#Royaljellohow many people really want to see an indiedev style internet? I've got a website, and I'm working on it at the moment so I want to ask: do people want to make a website indie web compatible
#RoyaljelloI don't want to, because it requires a lot of self sabotage
#[Rose]!tell aaronpk is there a special trick to get Overland to compile? Because I couldn't get it to do so yesterday on the plane, files are missing.
#sebselRoyaljello: can you explain what you mean by ‘self sabotage’?
#sebselAlso: the IndieWeb is really just a way to own your identity and content online. You need your own domain, maybe blog on it (or maybe not!). You don’t need to be ‘compatible’ if by that you mean implement all the building blocks
#sebselall those are optional, only there to help you if that is what you want
#Royaljellosebsel: I want to make a website that's pretty much indie web/blockchain, but in order to attract an audience (or if you're plain lazy), I've used Twitch for streaming video content
#RoyaljelloAh! then it's pretty much already indie web
#RoyaljelloWhoops. Just used to having to describe it
#RoyaljelloIt's pretty much solely an alternate to main web, using old methods and channels
#RoyaljelloBut I kind of want to incorporate the works. There just isn't much of a community yet internationally to do that though
#sebselyes, the blockchain is a bit out of context for the IndieWeb, but having a website as primary identity and using other sites to syndicate to / attract audience sounds IndieWeb to me
#Royaljelloisn't much of a community yet to justify it*
#RoyaljelloYeah. It's pretty much "the" hub for anything not centralised
#[Rose]Well, using it yourself is always an important step 🙂
#sebselso, yea, IndieWeb as described by the IndieWeb community on indieweb.org does not really try to create a new web, just using the ‘old web’ / stuff we already have.
#sebselfeel free to explore alternatives, but I personally like the approach to start with what works now :)
#[Rose]Starting with what we have now that works lowers admintax and barrier to entry usualy
#Loqiblockchain is a technology in search of a problem to solve, pushed by ideology into areas where the unsolved problems aren’t technological https://indieweb.org/blockchain
#sebseloops. that sounds a bit negative. but also points out that it’s really a different thing than the Web.
#[tantek]“Sans all social media fixation” sounds perfectly fine 🙌
#[tantek]Also please don’t confuse legitimate criticism with snark JeremeyCherfas
#[tantek]Blockchain has numerous actual serious harmful problems, some of which are explicitly mentioned upfront in the dfn to help steer people away from wasting time at best and creating (more) harmful things at worst
#[tantek]Being up front critical about things that are bad isn’t “being negative” it’s being honest and truthful
#sebselI was just trying to avoid saying “the IndieWeb is against blockchain”. But a lot of us are sceptical at least.
#@pvhAnyone of any skill level is welcome provided you're working on something with local-first/offline-first features. No slides or talks, but hopefully folks will demo things they're working on. Lastly, whatever your interests are, this is not the venue to discuss cryptocurrencies. (twitter.com/_/status/1181596431615811585)
#[tantek]“Cryptocurrencies” is a bit of a conflation but they do tend to both all be blockchain based and a prime implication of when people say blockchain
[Lewis_Cowles] joined the channel
#[Lewis_Cowles]GWG int10h… Amazing that someone captured that. I might get int21h lol
#[Lewis_Cowles]I must admit to having a post which mentions blockchain and is negative, but I think a lot of the negativity is around the hype-train surrounding tech
#[Lewis_Cowles]it’s not blockchain and I call that out. It’s the notion that tech is a magic hat, which defies physics or rationality I think a lot of people are against.
#[Lewis_Cowles]that loqi quote seems opinionated though
#[Lewis_Cowles]I also cannot get int21h because it’s taken… Probably a good thing
[qubyte] and jeremych_ joined the channel
#jeremycherfasNot entirely IndieWeb question: can one use, eg symlinks, to ensure that .bash_profile is the same on two different machines?
#jeremycherfasOr is there a different method you would recommend?
#[Lewis_Cowles]two machines unless mounting network won’t be able to resolve the symlink without it’s host file
#Loqiaaronpk: [Rose] left you a message 2 hours, 41 minutes ago: is there a special trick to get Overland to compile? Because I couldn't get it to do so yesterday on the plane, files are missing.
