[kevinmarks]Kaushalmodi: modelling as classes seems like overkill - they're fields really. Showing the parsed json is usually clearer than more abstract diagrams. If the audience is Hugo devs, then showing the toml -> HTML -> mf2 json may make sense.
kaushalmodikevinmarks: Thanks for you remarks. I agree that the class representation is an overkill. I'd still like to somehow represent the organization in a diagram. I'll keep exploring on the best way to do that. The blog post will also go into detail on how to add mf2 to existing Hugo themes.
vanderven.se martijnedited /posts () "(-1858) Limit talk of “kinds”, stick to “types”. Offload Inferring Post Type to [post-type-discovery]. h/t gerwitz" (view diff)
aaronpk[cleverdevil]: grantcodes: here's what RTs look like in monocle right now. i'm on the fence about this display though https://i.imgur.com/ySXSd3s.jpg
aaronpkalso microsub and the clients need support for that additional data of who reposted something so that the original post can be shown with its original author
snarfedwe support people hosting at all levels up and down the stack here! domain and data portability are the key starting points, not "own all the things"
skippysure, in some capacity you're relying on someone esle. but if you develop a toolchain using AWS Lamba, Amazon can change the API and deprecate whatever they want, whenever they want. Then instead of complainign about Twitter dropping 3rd party apps, we'll all complain about AWS mucking up our workflows.
snarfedskippy does have a point, there is some vendor lock in, but honestly less than many people think. eg app engine has a ton of lock in, but lambda and cloud functions (and that direction in general) less, since the interface is so small and clear
aaronpkif you use AWS properly, you're tying together their proprietary products. you end up with a great result, and it be far more effective than running on a VPS, but that's not exactly portable elsewhere
snarfedalways tradeoffs. lock in can be balanced by added value. if it's valuable enough, it's worh it anyway, even including the potential future cost of migrating
bearas long as the data flow of indieweb is implementation agnostic, then we can have multiple implementations that both prevent lock-in and also keep the ecosystem vibrant
bearyou can't get me to even look at Bluemix - their UI is horrible and the integration of their services on the backend has been a dumpster fire for months
sknebelIf you can't even tell me what resources I have and how that fits in what I am supposed to have I'm not going to trust them with the ability to bill me
[miklb]bear yes, the backend UI is horrible. I was playing with it for a hackathon for a non-profit. We were using Watson Assist, so I used their node sdk to run the bot.
[cleverdevil]One thing I was tinkering with before was using S3 to store all of my microformats JSON data and then using S3 Select or Athena to query it for my website.
[miklb]I’ve invested a lot of time into WP, both personally and professionally. I don’t see it going away, but I’m not sure I want to invest anymore personal time there. I kinda regret switching from Jekyll to be honest.
skippyHugo is interesting, but hard to grok in its own way. I want to do things that arent obviously easy. And adding all my tweets into my content has tanked rendering times; so I'm now reconsidering that decision.
[miklb]understood. But honestly, with how I have learned to tune my WP site using redis/static_cache and other performance tweaking (offloading images to s3, google pagespeed) I’m pretty pleased with it being dynamic site.
[miklb]but I believe that was the then janky way of checking/sending webmentions. Aaron rewrote the webmention plugin so I think it would be much better now.
kaushalmodiskippy: About the 20 seconds build time, you can wrap that in a conditional so that the tweet pages build only on server, and not on your local machine where you are writing new content
[miklb]my issue with build times was in using micropub to publish. I’d write a note in Quill, and have to wait for it to build/deploy before showing up on twitter.
Loqih-feed is a microformats2 draft specification with a top level feed object to contain h-entry posts and optionally a common author https://indieweb.org/h-feed
aaronpknot sure why, but maybe because of the leading sentence "These standards are interoperably implemented by the overwhelming majority of..." I was assuming that list was ordered by how widely implemented the specs are
jalcinelike if there's proposed ideas - to what extent have been fleshed out and some reference docs (probably back to the wiki and companion posts) about it
LoqiA content management system (often abbreviated as CMS) is software used to create, enter, edit, update, delete content on a website, even on indieweb sites https://indieweb.org/CMS
aaronpkoh the other slight difference between caniuse and our equivalent is that on caniuse, the expectation is that every browser will eventually support the same features. whereas with ours there might be a reader project that intentionally has no publishing aspect but would still support a bunch of indieweb specs.
aaronpkyeah like visually on the site, caniuse shows a bunch of red for features that have low support, or you'll see IE with a bunch of red. I wouldn't want Monocle to show up with a bunch of red on Webmention support because Webmention doesn't apply to Monocle
tantekI think jalcine is on the right track, something with the look and feel of caniuse would really help people (devs?) "get" the level of support in various tools
snarfedaaronpk: just fyi i'm trying a global (reactive) throttle for instagram scraping in bridgy and granary, since they're still massively rate limited. i can let you know how it goes if you're interested.
LoqiIt looks like we don't have a page for "eleventy" yet. Would you like to create it? (Or just say "eleventy is ____", a sentence describing the term)
@TwitterDevLast year we announced our plan to retire Site Streams & User Streams, and replace them with the Account Activity API (currently in beta). We are delaying the scheduled June 19th deprecation date. (twitter.com/_/status/982346370882461696)