Random instead of incremental post IDs

[eiclu]

For front end issues, use lemmy-ui

Is your proposal related to a problem?

Right now Post IDs start at one and will be simply counted up. This has some drawbacks:

• It is possible to scrape a whole Lemmy instance by simply iterating all IDs
• The amount of content on an instance is instantly apparent, some people hosting Lemmy may not want this
• It is impossible to have link-private posts, where they won't appear in the feed/search but you can still access them by an ID

Describe the solution you'd like

Multiple possibilities:

• Use the title as ID (issues when editing the title)
• UUID Strings (leads to pretty long URLs)
• base64 Strings (seems to be the best option. The length could be configurable)

[dessalines]

https://hashids.org/ is what you're looking for there, which we might potentially add at some point for the urls. But overall the scraping / iterative concerns aren't too big a deal to me, because everything is publicly available right now. Once we implement proper private communities, then access will be denied to even those who know a URL.

[Nutomic]

This is actually a duplicate of https://github.com/LemmyNet/lemmy/issues/1101 (both are talking about the same IDs).

[Nutomic]

Reopening this as its the general issue, #1101 is only about the activitypub part (which is really a followup to this issue).

[dullbananas]

Link-private posts can be done using a random string that must be included in the URL to see the post