A documentation how a capture agent communicates with the core system and what data is needed and exchanged and which APIs and methods could be used. This should be useable to provide alternative Capture Agents.
This should be done in a subsection of the Developer Wiki in a probably new section under the Developer Documentation:
Sorry if my comment seemed offensive to you, Rubén. I read it twice again and I don't know why you have the impression - at least this was not my intention. I was only explaining the purpose of this ticket.
I wanted to encourage you to start the discussion on updates to the CA API, as I know that many others think that it needs updates. And after a first discussion in Oxford nobody really took the initiative. And I would really welcome a more generic interface here. So if this seemed ironic to you - it should not have been. I would really appreciate if you would moderate such a discussion.
After I finally noticed Gregs hint with the existing documentation, I still feel the need that this documentation should be improved. But I reduced the points. And Rubén if you want to earn these points you are more than welcome. The purpose of the Test Fest is to improve the documentation too. And if you probably have the knowledge to update this part of the docs, please do it next week.
I am simply trying to prepare this event...
So sorry for any misunderstandings
No worries. It seems I have misinterpreted your words, so please accept my apologies.
There's no point to discuss here --we all agree that this is not the best place to talk about new features to the CA. I'm honored you considered me a good candidate to moderate such a discussion (even after my not-so-moderate reaction), but I'll seriously consider that idea, since I believe such a discussion will lead to a well-defined standard, which does not only benefit the community, the developers and the adopters, but also the vendors of 3rd-party solutions, who will certainly be grateful for having a solid ground to step on.
To go back to the ticket itself, yes imho what we need is an a technical extension to what Greg wrote. His original page is a good starter but it doesn't go in the depth of the transactional API calls (which endpoints? what order?), the details of the "zipped mediapackage" methodology, etc, an information flow diagram would be nice.
Basically imagine you're a hardware vendor trying to integrate your capture appliance with MH and don't want to reverse engineer MH. Also this document should serve as the standard reference for such cases and whichever Capture Agent software doesn't follow the doc will be flagged as such by the community.
Added an overview of the request you should do, fixed some smalleissues, added missing single request ingest method.
Aaand, this is why I'm glad I'm reviewing the docs. Thanks for the changes Lars, I'm going to be fixing the misunderstandings shortly