#dev 2022-01-18

2022-01-18 UTC
typ1cal_c0ffxe3, tetov-irc, frittro, rommudoh[m], diegov and justJustache joined the channel
#
jacky
heh implementing paging via Micropub is a bit tricker than I thought
#
jacky
doing it via SQL is oddly confusing
#
aaronpk
i assume you mean microsub
#
aaronpk
i used random strings in the examples, but the trick is you can use whatever you want, including for example, the integer primary key
#
aaronpk
or if you always sort everything by date, you can use a timestamp
solo joined the channel
#
jacky
oh no, I meant micropub lol
typ1cal_c0ffxe3 and darkkirb joined the channel
#
jacky
and support the `after` + `before` fields
#
aaronpk
same idea tho since it references the microsub spec paging :)
#
jacky
true! lol
#
jacky
probably overkill
frittro_ joined the channel
#
aaronpk
skimming thru that it looks like exactly the right thing to be reading :)
#
jacky
perfect
typ1cal_c0ffxe3 joined the channel
#
jacky
hm okay after reading this more, I think this works for me
#
jacky
leaning on `published_at` might be tricky since I can make that empty/null
#
jacky
but using something like `created_at` could be tricky (I plan to backfeed a lot of stuff from my archives)
#
jacky
I _could_ set the `created_at` to be the date of publishing from entries in my archives
#
GWG
I had similar paging trouble and ended up just doing offset for Micropub until I figured it out.
#
jacky
how did you figure it out?
#
jacky
or was it mainly trial and error for a bit?
#
GWG
I didn't yet
#
GWG
I just implemented offset and figured I'd come back to it later
#
jacky
ahh okay
KartikPrabhu joined the channel
#
GWG
I have this issue that I have the WordPress paging system to deal with, or not.
#
jacky
yeah so leaning on `published_at` breaks the potential cursor based approach I have
#
jacky
like I wouldn't be able to pull out things like drafts
#
jacky
but it doesn't necessarily have to
#
jacky
tries to think on this more
#
jacky
hm I could switch up the cursor for non-published posts to be based on their IDs
#
jacky
only other unique thing there I have
#
jacky
really avoiding adding a `created_at` column
#
[tantek]
what's wrong with keeping track of dt-created?
#
[tantek]
it might be one of the few sources of an actual unique id for active posts
#
jacky
chronologically, sure
#
jacky
but I haven't had an explicit need to publish it (nor can I think of a consuming case - maybe for an archive?)
#
jacky
I _techincally_ have it via the creation dates of the files I write to disk but those aren't indexed (obvs)
#
aaronpk
the micropub client-to-server communication is all private between the user and their own server, so it's ok to use data that isn't published
typ1cal_c0ffxe3 joined the channel
#
jacky
I think I can swap to created_at when post_status == draft and use it otherwise
jacky and gRegor joined the channel
#
aaronpk
This is a good bit of insight "Problem is, I think users tend to think “globally.” They expect all activity within a space or context to be visible to them. If a search for some user can return nothing because the subject is outside your local social graph, then your technology is less useful than other search models. This is why we tend to pool around well-maintained global indexes to give reliable
#
aaronpk
results, but then you have centralized authority emerging." via https://paulfrazee.medium.com/smart-contracts-without-blockchains-fc54603df754
#
jacky
oh I've been wanting to check out that video
#
jacky
the one re: Vitra
angelo joined the channel
#
jacky
I really wish the hypercore folks cared a bit more about language support
#
jacky
if they had bindings/implementations in things like Python or even in C (lol) and binded to other languages, it'd be easier to tinker on devices
ben_thatmustbeme and angelo joined the channel
#
[tantek]
awww still on Medium 😕
#
[tantek]
aaronpk, that email is false with one MASSIVE area: messaging
#
[tantek]
hah typo, that *insight* is false
#
[tantek]
"Problem is, I think users tend to think “globally.” They expect all activity within a space or context to be visible to them." <-- um no absolutely not. no one expects all activity within email to be visible to them, only emails sent to/from them. similarly with txt messages. DMs. etc.
