#GWGEmma2: No, thank you. You arranged for a venue.
sandro and [kevinmarks] joined the channel
#[kevinmarks]You can also use mention-tech.appspot.com to send them
#Emma2hmm ok so im trying to make a new page for the RSVP and it isn't working and ive notice that none of the text is green and blue, as it is on other pages that work
#cmalbasically it uses noapi to drain info from Twitter and then formats it as either json or html
#cmalnote that the JSON is neither mf2 nor AS2 compliant, it's the noapi format although they're probably going to use something standard in the future
#cmalsnarfed: @alct is going to rewrite social_stats using noapi I think we just need to focus on wrapping noapi properly (which I'm not really doing atm)
#cmalyup, the only difference here is we're just scrapping html from silos so paging is sort of difficult
snarfed joined the channel
#aaronpkI just added a simple view counter to my posts to track the number of times they've been viewed. I realized i didn't have that info anywhere other than google analytics, and it would at least be nice to know which of my posts are most viewed even if it's not super detailed stats.
#tommorrisaaronpk: at some point I might do that. just have it so each view leads to an event on a queue that'll be rolled into occasional DB updates.
loicm joined the channel
#aaronpkeh i just update the DB on every view, old-skool
gRegorLove joined the channel
#tommorrisaaronpk: I'd rather not give malicious actors a way to DDOS my database.
#aaronpki'm sure there are other ways they can already do that :)
#snarfedworrying about avoiding ddoses kind of reminds me of nutrition and trying to eat healthy. i suspect in practice the cost of the effort and stress far outweighs any actual benefit or harm prevented
#snarfedat least, for the project/site size we have here :P
#aaronpkthere are some pretty basic nutrition things you can do that actually do make a difference tho, likely analogous to some basic site architecture things you can do to prevent trivial DOSs
#aaronpkbut that sounds like a conversation for a different channel
#aaronpki use background queues for stuff where the tradeoff is in user experience. like if i make a post on my site and then i want to send webmentions, posse to multiple locations, etc, if the form I submit had to wait until all that was done before I saw the post it would be terrible. so I put all that on a background queue.
#snarfedoh definitely! UX and latency count. just mostly a different thing.
#aaronpkwell that was fun. Just fixed a hilarious bug. since I added support for multi-photo over the weekend, I discovered that a bunch of my bookmark posts had the screenshot of the site like 20-30 times in the post. this meant the pages were showing like 30 photos in the post and bringing my browser to a crawl.
#aaronpkturns out my cron job i set up at indieweb summit to make the screenshots forgot to check if it had already posted the screenshot, so they just kept stacking up
#snarfedAPI ideology is fine and all, and APIs do change too, but way less often, and with better warning and migration plans :P
#cmalagreed, but it also requires to be logged in therefore it's easier to get banned if you do stuff Twitter doesn't like
#cmalso this solution sounds perfect to me for people who'd like to add "social walls" to their websites while loading everything from there
#cmalthen how they get the data from the silo is their problem : can be a script switching VPNs to make it more random, could be a DHT-based pool of knowledge, for all I know :)
#cmalso I think although it's hacky and can break down sometimes, it's ultimately the most portable and anonymizing way to fetch data off silo that I know of
#snarfedyou can use the API without your users logging in, you just use an app-wide token
#snarfedand getting banned does happen, but way less often than people think
#cmalI mean better would be to have millions of users letting a silo app or a browser addon share in p2p the *public* content they receive, but we're not quite there yet :)
#snarfed(i know; i've had one app get "banned" from one silo, and many many more doing more "stuff they don't like" survive fine)
#snarfedin practice, the silos don't notice you or care until you get somewhat big, e.g. >10k active users
#cmalI don't know, call me whatever you like, but I don't like the idea of relying on the good faith of any kind of corporation or state
#snarfed(friends who've worked at four different silos have confirmed this)
#cmalnot that I'm afraid of getting banned, I just prefer never to rely on them
#snarfedbut that's my point, noapi started scraping from the beginning, regardless of the margin
#snarfedas an example, bridgy did have to switch to scraping instagram since they turned off much of its API for all apps, not just bridgy specifically...but we did that only when we had to
#snarfedand it's a pain, and we're very happy we still get to use all the other APIs
#cmalwell what you say makes sense in the case of a central hub a lot of people rely on (like bridgy), but if we're talking about a really decentralized setup the pros and cons of both sort of balance each other (having to register your app regularly to Facebook Known-style is really troublesome)
#aaronpkcmal: btw have you been able to sign in to the wiki? you should add yourself to https://indieweb.org/irc-people so your picture and website show up in the logs!
#cmalyup I have a userpage already and everything, I'll do this :)
#cmaljust atm my blog is broken (wrongfully edited a file on prod and not dev :()
#cmalalso what about ActivityPub, are some people actually working on implementing that as a "replacement" for webmentions + mf2 ?
#cmalbecause I think it's just brilliant, a PUBSUB architecture for the social web would make interconnection with XMPP-based social networks way easier
#aaronpkthere's some work on that happening in the w3c social group, but not a lot of implementations yet.
#bearyes, I do get what you mean and it does feel that way
#bearmy reason for stating that is, in my view, incremental implementation progress that leads to standards is soooo much saner than what has happened in the past
smcgregor joined the channel
#aaronpkindeed, "standards" which are made up before any implementation (especially before any interoperable implementations) are likely to turn out to be irrelevant in the long run