FOSS funding vanishes from EU's 2025 Horizon program plans | Hacker News Reader

FOSS funding vanishes from EU's 2025 Horizon program plans

theregister.com

80 pointsopenrisk21 comments

The open letter and sign-on opportunity can be found here: https://www.ow2.org/view/Events/The_European_Union_must_keep...

Does anyone know any good FOSS projects that were actually funded by this?

I feel like the people deciding what (research and technology) to fund in the EU generally have very poor taste in their selections. I assume that they are non-technical as they often seem to choose the projects that make the most ambitious claims about social impact and EU values, but have very little to none technical merit.

Stuff like Rust, that could be a real technological advantage for the EU, is not funded, but weird blockchain semantic web foo is https://www.ngi.eu/

NGI funded us to work on the GNU Name System (https://www.rfc-editor.org/rfc/rfc9498.html), that includes the Go implementation, updates to the C implementation, and the specification. Plus NGI funds work on https://taler.net/. Another example for a project they funded is https://cryptpad.fr/. https://nlnet.nl/project/ has an extensive list. If you actually care about privacy on the Internet and FLOSS, I'd be surprised if you did not find some project there that you know or use.

If you are working on improving privacy on the Internet with FLOSS, you can still apply for funding at https://nlnet.nl/propose.

One project that was funded that I can remember is https://stalw.art. It’s a very nice and compact e-mail server. The EU doesn’t take decisions directly about which projects to fund, but instead delegates it to more knowledgeable organizations. One such org is https://www.nlnetlabs.nl

I think the core of the question here is - do those funds bring any kind of return or are they one of thefriends of friends get handouts from government type projects that ultimately don't generate anything outside corruption and paperwork.

> The EU doesn’t take decisions directly about which projects to fund, but instead delegates it to more knowledgeable organizations.

They very much do. They invite volunteers to review applications and serve on juries for EIC and ERC, but the entire process, including definition of the funding criteria, is run by the EU.

You’re right, it’s a mix of both.

Note: It is not NLnet Labs, but NLnet [0] to which is delegated. They are not the same.

[0] https://nlnet.nl

Why would Rust create a technological advantage for the EU?

I don’t know why they said that. But I guess it’s the difference between a low certainty and a high certainty bet.

In 2024 it’s a safe assumption that a lot of the digital infra we use in 10 years will be based on Rust, so it may make sense to fund that.

On the other hand, something speculative like “semantic web” is less likely to be as important. It could be important, we just don’t know in $current_year.

I think it’s alright to split money between low certainty and high certainty bets, but reasonable people can disagree.

Oddly, this is partly by design.

EU SME funding programs have "non bankability" as a criterion (or variations on that, it changes every few years).

What that means is that if you have alternative sources of funding, even just potentially, you are disqualified.

What they're trying to do is 1) avoid competing with private funding and 2) fill in the gap where private funding is too risk averse but the project has huge potential.

In practice the result is as you say, this is basically a criterion to bias the selection to bad projects.

So I went through the 1090 entries on https://nlnet.nl/project/ and found 19 projects I've actually used:

Conversations.im https://nlnet.nl/project/Conversations/

Dolphin file manager https://nlnet.nl/project/DolphinAuth/

F-Droid https://nlnet.nl/project/Reproducible-F-Droid/

GPLv3 https://nlnet.nl/project/gpl3/

Jabber/XMPP https://nlnet.nl/project/xmpp/

Jitsi https://nlnet.nl/project/jitsi/

KDE Connect https://nlnet.nl/project/KDE-Connect/

LibreOffice https://nlnet.nl/project/LibreOffice-CRDT/

Marginalia Search https://nlnet.nl/project/Marginalia/

Mastodon https://nlnet.nl/project/Mastodon/

Matrix https://nlnet.nl/project/Matrix/

Nextcloud https://nlnet.nl/project/NextCloudSearch/

Nitter https://nlnet.nl/project/Nitter/

Oilshell/Oils for Unix https://nlnet.nl/project/Oils/

Okular https://nlnet.nl/project/Okular/

PulseAudio https://nlnet.nl/project/pulseaudio/

SeedVault https://nlnet.nl/project/Seedvault/

StreetComplete https://nlnet.nl/project/StreetComplete/

XWiki https://nlnet.nl/project/WikiActivityPub/

A few projects were funded multiple times for different features, but I've not necessarily used the particular feature that received funding, so let's say 2% of grants benefited me in some way. I suspect that's a different 2% for most people.

It doesn't look like they're averse to funding Rust projects (none of which I've used, like HN favorite Servo), so unless the Rust Foundation applied for a grant and was rejected, I don't think you can criticize NLnet for not funding them.

I periodically watch the status of Lanzaboote [1] to know when it will be upstreamed to Nixpkgs so that it will be easier for people (e.g. me) to switch to NixOS. It has been funded by the Next Generation Internet initiative that the article says is being eliminated.

[1] https://github.com/nix-community/lanzaboote

> Does anyone know any good FOSS projects that were actually funded by this?

SourceHut[1] comes to mind.

[1]: https://sourcehut.org/

bcachefs

There are many things yet to be done that don't fit as a product and are best served open source.

Short list of things that don't exist:

- universal datetime library for all major languages that can parse all formats, timezone aware, and is convenient to use.

- universal permissions and paths library. Filesystem, network filepaths, both for windows acls, Linux and Mac's extended perms. How can I know if I have read permission before reading? Or that I can traverse a directory before trying?

- memory safe openssl

- memory safe small footprint embeddable html/css renderer for universal GUI's and small arm computers. (not webview)

- USB c protocols for embedded devices.

- improvements to kicad to implement smart trace routing.

- any pain point or plumbing that really should be apart of a language/library already.

> memory safe openssl

This is closer to done than not done. Rustls is a memory safe TLS library that is compatible with OpenSSL in API and has comparable or better performance [2]. It has also passed security audits [3]

But your point stands. Rustls wouldn’t have been possible without open source funding.

[1] - https://www.memorysafety.org/initiative/rustls/

[2] - https://www.memorysafety.org/blog/rustls-performance/

[3] - https://github.com/rustls/rustls/blob/main/audit/TLS-01-repo...

This is the first I've heard of "Web 4.0," and the pairing of the terminology with "virtual worlds" is... interesting. Is the EU funding metaverse development in anticipation that it will be important (lol), or am I misinterpreting "virtual world" here?

The authors of such documents often inhabit virtual worlds, where the language is in different buzzwords than what we are used to in the real world.

Between this and the Cyber Resilience Act sheesh

> .. more than a third of NGI programs involve supporting compliance with the EU's GDPR law and its Cyber Resilience Act, while a further 23 percent are involved with implementation of Digital Services Act and Digital Markets Act rules.. some of it's important for helping the EU implement laws around digital goods and services. Without necessary funding, such implementations could be imperiled.

Funding cuts for CRA compliance and DSA/DMA enforcement?

Reply on news.ycombinator.com