Looking back at 2023

Our first prediction was that the community would see a spike in AI-generated material submitted as contributions. If that has indeed happened, the quality has been high enough that those submissions have not been identified as coming from machine-learning applications. That is perhaps surprising, given the common reports that large numbers of developers are working with machine-learning assistance at this point. Be that as it may, it seems fair to say that the flood we predicted has not, as yet, arrived.

The same cannot be said for submissions to LWN. We have, for many years, gotten offers of "articles" to publish, many of which come with promise of payment for publication (an offer we have never accepted in all of LWN's history, incidentally). Often it has been clear that not much human attention has gone into the creation of these offers. Over the last year, though, it seems that some of these people have figured out how to ask a language model to generate their emails, with the result that their solicitations are simultaneously more formulaic and less intelligible than ever. The flood has arrived, but it is going for the (seemingly) easier money first.

The prediction that new kernel functionality would be written in Rust and proposed for inclusion was reasonably obvious a year ago. Much of what has actually been proposed to go upstream is a reimplementation of existing functionality (the Rust Binder implementation, for example), which doesn't quite satisfy the prediction. Other work, though, such as PuzzleFS and the Apple M1 graphics driver, definitely does. As of this writing, a network PHY driver written in Rust, along with some supporting abstractions, is staged to be merged into 6.8, presumably shortly after the new year. As was predicted, no user-visible functionality written in Rust was actually merged in 2023.

The prediction of a "make-or-break year for distributed social networking" is somewhat hard to verify. Certainly, the distributed network often called (not entirely accurately) "Mastodon", has grown significantly over the last year and is the focus of a great deal of conversation. This network, it seems, will be around for the long term. Meanwhile, though, sites like the one formerly known as Twitter continue to exist, despite what appears to be a determined effort to kill it. How this story will play out overall is yet to be seen, but there is hope that centralized services may be giving way to federated alternatives based on free software.

Was it "the year of the immutable distribution" as predicted? The effort being put into the development of such distributions appears to have only increased over the last year (see our recent review of Project Bluefin for an example) , but it is fair to say that this approach to distributions has not yet taken over in general. Distributions built on SUSE's "ALP" platform are not yet shipping (though openSUSE MicroOS Desktop did make its debut). Instead, one could be forgiven for thinking that the commercial distributors out there have given up on innovation and decided to just make RHEL clones instead.

What was missed

While erroneous predictions are easy to point out, predictions that were not made at all can be a bit harder to identify. Nonetheless, one can certainly pick out some events that, in retrospect, we should have perhaps seen coming.

Red Hat's escalation of the enterprise-distribution wars is arguably one of those. The company has never made a secret of its intent to protect the Red Hat Enterprise Linux cash cow to the extent that it can, perhaps bending (though generally not breaking) the free-software licensing rules in the process. The company's decision in 2023 to stop making its patches generally available — and to reserve the right to cut off customers who distribute those patches — fits the pattern well.

Perhaps Red Hat's decision to drop LibreOffice was a bit more of a legitimate surprise; there was once a time when the availability of that sort of "productivity" software was deemed crucial for the future of Linux. In 2023, to the extent that office suites are still considered important, some distributors seem to feel that they can let others deal with packaging and shipping them — or just expect people to use centralized, online services instead.

LWN did not predict the end of six-year long-term-support kernels; once somebody starts providing a service, it is easy to imagine that this service will continue indefinitely. But if it seems that there are few users of that service, and even fewer who are willing to help keep that service going, it will eventually go away. The kernels that receive commercial long-term support are generally far removed from the mainline, so they benefit relatively little from many years of mainline support. Those of us using something close to mainline kernels are, as a general rule, better served by running relatively recent releases. In that world, there is not much call for six years of support for a mainline kernel.

After years of seeing BPF pushed into all corners of the kernel, it was only a matter of time until an attempt was made to put it into the CPU scheduler. That doesn't make it any easier to predict where BPF will end up next year, though.

While it would have been difficult to predict HashiCorp's license change specifically, it would not have been that hard to guess that some company, somewhere, would decide that this whole "open-source software" thing wasn't actually for them. As long as there are companies working with our software, there will be companies trying to figure out how to get the benefits of open source without actually having to free their code. We can't predict who will try this next year, but some company almost certainly will.

Another event that might have been foreseen was disruptions in the community resulting from the ongoing Russian invasion of Ukraine. Indeed, the surprising thing may be that the impact from such a major war has, thus far, been so small. Our community is good at staying above the political and economic fray, but there are limits. Depending on how events proceed in the more troubled parts of our world, we may see more turbulence in the coming year.

Surprises can come closer to home as well — and can also be more welcome. We first wrote about our desire to escape from proprietary accounting software in 2007. After 15 years (at the beginning of this year), it would have been hard to predict that this goal would be achieved in 2023, but it did indeed happen. We are now approaching the end of our first full year using GnuCash, and it continues to go well. We can easily predict that we will not be going back to proprietary solutions anytime soon.

In summary

While the free-software community tends to hold up well in difficult times, it is not entirely immune to those difficulties. In this case, the fierce competition to hire some types of developer has eased up to an extent, making lucrative positions harder to find, and some jobs have been lost. Opportunities remain, though; free software is sufficiently important to businesses that they can only hold off on hiring for so long.

Our community was hurt this year by the loss of Bram Moolenaar, Abraham Raji, and Satoru Ueda, among others; they will all be sorely missed.

Free software remains strong, in that much of the technology industry (and beyond) relies heavily upon it; as a result, there is more software being written and released than ever. It is still true, though, that many of the problems we have seen over the years — failure to live up to licensing requirements, unwillingness to support developers and maintainers, communities that do not feel welcoming to all, etc. — remain with us, and are not necessarily getting better. It is not surprising that people get discouraged and drop out, or never settle into our community in the first place.

If one looks, though, the spirit of what we set out to do remains alive and well. The growth of the ActivityPub network (Mastodon and such) in 2023 is a good example; a decentralized network based on free software has taken up much of the role that was once filled by a large, centralized, proprietary service. Applications like Home Assistant are restoring control over the systems installed in homes to the people who live in those homes. Open-source large-language models are refusing to cede that space to large, proprietary players. And the community as a whole continues to produce more software, some of which is high-quality indeed.

By now the world has seen enough examples of malign technologies, like printers that refuse to print without a subscription payment, "smart" devices that leak personal data far and wide, cars that brick themselves on command, and doorbells that spy on the neighborhood, to understand that control over the software in one's life is important. Our community has spent decades working toward a world where that kind of control is possible and, to a great extent, we have succeeded. Maybe, in 2023, we have also succeeded, in some little way, in showing the world why this control is important. If that trend can accelerate in 2024, then this year may yet look like an important inflection point with long-reaching consequences.

Over the course of this year, LWN published 292 feature articles (41 of which were from guest authors) and reported from 13 conferences. We have done our best to follow the ups and downs of our community and, in return, the community has supported us. As we head into the end-of-year holidays, we wish the best to all of our readers; thanks for sticking with us this year, and we look forward to seeing you in the next.