#[Lewis_Cowles]instead I’d either use dotfiles (search github many people maintain their own) or something lightweight like a gist to download and overwrite them
#aaronpk[Rose]: yes it uses a package manager thingy to install some dependencies. I forget what it's called. Pod maybe?
#jeremycherfasThanks [Lewis_Cowles] It was seeing an article about dotfiles that made me think that I've been thinking about this issue in a desultory way for a while.
#[Lewis_Cowles]when I upgraded catalina, I chose to `cp ~/.bashrc ~/.zshrc` rather than a symlink so that as differences cropped up, they did not alter my history of what worked / works in bash
#[Rose][aaronpk] That sounds about right, I'll have to figure that out. I did start adding a Help page which I wanted to link from the bottom of settings to explain things inside the app
#[tantek]I want someone to try out the “how to” instructions I added to /offline and improve it with feedback while the firsthand experience is still fresh in my head from IWC Brighton
#[Lewis_Cowles]TBH the only things I ever add for a serviceworker are blocked route paths
#aaronpknah i probably won't get to it for a while
#aaronpkalso i've already written an offline app once so proabbly better for someone who isn't at all familiar with the offline stuff to try the instructions
#[tantek]These should be opinionated for a personal site, based on real needs not extrapolated
#[Lewis_Cowles]If they are overly specific instructions, you might as well make a generic serviceworker people can use and modify
#[Lewis_Cowles]How would you envision a personal site service-worker as different to the times online
#[Lewis_Cowles]I think some offline tech would be different, but a serviceworker, would remain similar, perhaps avoiding push notifications etc, which none of our examples handle
[grantcodes] joined the channel
#[grantcodes]My site sort of theoretically works offline. If you install it as a pwa it should download the majority of the database too. In practice I don't think it works perfectly 😅
#[tantek]since my data store is static HTML, I'm considering loading the most recent storage file into the client, and then having JS that can actually render all the posts from that storage file locally
#[Sonny]with ServiceWorker you don't need all that, you can fetch and cache programmatically an URL
#[Sonny]and the browser will render that if offline
#[grantcodes]Mine is progressive enhancement too since I server render the js site 🙂 it's really nice to have tools with that built in
#[Sonny]what I mean is that there's no need to use JS to render anything, all you need is a service worker that tells the browser to cache url a,b,c,d,...
[chrisbergr] joined the channel
#[chrisbergr][grantcodes] is there anything affecting seo with this behavior?
#[tantek]Sonny, it's more efficient both in terms of network requests and in terms of # of bytes to only cache *one* static HTML storage file for 61 days worth of posts than to fetch *every single permalink* in those 61 days
#[tantek]in fact it may be so much more efficient that it is likely I could cache a full year's worth of storage files (just 6) than even a week's worth of posts as permalinks
#[Sonny]ah yes if your data store is HTML fragment then yes good point
#[tantek]right, if I was storing one HTML file per post then yes it wouldn't make much sense
#[Sonny]I'd be curious to see how a all-in-one solution works out - crawling the website for urls, fetch them, deduplicate html and cache
#[tantek]Sonny, the background here is that I've written most of my code in CASSIS, deliberately so it can run and assemble pages either serverside in PHP or clientside in JS.
#ZegnatWhat is the max cache size of a service worker? Honestly I find it a little weird that I may push an entire website into your local memory just because you visit my homepage once.
#LoqiIt looks like we don't have a page for "max cache size of a service worker" yet. Would you like to create it? (Or just say "max cache size of a service worker is ____", a sentence describing the term)
#ZegnatI think [tantek] is onto something with not assuming people want his caches when they come in straight to a permalink
#sebselI can still add the feature to /offline still, ofcourse. But it feels wrong to be in control of how much data I push to the user, even if they gave permission for notifications (because I won’t show any in this pattern)
#[tantek]hmm ok. would be great to get that documented somewhere as a how to or maybe brainstorming since it's not live
#[tantek]it's like a podcast app, except with less data
#[tantek]that's a great idea and it applies to *any* website, not just readers, feeds etc.
