Automatic Blurb Extraction

submitted by
edited

Meaning that a link post automatically extracts some meta info - every time it’s clicked, not only when it’s created. Because it appears that Lemmy doesn’t automatically do it once at posting time.

Examples:

https://piefed.social/c/mediatheque@feddit.org/p/2446966/la-la-land
https://piefed.zip/c/netzkultur@feddit.org/p/1874208/we-noticed-that-you-are-blocking-our-ads

I only see a link. When I view the original post on a Lemmy-based instance, I see an automatically, real-time extracted blurb.

Is pyfedi’s stance something like “it should be enough to do that once when you create the post”? But apparently that’s not common practice?

1
9

Log in to comment

Hot Top New Old
[any] edited

Is PieFed trying to pull in data on post creation? I think what people do is just post links and the platform can then go ahead and fetch a social media post and display it alongside, or fetch something like the Open Graph metadata of a news article and display the summary…

I haven’t looked at the code. Seems Lemmy fetches Mastodon toots. I remember this being discussed in the history of the Fediverse. Our instances already do lots of HTTP requests all day, so we should make it somewhat efficient. Idk if doing additional requests on each view would be great. On the other hand fetching it immediately on post creation isn’t great either, because we end up in a situation where we constantly DDoS websites because each time anyone posts a link, that website is going to be bombarded by requests by hundreds or thousands of Fedi servers… Idk where we’re at with that discussion… mainly remember this wasn’t as straightforward as it seems.

Would be useful to have those “blurbs” though. Both for cross-posts and other stuff.

 reply
1
Insert image

These files are not very private - piefed.nullspace.lol admins may review them occasionally.