Approved video domain list?

Tags: #<Tag:0x00007f0fa3ac62c8>


@trashHeap and I are posting video links in #scitech, from #fosdem! It is great fun, and we can watch the video and comment on the same page, this is kinda what web appy stuff is supposed to do, be useful, ne? :slight_smile:

Discourse has a whitelist/blacklist feature for "onebox"ing, that is the little app that scraps websites and embeds the info in the post.

I’m not entirely sure if it works as I expect BTW. I’m putting that out front, in case my understanding is completely different from what is possible. :slight_smile:

Anyhow, I’d like to encourage more people to watch and talk. But that means inviting folks to share links from all over. And not just direct video files, but the container pages/URLs and all the noise that brings along.

As it is, I block YouTube domains. It means the links get shared, but it doesn’t embed the YouTube player in the page. That was set years ago, but only that one site. In the past we just don’t share a lot of video.

Now I’m thinking about prohibiting everything by default, and adding embeds on a case-by-case basis.

Move slow and fix things.

When we visit the web interface for talkgroup I expect it to not betray us. I want it to ever scrub the input in smart, automated ways. I don’t want other internet sites tracking us on each page load.

Let’s look at this from the administrivia-site: we will need folks to ask to approve their domain before the website renders the video player on the site. Wow, that’s all, and that’s not bad.

I mean, I was expecting to play devil’s advocate, but that really isn’t that bad. It may even be preferable. But I am open to all feedback.

How I envision it is that someone has a domain where they or some froody group hosts a bunch of videos. So they drop some links, we investigate the creepy factor, and then add it (probably).

To start I’d add things like, Wikipedia stuff, our personal domains, things like that. Thoughts?


I also wonder how this changes when we start getting into PeerTube. Torrenting has inherent networking info sharing that can be invasive, but is also an efficient way to distribute video among us non-corporate masses…

Will I be able to embed a webtorrent video file hosted on PeerTube site into a web page view of Discourse? It’s like, difficult to track what is happening during that transaction, in my head.


Wikimedia, and personal domains I htink cover like 90% of my use cases I think. Of course or a specific agreed upon mirror like the OSUOSL site would be good at the moment.

For all the hype of Peertube; im personally passing on peertube at the moment. I like the idea of distributing this content via bittorrent to scale bandwidth and related issues; but the whole rest of the peertube scene im not sure is for me. The server software looks far more complex and heavy than I typically like to host or maintain. Im not sure duplicating youtube UI and comment culture are goals im invested in; etc.

MediaGoblin is an alternative in that space that is working on ActivityPub federation too worth keeping an eye on; but again; a lot of complexity (to me anyway) on the back end.

Im currently thinking I kinda prefer embedding bog standard web video in things like discourse or elsewhere for the time being.

Archiving/hosting AV

I may have misunderstood the list features, and may not be able to prevent generating the video links when a file is directly linked to… hmmm. Here is my concern: when I visit ActivityPub panel, FOSDEM 2019 I see a video box with a screencap/thumbnail for the video.

I don’t really know what data is going back and forth, what those other servers are learning about visitors to this site, etc.

It may be that as a public data store with a web interface, talkgroup/Discourse will always straddle the line concerning web safety. On the other hand, maybe a separate client would be useful, a user agent accessing the data in talkgroup via the API… :thinking:


Talkgroup is supposed to have semi radical transparency, no?

Is there anything which could be learned from this, that also wouldnt also be learned by scraping the site?


Anonymous site visitors. I’ll think on that, though.


Yes! :slight_smile: