#dev 2024-05-29

2024-05-29 UTC
geoffo joined the channel
#
[tantek]
πŸ‘€
gerben and AramZS joined the channel
#
[KevinMarks]
Maybe it was just my account he disabled
#
[tantek]
whoa what
jacky joined the channel
#
jacky
perfect place for something like this to exist (with websub as a place to experiment) https://rachelbythebay.com/w/2024/05/27/feed/
alephalpha0, GWG and gRegorLove_ joined the channel
#
[Jo]
I didnt know browsers used to show you the sites feed automatically :0
#
claudinec
one of the things we have lost from the old web (though I think there are extensions for various browsers that will display feeds)
#
superkuh
It still seems crazy to me that Mozilla removed HTML feed header indications from Firefox.
#
superkuh
I guess whatever is not needed for shopping or watching netflix had to go.
#
superkuh
pcarrier, server side includes allow for just enough templating power with html fragments to make everything very easy but also easy to mantain and secure.
#
superkuh
I know they sound ancient but w/nginx and many other webservers they're first class citizens still.
#
superkuh
And old does not mean bad.
#
pcarrier[d]
was this meant for me?
jacky joined the channel
#
superkuh
The bit about ssi/server side includes, yes.
#
superkuh
"<pcarrier[d]> something that would allow for "CMSes" without eg arbitrary PHP for all visitors"
nertzy joined the channel
#
pcarrier[d]
aaah, yes, thanks
[Ros] joined the channel
#
[Ros]
Pinboard: nooooo! I only just discovered Maciej because [snarfed] mentioned applying to go to the last XOXOFest a week ago, then I discovered the XOXOFest and the treasure trove of talked, and listened to a bunch of them including https://xoxofest.com/2013/videos/maciej-ceglowski/ which is was hilarious in. And only just discovered Pinboard that I thought looked great and wanted to use later 😒 I am the latest bloomer in this IndieWorld. I
#
[Ros]
only just got psyched adding hover element in CSS this week
#
[Paul_Robert_Ll]
Fun fact: `:hover` was the gateway drug to my becoming a web developer. Can still remember the day I found the underlying CSS that was making links change colour on hover on the BBC News website waaaay back in 2001 or so. πŸ‘΄
AramZS joined the channel
#
capjamesg[d]
Wow! That's amazing!
geoffo and jacky joined the channel
#
Zegnat
[Ros]: if you like Maciej talks, https://archive.org/details/dConstruct2013MaciejCeglowski is by far my favourite
#
[Paul_Robert_Ll]
He has some great take downs on the lunacy (pun intended) of NASA’s Moon and Mars projects, too:
dz4k joined the channel
#
dz4k
how much of a problem is it if my mf2 parser doesn's support legacy microformats?
[schmarty] joined the channel
#
[schmarty]
dz4k: depends on what all you want to parse. a lot of old wordpress themes have "mf1" microformats
#
[Joe_Crawford]
There's no Maciej talk that's not worth watching. Web bloat. Toxic data. Great stuff!
#
[Joe_Crawford]
Reading old code always gets me musing. https://artlung.com/blog/2024/05/29/the-web-is-better/
#
[Ros]
Thanks guys! I’ve burnt myself out on 90 min non-progress on css centering so this is just what I need before I try again
#
capjamesg[d]
(posting here since the post talks about making one's own blog engine)
#
Loqi
[preview] [Jeena] 20 Years of Blogging On my own website
barnaby joined the channel
#
[tantek]
new idea for a minimum interesting (and storage impact) service worker: the mirror router
#
[tantek]
basic outline: you have a service worker on your personal site, installed from any page on your site, that all it does for future requests to your site is check to see if the page is available and serves it, otherwise it displays the "site appears to be unreachable" type message that a lot of service workers provide, AND provides an algorithmically constructed link to the page on a static mirror of your site (typically at another domain),
#
[tantek]
where the copies there use rel=canonical to link back to the original obviously.
#
[tantek]
that way if someone has ever visited your personal site, in the future, whenever they click a link to your pages / posts, if your site/domain etc. is unavailable for ANY reason, then they see a notice and a link to a static mirror to view instead, thus providing a soft-repair to any otherwise apparently broken permalinks
#
[tantek]
you can either deploy that static mirror yourself with a lazy SSG to a GH repo to a domain
#
[tantek]
OR you could also (even more minimal) generate a link to the internet archive copy of a page for folks to click through to view
#
[tantek]
your service worker* could generate a link to the internet archive copy (these are algorithmically constructable from a permalink path)
#
[snarfed]
interesting! I like it
#
[tantek]
now I'm wondering if I can write this as a generic (rather than site-specific) service worker than literally anyone with a personal site could "install" as is (as in one .js file, and one link rel in page head or HTTP headers) and it would figure everything out from the context it is running in, unchanged (zero config)
#
[tantek]
ideally you want rel=canonical on your pages to your domain so browsers / user agents / add-ons could offer to try to view the canonical live version of the page (when viewing an Internet Archive copy)
#
[tantek]
(or a static mirror you setup somewhere else)
#
aaronpk
id love to test that with a static copy of a site
#
[tantek]
this is what happens when I brainstorm with [Joe_Crawford] during espresso live streaming
#
[tantek]
aaronpk, same here. it's also another use-case for SSGs, literally SSG for backup viewing, e.g. for creating a static mirror on GH, so you don't have to worry about "SSG build time" because there's an expectation of lag in mirrors which does not apply to wanting your own site immediately up to date as soon as you click "post"
#
[Joe_Crawford]
It's great stuff! It also makes me think about markup or other conventions in markup to denote that, say, a comment form, would not work. Like `if-page-static-disabled` or something. I like this idea. Folks use internet archive like this, or a CDN... I'm not even sure Google exposes cached versions of pages anymore.
#
aaronpk
Makes me think about what it would look like to make a static version of my site. Technically it's huge because of all the tag pages and pagination, but maybe those aren't important for this use case and I can focus on just permalinks instead
#
aaronpk
and then maybe even just the permalinks of articles/photos/notes rather than all the posts like food and travel
#
aaronpk
So maybe I could build it externally, like spiderpig, where it crawls specific feeds and archives only certain posts, rather than trying to somehow add an SSG build process into my actual website source code
#
[tantek]
right, step 1 is automating sending "only certain posts" to the Internet Archive as the ultimate fallback / mirror
#
[tantek]
and then you can decide how much "nicer" a mirror you want in comparison, trading off with the effort/maintenance to do so
#
[aciccarello]
I've limited the number of tag page variations I have on my site since they can explode the number of pages in my build
#
[aciccarello]
For example, I don't have a way to view tags filtered by post type
#
[aciccarello]
And I've considered moving to date base pagination so a new post doesn't change every page in the pagination
#
[tantek]
that needs a meme aciccarello
#
[tantek]
βœ‹ database pagination; πŸ‘‰ date-based pagination
chxshire, jeremycherfas and Xe joined the channel
#
[aciccarello]
Some day I'll implement that. It does feel strange to have variable length list pages
barnaby and daiyi[d] joined the channel