#dev 2023-04-27
2023-04-27 UTC
geoffo, gxt__ and [schmarty] joined the channel
# 
[schmarty] just did my first POSSE copy of a big post to bluesky so looking forward to bridgy responses backfeed support there lol.
[snarfed] joined the channel
# 
[snarfed] ooh I think you just volunteered for https://github.com/snarfed/bridgy/issues/1453
{{lifeofpablo}}, [schmarty]1, [snarfed]1, [snarfed], [schmarty], gRegor, BinarySavior_, [schmarty]2, lockywolf, palle, [marksuth], [marksuth]1, [jeremycherfas] and [tantek] joined the channel
# 
[tantek] emoji << Issue: pitfalls to consider, especially if considering using emojis in IDs: 2023-04-26
{{schmarty}} [https://martymcgui.re/2023/04/26/bad-web-dev-ideas-emoji-as-ids-in-urls/ Bad web dev ideas: emoji as IDs in URLs]# 
Loqi ok, I added "Issue: pitfalls to consider, especially if considering using emojis in IDs: 2023-04-26
{{schmarty}} [https://martymcgui.re/2023/04/26/bad-web-dev-ideas-emoji-as-ids-in-urls/ Bad web dev ideas: emoji as IDs in URLs]" to the "See Also" section of /emoji https://indieweb.org/wiki/index.php?diff=87639&oldid=79242 # [schmarty] ^ 100%. As grumpy as I might be about language changes from php 7.4 => 8.x, in general the maintainers strive for stability. I update minor releases without fear.
# bkil tantek++ I fully agree with being cautious of backend hell. Hence why I encourage everyone to adapt the lightest feasible backend for their services possible (backend-optional even). Frontend quickfixes can be rolled out automatically and immediately (i.e., without bundling and just depending up to the minor version number of a given library, not to a patch level of it).
# [schmarty] The Churn is real
[Jo], lanodan and [KevinMarks] joined the channel
# 
[KevinMarks] I think it's the Jevons paradox. The dependency solution that npm built was so good that it encouraged use of more dependencies, reducing friction.
# [KevinMarks] That and the continuous release model that the web enabled over the previous annual stable release model that system software went with.
# [KevinMarks] Then we got aggressive deprecation of old versions, so the treadmill is real.
# [KevinMarks] Which was a response to the stability problem
# IWDiscordRelay <Lunarequest#0122> Hi, I've been considering making a test suit for local development/ci for webmentions, would people be interested in creating one?
holiday_medley and geoffo joined the channel
holiday_medley joined the channel
# IWDiscordRelay <Lunarequest#0122> I'm thinking of mostly creating a set of html templates along with a server that can send a set of webmentions to a server and the user can confirm if the right amount are rejected/accepted
# [schmarty] Lunarequest: you may want to check out https://webmention.rocks/ which is hosted, but open source (https://github.com/aaronpk/webmention.rocks)
# IWDiscordRelay <Lunarequest#0122> I personally was thing either python or rust of php since rust can run without any host deps and python is preinstalled on most linux distros
# [schmarty] You may find the test cases in there to be of use, independent of the programming language. :}
# IWDiscordRelay <Lunarequest#0122> php is not my strong suit would you mind pointing out where the test cases are?
# 
@rovenclasher Testing for webmention: https://rjitsu.github.io/writing.html (twitter.com/_/status/1651592992489103364)
# @rovenclasher testing for webmentions: https://rjitsu.github.io (twitter.com/_/status/1651595738772566017)
# bkil [tantek]: We still have a single ~3.6Mb/s 3G modem shared over wifi per carriage on most of our trains. It was top notch when deployed around 2013, but it's not considered very good today and 3G switch-off is also imminent. They started rolling out updates to 4G in 2021-2022 on premium class, but I don't think they are quite finished everywhere.
# bkil Just had a second look, some have been running since 2011 and their maximal link speed is 7.2Mb/s. Considering that both portable device usage and bandwidth requirements had only increased since then, I don't think that hundreds of people sharing ~300Mb/s LTE would allow websites to sit back.
# [tantek] I want this for webnotification reply notifications https://tantek.com/2011/239/t1/gmail-ui-feature-request-sort-smallest-first-email
# bkil Some of our buses also have it, although only about 40 people need to share it over there and they usually drive by closer to dense, covered areas.
# [tantek] a few posts about it here (click the next arrow to see the follow-ups) https://tantek.com/2011/239/t2/always-slow-connections-minimize-http-and-bytes
# bkil Interesting idea of sorting messages by length (word count even). We had a case with payed GitHub support and guess what: their reporting interface tells you that if your inquire is shorter than X characters (a hundred or two?) it will be "starred" and usually responded to faster!
