Hi Mark, all
There is a difficult issue to be traversed. Namely that open source tooling needs to be fully ring fenced from non‑open source tooling in order to stay safe and usable.
I am not talking about an individual study that uses both open and closed source tooling to address a particular research question and then goes no further. I am talking about initiatives that run these two camps together under common workflows, possibly even code transfers, and invariably with an accompanying lack of discipline.
The underlying issues are legal and practical. To which one could add ethical, but I am going to omit that particular prism.
On the legal front, it is rather too easy to compromise the open source material in this context.
We know this from open source software more generally. That wider community is completely dedicated to rigorously maintaining that separation. Not because they are zealots (although some clearly are) but because that is the only effective way to manage the maze of legal risks present — stemming primarily from copyright law and patent law. I mix with the open source law community (for six years now) and defending the integrity of open source projects is their paramount concern. That includes protecting open source projects against misappropriation, license abuse, strategic litigation, and various forms of trolling.
On the practical front, researchers and analysts working outside of open source can become entirely reliant on both the goodwill and the terms‑of‑use set by the various institutes, consortiums, and multi‑lateral agencies who own and control these proprietary software assets. That will often work where relationships have been or can be established, but it cannot work more generally. If it would, then these various parties would simply add OSI‑approved open licenses to their tooling and be done with it. And in many cases, the inevitable costs, contracts, and related overheads are prohibitive.
It is easy to say that we, as an open modeling community, can accommodate some gray middle ground and that the world is moving toward open in any event. But that requires a level of trust that I personally am not prepared to countenance. And, as an open modeling community, we should be clear that only open provides the freedoms and permissions that we need to work efficiently, effectively, and collaboratively.
And to provide examples, FinPlan, MAED‑2, and LEAP are all non‑open. Indeed a presenter at the COP‑27 Climate Compatible Growth side event described MAED‑2 as "open source" when it is indeed proprietary and only available to selected parties (we talked through that misunderstanding afterwards quite amicably).
I would hope the hard lessons learned by the Linux kernel developers and the GNOME Foundation as litigants — and the various non‑profits such as FSF, FSFE, the Open Source Initiative, the Software Freedom Conservancy, OpenUK, and quite a number of others — can become recognized and accepted within our community.
Otherwise, I predict a rocky road ahead. As energy modeling rapidly spills out from niche research into policy‑relevant applications. And just for context, I sat next to a pro bono barrister for three days in 1995 in district court proceeding to protect our PlaNet Wellington Trust, an early community‑based ISP, from commercial intrusion. So I know the degree of effort required to defend such attacks, that creating legal vulnerabilities is all too easy, and that just one adverse court ruling would be terminal.
I also got caught up in the UNIX wars and my development environment (for the deeco framework) was summarily abandoned. Another example of how these risks can materialize.
Particularly in the context of analysis in the global south, we researchers mostly from the global north should only be offering best practice. For the reasons outlined, I cannot circulate programmes that fall short on open. I am sorry, as I know that this is vital work.
For those seeking background, I recommend this open access publication which you can download for free (you many need to remove some U+200D chars from the URLs to get them to work):
with best wishes, Robbie
You received this message because you are subscribed to the Google Groups "openmod initiative" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openmod-initiat...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openmod-initiative/DB9PR04MB823545DB585B3B2CE8BA9CF1F8FA9%40DB9PR04MB8235.eurprd04.prod.outlook.com.
-- Robbie Morrison Address: Schillerstrasse 85, 10627 Berlin, Germany Phone: +49.30.612-87617
Hello all again
A follow up — this time on open and non‑open data and with connections back to Africa and Germany.
