#LoqiText-first design refers to the practice of designing information so that it is usable/actionable in its most basic plaintext form https://indieweb.org/text_first_design
sl007, tantek, j12t, wolftune, nitot, KevinMarks, miklb and friedcell joined the channel
rickygee, jeremycherfas and Kopfstein joined the channel
#ZegnatHmm, I need to find a way to own my YouTube history. Any ideas?
#ZegnatGoogle’s own watch history misses the one important data point: when I watched it. Why?!
Pierre-O, leg, hs0ucy and jeremycherfas joined the channel
#jaduncan[m]Zegnat: if you're fetching it often enough you can just pick up the video names and you know the time slot in any case.
#jaduncan[m]It depends how granular you want it to be, but the difference between the old and new responses provides you with a rough time.
mlncn joined the channel
#ZegnatI am currently just doubting if it actually even includes everything
jeremycherfas and Pierre-O joined the channel
#jaduncan[m]The other way to include everything would be to have a userscript that added each video you watch to a special favourites list, and an import script that then removed them.
mlncn and jeremycherfas_ joined the channel
#ZegnatYeah, I was thinking of just rolling up a WebExtension to keep track. Probably a lot easier than trying to scrape the youtube history page. But it will not be able to catch stuff I watch outside the browser.
friedcell and jeremycherfas_ joined the channel
#Zegnatacegiak, you were doing YouTube logging, right? What are you using?
#Zegnat!tell acegiak you were doing YouTube logging, right? What are you using?
#sknebelsebsel: more that you share your reputation with other users on the platform (which admittedly also could be good)
#sebselyeah that's right! the good/bad content discussion is not really indieweb relevant, but having bad content of other on the same domain as your (good?) content is a good reason for indieweb / your own domain
#aaronpk"Google was able to block it, because each querying machine broadcast a user agent that contained the words "WordPress pingback," which Google engineers promptly blocked."
mlncn, acegiak and begriffs joined the channel
#bearI worry about that style of attack being aimed at webmention
#aaronpkthe good(?) news is that they first tried a bunch of other unrelated attacks. pingback was not their first attempt.
#bearsure, but the distributed nature of webmention and the fact that most folks using it run on tiny servers means that any attack, even partially successful, is going to hurt a *lot* of people
#bearone of the many things on my todo list: create a list of attacks to try against webmention and show how it either already prevents them or how they can be mitigated
#jaduncan[m]Bear: I would imagine that the defender would promptly filter on the webmention headers. It would however maybe be good to ratelimit webmentions on the plugins etc. If something that isn't on a whitelist is (or appears to be) sending 5 webmentions a second, it's probably good to cut that out on the respondent end. Maybe even to just require the user to authorise continued mention display.
#jaduncan[m]At least that way the DoS is on webmentions rather than tying up every resource on the server and DoSing the whole site.
#bearjaduncan[m] yep - that's the attack vector that worries me the most - inbound webmention processing as a DOS
#bearI need to get my nginx config cleaned up so I can post how I use it to rate limit my dynmaic endpoints
#bearsadly today, instead of fun coding, I am working on terraform and packer configurations for work
#GWGI have rate limits on my XMLRPC endpoint, but not Webmentions yet
#jaduncan[m]bear: Yes. I was idly wondering about getting around bans via spoofing. There's also an issue of spam when webmentions scale, which probably means extra calls out to a spam detection service. The other thing I've idly wondered about since reading the text is just proof of work from the sender; it means you would at least cut out the ability to do viable attacks without a botnet.
#GWGbear, I am sitting here waiting for them to finish working on my car
#jaduncan[m]Absent that, you're getting the server to write the webmention comment to the page, probably call Akismet or similar, and then carry on. That's a lot of cost for the server when compared to the cost of the attack.
wolftune joined the channel
#bearthat type of spam handling I think is already covered by the vouch handling
#bear(which is to say that you could have as part of your vouch handling a callout to akismet or the like)
#jaduncan[m]Ah, OK. I've only just started looking, sorry. Does this mean that the attacker can currently also use webmentions against many servers to take down the main vouch server?
#bearpart of the learning process is asking questions
#bearand part of the learning process for other members of the community is learning where FAQs are answered
#beara big part of IndieWeb is the guiding principle that things are implemented when they are needed - so while thought experiments are encouraged and useful, they are secondary to experience
#sknebelthe pingback attack referenced in the article is the other direction, isn't it? using verification traffic?
#jaduncan[m]Eh heh. I salute your effective creation of a proof of work for the spammer.
#KevinMarksYes, it is. They sent the ping backs to lots of wp sites so that they would fetch from the server to check for the link
leg joined the channel
#jaduncan[m]bear: Vouch does indeed seem quite good. So for DoS prevention you could just have a strictness percentage on the servers that activated Vouch partially based on ratelimiting in suspicious situations before stopping and just logging/creating a rule to temp ban webmentions for increasing timeouts. Maybe even based on server load.
#bearyes, the benefit with webmention+vouch is that the normal attack surface mitigations can be used
#bearright now the issue is how to communicate that to a group of people who do not want to be ops :)
#jaduncan[m]Is this something you're planning to work on in any particular plugin or CMS? I'm looking at how much I can harden Known at the moment, since I want to use it for a blog whilst I'm cycling Africa and won't be able to sysadmin from moment to moment (the best of of a few OKish options, given that I'm heavily suspicious of WP security).
Lana joined the channel
#bearjaduncan[m] I probably won't as I don't use any CMS - my site is static with a dynamic part just for webmentions
#bearI do plan on working on some general guidelines, just nothing specific to a CMS
#[dgold]bear Yeah, I've been following eliot's tutorial - that's gotten me to a blank screeen
#bear[dgold] sorry then :/ you may want to ask in the known irc channel
#Loqi[Sebastiaan Andeweg] Inzichten op de late avond: het probleem is niet dat ik niet op Facebook zit. Het probleem is dat jullie er allemaal wel op zitten, en dat normaal vinden.
En, min of meer door Tegenlicht van vanavond: zwart-wit denken los je niet op met grijstinten....