# bkil Sorry if this comes through as a bit harsh, but having to go through multiple round trips of multiple kilobytes to read 100 byte of your posts at a time, also taking time to decipher when and where a reader needs to click sounds a bit wasteful. Yes, I know this is a cultural thing coming from Twitter, but still feels a bit considering the topic you have chosen.
# bkil At the same time, if you offered a side file that assembled such posts fragmented to threads or labels, a single kB of fetch could render the whole thing on screen.
pharonix71 joined the channel
# bkil I see you also serve a feed file from the wiki, but it would be nice to also mark this up via alternate links in your HTML https://indieweb.org/feed_file#Inefficient_Representation
# bkil You inline JS and stuff, so it's 2kB gzip'ped. Note that it's still considered very tiny by today's standards, just wanted to highlight the fact that there's always room for efficiency improvements. I also view round-trips & human processing much more processing than a few kB of extra transfer by the way.
# bkil more overhead
# bkil If you don't wish to mark it up as an alternate, consider just advertising its link in the header perhaps. Otherwise, if I wanted to subscribe to you, I didn't see where to look for (better browsers and addons give you a button if they found it on the given page viewed) - I had to visit the home page.
# bkil As I read most web pages through less, I pay close attention to details. There's a little inline CSS for the search form that is not too much. There are these `<indie-action>` tags that are unusable without JavaScript loaded, hence redundant. Then there's JS for select all inlined. Again, this few kB of cruft doesn't bother me at all, just wanted to elaborate now that you've asked.
# bkil I'm in the middle of transitioning of my publishing flow from side file-first to a custom HTML-first system and also working towards converging all of my content to a single site.
[kimberlyhirsh] joined the channel
# bkil I actually came here with a question today originally. I'm designing my favicon an wanted to know what all of your insights are about this. Already checked https://indieweb.org/icon https://indieweb.org/profile_photo
bret joined the channel
# bkil I planned to use it across sites as an avatar as well, but this is complicated by the fact that just within Element itself, the avatar is scaled and showed in various versions simultaneously: 14px, 16px, 30px, 36px, 325.2px and also supporting full size when clicking on it. And this is on a single platform! I'm sure it will process it to further sizes on Retina displays.
# bkil Surely HTML makes your job easier that you can specify a separate file for each size. In general, you can choose between: a big photograph, a small mug shot photo of your face, vector art of pixel art. This is especially tricky with pixel art I am working on. I could tweak a single file to be legible in 14px, 16px (desktop favicon) and 32px (favicon on HQ or avatar) at the same time, but the effort seems futile.
# bkil Also, when designing a favicon, do you consider that data consumers and renderers may be cutting corners so to speak (i.e., border radius)? If yes, to what extent? I strive to test with the worst case (encirclement) as some platforms are handling avatars that way.
# bkil Indeed. Interesting that I haven't seen an app or websites that could help you in designing your favicons & avatars with all the above criteria (or if one exists, it's not linked from the wiki).
# t0nic So I've got a completely flat file website, generated using an SSG; I'd like to add a newsletter as a functionality
# t0nic What I'd like to do is once a week, generate a template that includes all my posts from the last week, and send that out as a newsletter
# t0nic Anyone here do something similar already, or have recommendations for hwo to go about it?
# [snarfed] sounds a bit like https://github.com/indieweb/this-week
# bkil I've also answered my own question from last week. I made a simple site that is 1st on one SERP and 5th on another for two common words. The feat here is that I haven't done _any_ SEO activity at all other than producing semantic and tiny sources. I haven't even **linked to** the website using `<a href=...>` anywhere! I did place a clear text mention of it in UGC on a high PR local site. It seems certain crawlers do follow such references.
gRegor, angelo, motheki and gRegorLove_ joined the channel
[jacky] joined the channel
# [tantek] [snarfed]++ great post, nailed it: https://snarfed.org/2022-03-10_were-drowning-software-dependencies
gRegorLove_ and [Tim_Nolte] joined the channel
# [Tim_Nolte] [snarfed]++ yes, so true. https://snarfed.org/2022-03-10_were-drowning-software-dependencies
gRegorLove__, [lifeofpablo], pharonix71 and geoffo joined the channel
# [tantek] Some good thoughts about who gets to have replies to posts show up when/where in AP-like systems with some mitigations being proposed: https://queer.af/@erincandescent/110271938380622364