#
[tantek]
"we tend to pool around well-maintained global indexes" <-- really? so what global index of email addresses are we pooled around? (hint: none)
KartikPrabhu1 joined the channel
#
[tantek]
still felt like a bit of solutioneering
#
Ruxton
we tend to pool around indexes.. even poorly maintained localized ones :P
#
Ruxton
mailing list = e-mails index group
#
Ruxton
newsgroups
#
Ruxton
forums
#
Ruxton
community groups (offline)
#
capjamesg[d]
GWG jacky I have had issues with paging in Microsub too. I still have to fix my paging implementation. I can see two pages but that's it.
wagle, IWSlackGateway, tetov-irc, jessealama and jacky joined the channel
#
aaronpk
[tantek]: read it again, it's not just about the content, it's also about user identification. In other words yes we do have a global namespace of both email addresses and telephone numbers
#
jacky
an acceptable global namespace
kimberlyhirsh[d] and jacky joined the channel
#
[KevinMarks]1
So how does vitra compare with Trillian? https://github.com/google/trillian
#
jacky
it's written in JavaScript :)
#
jacky
also doesn't need MySQL
jacky and Christian_Olivie joined the channel
#
jacky
you know
#
petermolnar
trillian is a messenger... and I miss the old one
#
jacky
I think what would make me really interested in some dweb stuff would be a way to transparently handle files without having to deeply change too much of my setup
#
Loqi
misses the old one too
#
jacky
like having a small abstraction over the filesystem and it'd "magically" sync across devices/systems
#
jacky
right now, a lot of that requires loads of rewriting and is usually stuck in JavaScript
#
jacky
like I do like the idea of a Dropbox-esque approach for stuff
capjamesg and srushe joined the channel
#
petermolnar
related: https://ansiwave.net/blog/semantic-web.html (title and slug is misleading)
#
jacky
> Doctorow's point was that you can't rely on people to provide accurate or comprehensive metadata on their web pages. We're lazy, and your utopian idea is just more work for us with little short term reward.
#
jacky
very true
#
jacky
we've trained people to just tap and click
#
jacky
versus helping them curate experiences
[schmarty] joined the channel
#
[schmarty]
ah, i see a fun point with the public SQLite DB-over-HTTP as well - the API is now SQL
#
jacky
GraphQL, eat your heart out! lol
#
[schmarty]
so with that translation layer, making mashups of data in DBs published this way is a lot more like building them locally
#
[schmarty]
lol totally
jessealama and jacky joined the channel
#
[KevinMarks]1
can you use FUSE to wrap the abstraction you want?
#
sknebel
[jacky]: didnt dat/hyper have that?
#
jacky
sknebel: yeah but it's locked in JavaScript (by implementation) so it's impractical for me to use
#
jacky
despite me really wanting to
#
sknebel
I dont think "dweb" and "not locked into 1-2 implementations" is really a thing?
#
jacky
[KevinMarks]1: sure it can; a bit more low level than I can muster at the moment and I don't know if I can port that onto Android as well
#
jacky
sknebel: heh yeah
#
jacky
I think only IPFS has more implementations mainly b/c of their use of cgo
#
sknebel
yeah. hyper had a rust thing under way, but not sure that ever got to working
[snarfed] joined the channel
#
[snarfed]
jacky there are definitely more good (non-dweb) data sync libs/services now that are JS/browser friendly! Noms, Replicache, etc. also lots of history in https://indieweb.org/sync
#
[snarfed]
generally higher level models, ie objects/user data, not backend file oriented like rsync
#
[snarfed]
(apologies, I misread your JavaScript comment 😁)
[tw2113_Slack_] and jessealama joined the channel
#
GWG
jacky: Do you do Android development?
#
jacky
GWG: recreationally (lol)
#
jacky
[snarfed] nah this stuff is good
#
GWG
jacky: Could I talk you into something?
#
jacky
tbh like if it means I have to implement some sort of 'bridge' that's just a bunch of socket calls
#
jacky
then _maybe_ I could try it out
#
jacky
would just want more of it to be in the single binary/app I use (for ease of use/deployment)
#
jacky
GWG: depends on the complexity, lol
#
GWG
jacky: It is a quick project idea
#
GWG
I want an app that can register a share target and send it to a server of my choosing along with a token that I either provide or it gets by IndieAuth. If the server is unavailable, it saves and allows retry later.
#
GWG
And optional tagging
#
GWG
So effectively, it's a Micropub app for bookmarks
#
GWG
Just customized for bookmarks
#
GWG
I'm working on owning bookmarks
#
GWG
jacky: What do you think?
#
jacky
I don't know if a dedicated Android app would be needed for that tbh
#
jacky
Like I could see you getting away with this with a PWA that embraces local storage
#
jacky
but yeah it does sound feasible
[aciccarello] joined the channel
#
GWG
jacky: I thought of a PWA
#
GWG
But native apps have their advantages
#
GWG
I'm working on owning my bookmarks again after pinboard had an outage and remained largely silent..I think I can reproduce a minimalist aesthetic
#
GWG
I'm trying to decide what constitutes MVP for me, and that's...
#
GWG
A simple UI for management...a way to send articles into it quickly, with or without tags.
#
GWG
And a way to search the stored bookmarks and use them in my post editor to post.
