Linux: no money for documentation

Linux has been an open source flagship project for more than 30 years. It’s also worth noting that the leader is still there – in his original role: Linus Torvalds releases the Linux kernel and has no say in what happens to it. In many ways, this project is different from most other projects in IT. This includes the development model itself, the number of employees, the financing of the work and of course the duration.

Advertisement

Unfortunately, there is also a problem that often occurs in the project world: the importance and evaluation of documentation.

Cry for help and resignation

Shortly before the EU Open Source Summit in Vienna in September 2024, a small earthquake struck the region. Alejandro Colomer Resigned as chief administrator of Linux Man Pages. The reason for this is simple: he can no longer afford this voluntary role. In other words: either a sponsor can be found to maintain the support pages or Alejandro will devote his time and effort to tasks that require payment.

This problem is not new, as Linux Weekly News (LWN) editor Jonathan Corbett has previously reported in april 2018 After that. “Nobody wants to pay to write a document,” he said. Corbett went even further. In his opinion, the quality of kernel documentation is not as good as it should be and better financial support is needed. Corbett is no longer an ordinary name in the world of Linux. still wait his cry for help Probably unheard of.

WebAssembly, WASI and Rust: Dream Team for Microservices

At the Open Source Summit EU 2024, the topic came up again in a panel discussion – but only as a secondary aspect: it was really about the community of kernel developers, administrators of various subsystems, and ultimately the Linux world. Looks like a post-Linus Torvalds era. After all, Finn has been there from the beginning and still has the final say when it comes to the kernel code.

Interestingly, Torvalds has had several proxies over more than 30 years. Well-known names include Alan Cox, Andrew Morton and, these days, Greg Kroah-Hartman. Representatives came and went – ​​but the project manager stayed. One possible creation of the post-Linus Torvalds era could be the replacement of the “benevolent dictator for life” (BDFL) by a kind of supervisory board. So a group of people who have the final say on the kernel code.

Then there could be implemented a type of editorial team with an attached editor-in-chief. It’s not that kernel developers don’t write documentation. In such a huge project, a uniform and consistent consumer sentiment on the user side is not a sure success. An editorial framework greatly expands the possibilities. But it also requires healthy financing.

known vulnerabilities

The editorial approach also solves another problem that has raised its ugly head more than once in the open source world: the dependence of an important project on one person. Famous cases include OpenSSL with Stefan Hansen, GnuPG with Werner Koch, and NTP with Harlan Stein. But that’s not all, because here also the work was done on a voluntary basis. Income from donations can be safely ignored. There were clear calls for help in the mid-2010s, and the Linux Foundation responded. Previously there was the Core Infrastructure Initiative, which has now been replaced by the Open Source Security Foundation (OpenSSF). It financially supports the mentioned software projects and now others too.

A similar solution may also work in the Linux documentation area. The first step would be to appoint Alejandro Colomer as a paid administrator of the help pages, followed by the installation of a deputy. This way you can gradually reach an editorial structure. The whole thing can also be integrated into the Linux Foundation using the LF documentation. However, the subjective perception during the Open Source Summit EU 2024 was that not enough people with decision-making power think that Linux documentation is a serious problem. This means that it is less likely to be dealt with in a timely manner – but hopefully that subjective perception is misleading.

(For)