Seeing I quoted the book edited by Amanda Brock, I should point out chapter 14 on "Sustainability and open source" has very significant shortcomings:
The key deficiency is that section 14.5 on open data covers nothing of the sort. The quoted United Nations Sustainable Development online database named "Open SDG Data Hub" operates under terms‑of‑use that state: (https://www.un.org/en/about-us/terms-of-use) (emphasis added):
The United Nations grants permission to Users to visit the Site and to download and copy the information, documents and materials (collectively, "Materials") from the Site for the User's personal, non-commercial use, without any right to resell or redistribute them or to compile or create derivative works therefrom, subject to the terms and conditions outlined below [which add further restrictions], and also subject to more specific restrictions that may apply to specific Material within this Site.
The mostly widely accepted definition for open data comes from the Open Knowledge Foundation via the Open Knowledge Definition. And recital 16 in the preamble of the European Union Open Data Directive broadly aligns.
So I will repeat recital 16 here — as this serves our purposes quite well and is also applicable to Germany:
"Open data as a concept is generally understood to denote data in an open format that can be freely used, re-used and shared by anyone for any purpose."
So describing material issued by the United Nations Statistics Division (UNSD) as "open data" in the absence of suitable open data licensing is very odd. And the stated presence of an "open API" is unlikely to be either true or material (it is certainly not true under US law, there was a recent Supreme Court case). And seeing the Linux Foundation (LF) is mentioned in that section 14.5, the LF have zero commitment to adopt genuinely open data licensing on any of their sponsored LF Energy projects (that the result of discussions on a mailing list under CHR which I am not at liberty to disclose).
And finally despite the phrase "open source" being in the title of this chapter, that concept is not referred to substantially in the text even once.
If you think we, as energy system analysts, can simply skate along on this doubtless thin and melting legal ice, then check this recent protective litigation in Germany:
This case is focused on press freedom but the wider agenda is usable and reusable data for any purpose. And the information in dispute is the remaining area for the construction of wind turbines under the respective distance rules for various federal states. The Bavarian state government claims its 96/9/EC database producer right was compromised by an earlier story in the taz newspaper, which also contained online information to support that story. And to complete the irony, some of the datasets present are under CC‑BY‑4.0 licensing in any case.
We owe it to ourselves to walk forward across this legal thin ice with our eyes open and our terminology accurate. And if the Bavarian state government can issue legal proceeding against public interest data users, then surely any other government or legal authority can also? We need widely accepted and interoperable open data licenses on this material. Nothing more and nothing less.
with best wishes, Robbie
To view this discussion on the web, visit https://groups.google.com/d/msgid/openmod-initiative/f14d5ef2-b77a-4759-781c-4aaff8851a8e%40posteo.de.
** THIS MESSAGE ORIGINATED OUTSIDE LOUGHBOROUGH UNIVERSITY **
** Be wary of links or attachments, especially if the email is unsolicited or you don't recognise the sender's email address. **
Thanks Mark, hi everyone
That all sounds very satisfactory. And perhaps some of the agencies covered will add open licenses and upload their code to GitHub or GitLab in due course. I know that at least one agency is currently formulating policies in that context but those discussions remain necessarily in‑house at present.
To add that these separation issues need to navigated with quite some care and insight. For those not aware, I have participated in the Free Software Foundation Europe (FSFE) Legal Network (LN) for six years and regularly mix with open source lawyers and technologists. Indeed I have a proposed presentation on whether the use of non‑open data models (like the IEC Common Information Model) can inadvertently transfer intellectual property (IP) protection to conformant codebases and datasets. More on that question in late April 2023 after I have run that thesis past real IP lawyers at the annual meeting of the LN.
To change tack, nice also to see recent cooperative use of software from the OSeMOSYS and PyPSA camps. I am not really at liberty to say more but someone from Simon Fraser University might like to comment?
kia ora, Robbie
PS: Mark, probably owe you another beer for being quite so blunt .. just a part of my Antipodean upbringing I guess?
To view this discussion on the web, visit https://groups.google.com/d/msgid/openmod-initiative/DB9PR04MB82351420226D0A34DD179D8BF8F99%40DB9PR04MB8235.eurprd04.prod.outlook.com.