[xhkkffbf]

It's not nonsense, but it dates from an earlier era and we may not want to continue in the tradition. In the past, professors who wanted to preserve a report of their work sent drafts off to publishers who reviewed, edited, typeset, printed and distributed. There are still plenty of people in this chain who do the work and aren't paid by the government. They get their paychecks from the subscription revenues to the journals. These are still, usually, paid indirectly by the government from the so-called "overhead" charged on grants. You can think of this system as "reader pays".

As more professors developed the ability to typeset their own work with LaTeX and email/ftp/http, the costs of the old system have started to become more difficult to sustain. Some researchers do the typesetting themselves but the richer ones hire their own production team to produce better LaTeX reports. You can think of this model as "writer pays." Either with time or money.

I think both sides have strong points. The old system is expensive and slow, but it does produce a very well-curated record of what happened. The new system is generally cheaper, but only because the researchers handle much more of the workload. I also worry that research will disappear when people retire or move on. A professor's web page may be really convenient, but they tend to disappear or die from bitrot after time. I also think the "writer pays" model tends to encourage over-producing some basic research notes that might not be ready to publish yet.