this one-way transition clause is actually very familiar to us, especially to me, because the first open-source license that I encounter was the Artistic License as used in the Perl community.
Larry Wall designed it, Artistic License specifically, to make people using the GNU Public License happy, because they can, if they want, transition from Artistic License to the GPL.
At the time, GPL has a very large number of users, and Artistic License is only used by Perl or a few other software packages. He wanted to make sure that it’s considered part of the larger free software movement, so this device is promoted to us.
As why do we not just choose CC 4.0? Was that the question?
At that time, there is no Chinese version of Creative Commons 4.0. It was not translated until 2016, if I am not mistaken. It wasn’t an option. While we could of course use CC BY 3.0 Taiwan, but that doesn’t really work.
Until we have a truly international Creative Commons license, I don’t think it’s easier or any more comfortable than a local license with a transition clause. That was back then. I’m not sure if we do the whole exercise now, wouldn’t we just choose CC 4. Maybe we would, but that’s beside the point.
That’s exactly right. We take, for example, the data has its origin URL. You use that as the attribution, and that will completely state the license quite well.
As long as it’s fitting the open definition. For example, they can also choose the CC Share-Alike license.
Sure. I’m pasting you that. Actually, all the relevant documents is listed in one single page in the open data platform. I’ll paste you that. You can follow all the links. Here we go.
It also talks very clearly about how all the new ICT systems must produce open data by default, unless its maintenance cost over the past three years is one million euros or more, or things like that. You can see the open definition is our definition of open at the very first paragraph in section two.
We actually do. Let me find a power adapter. I’ll be back in a minute.
I’m back. Batteries are the only technology that doesn’t improve exponentially.
We do have an automated machine-to-machine protocol to syndicate between data portals. That is to say both the regional city portals and national city portals exchange through this machine-to-machine interface. That answers the technical part of the question.
As of whether the agencies themselves still publish the URL to this page that clearly says the license applies to all the data provided online. They are more and more doing so. We are not 100 percent yet, but whenever someone sees something, then we send a letter for them to change. So far, there’s no resistance.
The same data is provided often in two forms. It’s provided as data set over at data.gov.tw, which we handle through M2M, machine-to-machine APIs.
On the other hand, it may be also offered as a download link, for example, to a ZIP file or to a CSV file, from its normal website. I’m referring more to the latter, where the term of use of one specific website also overlappingly refers to the data that may be followed through their website.
We’ve updated the template. We are at the moment not strictly speaking saying that all the agencies must switch over to the new license. Sometimes, an external user or experts would make the suggestion. So far, all the ministry that received the suggestion, that I’m aware of, made the change.
In our case, it’s actually, we don’t provide a way for agencies or city governments to upload their data, per se. The national Open Data Platform only uploads the metadata. The actual download still goes to the respective agencies.
Again, because the community wants it. I’m just the sounding board for the community. In this case, it was proposed by Lucien, the same person who originally drafted this license.
It’s important because we are moving toward defining open data as fitting the open definition, and perhaps -- I’m guessing, here -- that he wants not just the one-way transition clause part, which automatically fits the open definition, but the design itself to be verified by the open knowledge community.
Yeah. That’s documented in the OAS standard. It’s a updated version of Swagger. We actually also adopted that as a national standard. We’re the first, at the moment, the only one to do so.
Again, for all the newly-constructed systems, the agency can, as part of IFP, to require at zero or very little cost, a production of the OAS schema, in addition of the open data set -- not open, or any structured data set.
Sure. It’s just fine. All right.
Yeah, you too. Thank you for the interview. Bye.
Since October 1 last year.
More or less.
Sure.
Or I can practice my sign language.
Yes. [writes on tablet and holds it up] That’s my sign language.
Yeah, it’s a sign.
A what?
Yeah, I know that work.
The main challenge is the size and weight of the battery. Once they solve that, it’s...
Everything’s exponential but batteries.
No?
Yeah, I’m aware of that...
We were just working here on the OpenCV recognition of post-it notes...
...so that we can paste on those whiteboards and have it all magically transcribed to online versions. They’re coding that right now. We were just joking because if you wear something like this, like an optical camouflage, OpenCV will see all these faces...
That’s right. That’s what we’re doing too.