[jgmac1106] joined the channel
#sebselI have no idea how that works, but I’ll be in this train for another while, so I’ll look into it :)
gxt joined the channel
#[tantek]specifically a feature request for Firefox UI that provides a list of offline cached websites to browse
#[tantek]as perhaps an enhancement to the existing "Work Offline" menu item in the "File" menu
#ZegnatI had never noticed that Work Offline menu, and my default browser is Firefox Nightly. Thankful you used that one in your demo, [tantek]!
#[tantek]heck one minimal way to do it would be to provide a list of the N most "recent" offline cached websites as a *submenu* of that Work Offline menu item, with perhaps an item at the end like "Show All Offline Sites"
#[tantek]Zegnat, "Work Offline" has been in Firefox like forever! 🙂
#[tantek]It also loads pages from the cache that would otherwise force a network request
#[tantek]since you've specifically asked it to do so by choosing that menu item
#[grantcodes]I only do my database download if you explicitly install the pwa. If you just visit the site it does less full on caching
#Zegnat[grantcodes], that seems fair. I just don’t know how much I actually would be comfortable with. It seems to be one of those weird questions of what a visitor expects you to be able to do, and how much you should be doing that they do not expect.
#[grantcodes]Well my expectation is for literally noone apart from me to seriously install my website 😛 maybe some indieweb people to see if it works and then delete it later
#[tantek]ok three level approach is what I'm thinking (of which I do nearly two)
#@beckycodes↩️ Less effort solution: Commento. Forget about that in my webmentions excitement. See what you’re saying about Twitter, but the JSON response is mine to do whatevs with forever I guess (twitter.com/_/status/1186732066991460352)
#[tantek]1 visit a permalink from a search engine -> only install service worker and cache offline page, plus cache-as-you-go for *future* pageloads
#[tantek]2 visit home page -> install service worker, cache offline page, home page, contact & pay pages, plus cache-as-you-go
#ZegnatI do wonder if the cache-as-you-go should have some sort of freshness rating for caches, so you do not run up the storage?
#[tantek]3 install PWA -> pre-cache entire site storage files (or most recent storage files up to 5MB max)
#[tantek]Zegnat or even better, just a limit so that when a new page is cached, the oldest is thrown out
#[tantek]this may require better image optimization for my photos
#[tantek]or maybe to give them their own cache with its own limits
#ZegnatYes, [tantek], freshness can be anything I guess. Either time-of-visited (as in, FIFO) but you could maybe try something more fancy like trying to prioritise dropping other things before you drop your latest X posts if people are more likely to return for those.
#ZegnatI do think the idea of considering your user’s data limits as well as storage is important. Because you can do something (e.g. cache 50MB) does not mean you should :)
#[tantek]I'm thinking 2MB of text/css/js plus 3MB photos max for cache-as-you-go
#[tantek]time of visited makes more sense from the user perspective
gRegorLove and [qubyte] joined the channel
#[qubyte]A lot of my site hinges, oddly, on the CSS. It all gets bundled and the content hashed to form part of the file name. The browser is instructed to cache it immutably. The knock on effect is that the CSS changing effectively invalidates all HTML files which reference it. Right now I have a pretty basic service worker based on one a while back by Jeremy Keith, but I could rework it to spot CSS changes and clear out/refresh everything using old
#[qubyte]Then again, that could get network heavy and might be considered bad behaviour.
jbove joined the channel
#[qubyte]Perhaps I could keep N versions of the CSS or use different versioning like semver.
#[qubyte]I suppose another interesting option (not sure if it’s possible but I think it is) is to perform surgery on outdated cached HTML to fit the new CSS file name into them and update their etags.
#Loqi[Jeremy Keith] Paris Web 2019 - 10 octobre après-midi - Amphithéâtre - YouTube
Here’s the livestream of the talk I gave at Paris Web—Going Offline, complete with French live-captioning and simultaneous interpretation in .
#[qubyte]Hmm. Definitely something I should look into. It seems to be a little too clever though. Very use case specific and probably fragile (if possible).
#Zegnat[qubyte], the better solution may be to not do it streaming, but to grab the full cached resource as a string and work on that. Then you have the option to run it through DOMParser, which should make it a lot more stable. That could have some sort of performance impact of course.
#LoqiIt looks like we don't have a page for "silent push" yet. Would you like to create it? (Or just say "silent push is ____", a sentence describing the term)
#[tantek][sebsel] ^^^ that's the term for the technique you were doing in Berlin