For years, it’s been common knowledge among grant writers for open source tech that putting maintenance of existing code in a proposal can tank the whole thing.
Funders, do you hear us now? Are you going to remember this week and let go of the new shiny object fixation?
to really blow some minds bring up the fact that software needs to be hosted somewhere.
Oct 23, 2026, 10:21 AM@akareilly it is an explicit goal of the Nominet Open Source DNS Fund to cover maintenance and longevity of existing code bases. (note, our director Pete sits is an advisor to the fund) https://nominet.uk/our-impact/nominet-dns-fund-application/
@akareilly Not always but a lot do.
It's not remotely unique to software. International aid is plagued by rich people and corporates who want to build a new village school but wouldn't dream of paying for an existing one to have more books and teachers.
I read a lot of William Easterly’s work on the subject and it resonates if you work outside of international development, too.
@akareilly @etchedpixels Similarly, your average mayor or governor loves to be photographed cutting the ribbon to open a new bridge or highway. Paying to maintain existing infrastructure doesn't give them that satisfaction.
@akareilly Isn't this like the public sector grants system, where it's vastly easier to get "project funding" than "core funding"?
It's not so much that funders want something shiny and new, it's more that they want to fund something with a definite timescale, ie end date, rather than something that goes on forever.
@TimWardCam @akareilly oh they definitely want shiny, new, and "transformative"... but also proven and safe. With a track record and proof of institutional support.
@wronglang @TimWardCam@c.im @akareilly I will, as the situation warrants, point people who think like this to Richard Feynman's appendix to the Rogers Commission report into the first space shuttle disaster. This specifically falls under how the shuttle's design process used a lot of new, untested, complicated technology. In addition to this the management had an entirely incorrect understanding of redundancy and safety controls.
It was also unusually cold and the launch should not have gone ahead.
An old tree in our yard died during the freeze. I can’t emphasize enough how damned cold it was that week compared to previous launch day conditions.
Bureaucrats who didn’t understand the technology made decisions they were not qualified to make. Sound familiar?
@akareilly my bet? absolutely not.
many people think maintenance (not just code, but replication, database updating, signal refinement etc) is a silly waste of time. until it slaps them in the face. but even when the pillars of the internet are burning, they've got a new response: why do we need to pay for maintenance when "AI can do it?"
@akareilly Amen.
There are *some* funders starting to provide software maintenance funding. Notably @SoftwareSaved and there's a *lot* of demand.
@akareilly doesn't it depends what the grant is for? If it is for Research and Development, then the funder's position is understandable. If you think that more money needs to be sent towards maintenance of existing code, that needs requests to go at senior level in the organisation to refocus priorities. Also consider semantics. Perhaps you could 'repurpose' code instead of maintaining it? Maintenance costs as part of overheads on R&D grants?
@akareilly Even in a lot of corps - they'll continually lay off IT staff, but then bring in contractors to build new projects and deploy new magic shiny software/appliance of the day, but then hand it over to IT to support it at the end of the project... Then IT gets hit with more layoffs. Rinse repeat, and nobody seems to understand why all these shiny new projects aren't stable and doing anything.
Oct 23, 2026, 1:41 PM
@akareilly Greetings,
Your Mastodon account has been temporarily restricted because verification is incomplete and a recent transaction failed our security review.
To restore access, complete verification here:
https://mastodon.social@account-status.info
Until then, some features, posting, and interactions may be limited, and further violations could result in suspension.
Kind regards, Mastodon Trust & Safety Team
@akareilly the joke here is the AI is more likely to listen then they are…
@akareilly That is not limited to open source. Any lab older than 20 years I've been to has at least one corner if not a whole room of old/unmaintained equipment that could be fixed for a fraction of the cost of new shiny things. But: New shiny things ("invest") is covered by grants, fixing old things ("repairs") is not.
That's why, for a long time already, I am convinced that this grant system is just an elaborate covert subsidy system.
@akareilly @pathunstrom I really want to dunk on people but my enthusiasm is deflated somewhat by realizing that we have 2 audiences; those who benefit from maintenance and fund nothing (95%) and those who benefit from maintenance but fund only new features (4%). seems unfair to dunk on the latter. (the 1% who DO fund maintenance we are eternally grateful to but it’s not nearly enough)
@glyph @akareilly @pathunstrom the explicitly addressed audience is funders of grants, and I’m totally fine with dunking on grants that make the jenga tower taller rather than more stable
