Conference proceedings are the collected written record of what was presented at an event, published as a volume with an identifier attached. They are standard in academic conferences, where the paper rather than the talk is the artefact, and they are almost entirely absent from technical conferences, which produce something else instead.
What proceedings actually are
A published volume containing the papers accepted for a conference, usually appearing at or shortly after the event, with an ISBN or DOI and a permanent home. The paper is peer reviewed before acceptance, the talk is a presentation of it, and the citable artefact is the paper.
That ordering is the whole difference. At an academic conference the talk exists to present the paper. At a technical conference the talk is the thing itself, and anything written is a by-product if it exists at all.
Why technical conferences do not produce them
Because nothing in the ecosystem needs them. Nobody’s career depends on a citation count, there is no tenure committee counting publications, and the audience wants to see how somebody thinks rather than to read a formal account of it.
Writing a paper also takes weeks that a working engineer does not have, and requiring one would change who is able to speak. The call for papers at a technical event asks for a proposal precisely because it is not asking for a paper, which is set out on the call for speakers.
What technical conferences have instead
A recording, usually published within weeks. Slides, sometimes. Occasionally a write-up by an attendee, which is the closest thing to proceedings the ecosystem produces and is entirely voluntary.
A video on somebody else’s platform is not an archive.
Reading the chart, this is more accessible and considerably less durable. A recording is easier to consume than a paper and far easier to lose: it is not citable in any stable way, and it lives wherever it was uploaded for as long as that platform chooses to keep it.
The durability problem
The talks of the small conferences of the late 2000s are substantially gone. Some were never recorded, some were recorded onto hosting that has since closed, and some sit on a video platform under an account nobody maintains. What survives is patchy and mostly accidental.
That is a real loss and it is the argument for the least glamorous thing an event can do: keep its own copy of the recordings and the slides, on storage it controls, with the programme alongside them. It costs almost nothing and it is what the events documented in the regional conference index would need to have done for their record to still exist.
The middle ground some events found
A few technical conferences publish something between the two: a written companion to each talk, usually a page or two, produced by the speaker after the event rather than before it. It is not peer reviewed and it is not citable, and it solves the two problems that matter most in practice.
The first is that it is searchable, which a recording is not. Somebody looking for the talk that described a particular approach can find a page of text and cannot find forty minutes of video. The second is that it survives: a page of text is trivial to keep and to move, where a video file is neither.
The cost is that it asks the speaker for another afternoon after the event, which is the point at which most people have moved on. Events that get it consistently are the ones that ask for it as part of the acceptance rather than as a favour afterwards.
What it is confused with
A programme. The list of what was scheduled, not the content of it. Proceedings contain the work; a programme contains the timetable.
A call for papers. The invitation rather than the result, and at technical events it is a call for proposals wearing an academic name.
An archive. Proceedings are a published volume; an archive is whatever an event kept. Most technical conferences have the second and none of the first.
The rest is in the glossary.
Questions
What are conference proceedings?
The collected written record of an event, published as a volume with an identifier and a permanent home. The papers are peer reviewed before acceptance, and the paper rather than the talk is the citable artefact.
Why do technical conferences not publish them?
Because nothing needs them. No career depends on a citation count, the audience wants to see how somebody thinks rather than read a formal account, and requiring a paper would take weeks a working engineer does not have, which would change who can speak.
What do technical conferences produce instead?
A recording, usually within weeks, sometimes slides, and occasionally a write-up by an attendee. More accessible than a paper and considerably less durable, because none of it is citable and all of it depends on a platform staying up.
What happened to the talks of older conferences?
Substantially gone. Some were never recorded, some went onto hosting that has closed, and some sit under accounts nobody maintains. What survives is patchy and mostly accidental.
What should an event do about it?
Keep its own copy of the recordings and slides on storage it controls, with the programme alongside them. It costs almost nothing, and it is exactly what the events whose records have disappeared would have needed to do.