#dev 2021-12-26

2021-12-26 UTC
Eddy04[d] and jeremycherfas joined the channel
#
GWG
How is a page this long classified as a stub? https://indieweb.org/Webmention-developer
#
GWG
And https://indieweb.org/webmention-spec seems a bit dated as well.
edburns[d] and maxwelljoslyn[d] joined the channel
#
sknebel
because nobody thought of removing it
#
sknebel
happens quite a bit
tetov-irc, neceve, Christian_Olivie, [jacky], KartikPrabhu, chenghiz_ and jjuran joined the channel
#
GWG
Why do /Webmention and /Webmention-developer both have an implementations/examples section? Isn't that duplicative?
#
petermolnar
probably. I'd say go ahead, clean it up
#
GWG
Just wondering if there is something I'm missing
#
GWG
The page had a todo about generations, so definitely dated
neceve, IntriguedWow[d], razinsyahmi[d], Help_My_Doge_Pls and omz13 joined the channel
#
petermolnar
!tell capjamesg[d] did you see https://yacy.net/ when you started to work on your search stuff?
#
Loqi
Ok, I'll tell them that when I see them next
#
capjamesg[d]
I did but I didn’t look into it that much.
#
capjamesg[d]
Is there anything in particular that interests you petermolnar?
omz13 joined the channel
#
petermolnar
just curious :)
#
capjamesg[d]
I love the idea of a decentralized search engine.
#
capjamesg[d]
Although I think what interests me most is decoupling ranking with indexing and search results representation.
#
capjamesg[d]
Ranking is the hardest one.
#
capjamesg[d]
I haven’t learned how to build reverse indexes that are fast yet.
#
capjamesg[d]
I got stuck the last time so took a break.
#
capjamesg[d]
I have an inverted index but I could not figure out how to add boosts to make a ranking algorithm. Doing so does however greatly interest me.
#
capjamesg[d]
The trouble with elasticsearch is that it is very abstract. That’s good for most use cases but I want to know how everything key works.
omz13 and gRegor joined the channel
#
petermolnar
ES is probably the heaviest thing you can throw at this problem
#
petermolnar
doesn't necessarily the best though
#
Loqi
[prabhatsharma] zinc: Zinc Search engine. A lightweight alternative to elasticsearch that requires minimal resources, written in Go.
#
petermolnar
the p2p search idea is quite an interesting one. What do you (this channel) think of a p2p search engine system where each domain brings their own index of their own content?
#
Zegnat
The hard part is mixing the indeces in a logical way. I did a meta-search-engine thing at IWC Utrecht where I used WordPress APIs to run a search on multiple people’s blogs and then combined the outputs.
#
GWG
Zegnat: And there are new APIs
omz13_ and KartikPrabhu joined the channel
#
capjamesg[d]
petermolnar I thought about this from the perspective of allowing individual sites to send indexes of their own content to a centralized server.
#
capjamesg[d]
But that would not be too scalable due to demands on resources.
#
capjamesg[d]
p2p is interesting, can you explain more about what you have in mind? I don’t have a good understanding of p2p network architectures
#
petermolnar
neither do I, I just came across yacy, and haven't started looking to be honest.
#
capjamesg[d]
Has anyone thought about auth for static sites?
#
capjamesg[d]
What is htaccess?
#
Loqi
htaccess is the file used for configuring particular sites on apache (as opposed to httpd.conf which is global) https://indieweb.org/htaccess
#
capjamesg[d]
Maybe there could be a third service that verifies you with indieauth and gives you a password to a page / set of pages.
#
capjamesg[d]
Then you could revoke that password at any time so that users access would be revoked.
#
capjamesg[d]
This is contingent upon htaccess allowing multiple passwords for a file. That might not be possible.
#
Zegnat
I always feel like "auth for static site" is a bit hard to define. When do you consider something no longer static? I think https://www.authelia.com/ integrates with nginx, so you can have your server redirect the user to a separately hosted auth provider. But then that might actually be more stuff for you to host than having a static site that also supports PHP includes for a login
#
PK[d]
capjamesg[d]: Indexes as in the URLs on that site?
jamietanna joined the channel
#
capjamesg[d]
PK[d] all of the content on the web page.
#
PK[d]
Hmm.. Oh yes that’s right..
#
PK[d]
I wonder if it’d be easier to create an index of marked up websites
#
[jacky]
nice work! digging the 'random' otpion
#
[jacky]
*option
#
capjamesg[d]
Thank you 🙂
#
capjamesg[d]
Truthfully I haven’t worked on it for a little bit as I have been doing stuff with my own site. My dream is to run the search index too rather than using elastic search.
#
capjamesg[d]
Every page is a h-feed and comes with a JSON / jf2 feed too.
#
capjamesg[d]
No rss just yet as I haven’t modeled system resource usage for that.
#
[jacky]
you could use granary to stub the RSS part until you implement
#
[jacky]
how do you get to the alternative feeds?
#
[jacky]
also https://indieweb-search.jamesg.blog/about seems to use white text when in dark mode
#
[jacky]
(hope I'm not overwhelming!)
omz13, rototo, Murray[d], tetov-irc and [tw2113_Slack_] joined the channel