#
GWG
I have tags, the ability to store extra data for searching like author in there... but more to do
#
[tantek]
jacky: "small abstraction over the filesystem and it'd "magically" sync across devices/systems" <-- yes this. I keep moving more and more things to "Notes" in iOS / MacOS because it actually does a decent job at this and they're basically plain text (the way I use them at least). I'd prefer if they were directly .txt or .html etc. in the file system though.
#
GWG
Anyone else have thoughts on bookmarking caches that can be turned into posts...I have issues making them posts themselves
#
[tantek]
I think it's a reasonable workflow to go from quick /bookmark posts to /tag-reply to those to using a bookmarks or two as a citations in a /note to gathering related notes into a longer form /article that is making a larger point (out of a pattern of smaller points)
#
GWG
I just bookmark a lot of things compared to how many I post about. Pinboard alone has 44000 urls I wanted to remember
#
[aciccarello]
I experimented with a PWA for web share. I think it would work if you're not worried about teaching people how to install it
#
capjamesg[d]
I don’t know how this would play out but I kind of want to make a web page that just lists meta tags you can add to your site.
#
capjamesg[d]
There seem to be so many cool ones.
#
capjamesg[d]
Open search, web monetization, PWA manifests, rel=me, etc.
#
capjamesg[d]
Oh, it’s been done haha.
#
capjamesg[d]
What is a meta tag?
#
Loqi
mo’ <meta>, mo’ problems -- Tantek Çelik, 2015-01-23 https://indieweb.org/meta_tag
#
Loqi
ok, I added "https://github.com/joshbuchea/HEAD" to the "See Also" section of /meta https://indieweb.org/wiki/index.php?diff=79158&oldid=54486
#
capjamesg[d]
Haha. I think we need a serious definition for that [tantek] 😂
#
capjamesg[d]
Hey! That HEAD list has some IndieWeb tags there too like Micropub.
#
[tantek]
lol. you'll note that I was being quoted, I didn't actually edit that page 🙂
#
capjamesg[d]
I’ll submit a PR to add microsub too.
#
capjamesg[d]
Yes, indeed.
#
[aciccarello]
Oh, this is an interesting option:
#
[aciccarello]
```<!-- disallow Twitter from using your site's info for personalization purposes -->
#
[aciccarello]
<meta name="twitter:dnt" content="on">```
#
capjamesg[d]
Can I replace the definition for that tomorrow? 😅
#
jacky
smh they really prefixed `dnt` with `twitter:`
#
jacky
twitter--
#
Loqi
twitter has -2 karma over the last year
#
capjamesg[d]
Sorry for distracting from your point [aciccarello] re: web share.
#
[aciccarello]
jackey, I'm surprised they even give an option
#
[tantek]
jacky, re: IANA link-relations, you mean https://microformats.org/wiki/existing-rel-values for HTML at least :D
#
jacky
ahh true :)
[philbowell], Gotcha, [chrisaldrich] and BinarySavior joined the channel
#
npd[m]
what are folks using for avatar image identification?
#
Loqi
It looks like we don't have a page for "folks using for avatar image identification" yet. Would you like to create it? (Or just say "folks using for avatar image identification is ____", a sentence describing the term)
#
npd[m]
this script/service looks for a few different icon `link` relations: https://github.com/11ty/api-indieweb-avatar
#
Loqi
[11ty] api-indieweb-avatar: Return an optimized avatar image from a domain name input.
#
npd[m]
hcard microformat has a "photo" class for an image, but I don't know if that's regularly used as the "avatar" or smaller icon
#
[tantek]
what is icon
#
Loqi
An icon in the context of the indieweb typically refers to a home page icon for an indieweb site, or small decorative images to indicate types of posts https://indieweb.org/icon
#
[tantek]
^ npd[m] more there on how folks are identifying someone's icon/avatar/profile-photo
frittro__ joined the channel
#
npd[m]
thanks
#
npd[m]
I'm not sure I want my website's favicon to also be the avatar I use when I'm writing or commenting elsewhere
#
[tantek]
what's the difference?
#
[KevinMarks]1
Mastodon used to use the Portable Contacts xml namespace for avatars
KartikPrabhu joined the channel
#
npd[m]
while my website has a card about/depicting me, I don't want visiting my site or saving a bookmark to consist of a picture of me
#
aaronpk
same, i use a logo as my favicon, also because it's usually displayed a lot smaller than a profile photo is displayed
[fluffy] joined the channel
#
[tantek]
seems like extra work to come up with a second graphic
#
[tantek]
aaronpk, not last time I checked. when I added you to my home screen I got your face icon
#
aaronpk
that's because that's not the favicon
#
aaronpk
because i know it's going to show up bigger
tetov-irc joined the channel