#dev 2021-08-06

2021-08-06 UTC
[tw2113_Slack_], nertzy, gerben and wagle joined the channel
#
GWG
capjamesg[d]: I fixed the form fills and added a bunch more tweaks. Haven't done the menu yet.
#
capjamesg[d]
GWG πŸ™Œ
#
GWG
Thinking of editing the next previous links to only cover the same type of post
capjamesg and nertzy_ joined the channel
#
@vishae
Day 5 and 6 of #TheShapeOfShadows: Things I learned: * that as much as I love journal pages filled with sketches and doodles, I am not good at sketching or doodling * the Northern Pacific sea star, the Japanese kelp, the European fan worm, and the… https://www.brid.gy/publish/twitter
(twitter.com/_/status/1423548546486243332)
neceve and hendursa1 joined the channel
#
vikanezrimaya
aaaaaa I need to write instrumentation for a connection pool library because it doesn't contain ANY kind of logging and has a hard-to-debug problem that triggers whenever I use my database too much
#
capjamesg
Are there any benefits to adding JSON-LD to a page when you already have microformats?
#
capjamesg
I'm thinking of adding JSON-LD to my site to explore structured data (and parsing it for my index).
#
capjamesg
(with help from a parser, of course)
#
vikanezrimaya
I'd say it violates DRY principle but if you absolutely need it for something feel free to make it somehow generated - preferrably on-the-fly so you don't forget to update it
#
capjamesg
I agree.
#
capjamesg
DRY was my main hesitation.
#
capjamesg
(And I want to support microformats over other types of structured data in general :D)
#
vikanezrimaya
I was using Granary to generate RSS feeds for legacy clients by inserting a Granary converter link that would simply fetch the page and then convert MF2-HTML into RSS or Atom
#
vikanezrimaya
This way RSS feeds are available for whoever needs them and I don't incur an admin tax for having them since Granary handles all the dirty work on-demand
[KevinMarks] joined the channel
#
@hondanhon
I purposefully bought my kid an MP3 player so he could listen to podcasts without looking at screens and I HATE IT THAT IN OUR YEAR 2021 IT'S SO DIFFICULT TO DOWNLOAD JUST THE GODDAMN MP3 FILE FOR A GODDAMN PODCAST THANKS PLATFORMS YOU RUINED EVERYTHING
(twitter.com/_/status/1423476490063147010)
#
jeremycherfas
Amen to that.
#
vikanezrimaya
I definitely agree with the fact that platforms ruin everything to lock their users in
#
vikanezrimaya
I am considering throwing away all of my storage backend code and writing something that doesn't depend on Redis
#
[KevinMarks]
For the JSON-LD and microformats question it's like any of these things - do you want a specific silo to have markup in the format they insist on?
#
vikanezrimaya
Since the current Redis backend is utterly broken (thanks to a buggy dependency that I do not currently have the energy to debug)
#
Loqi
Decaying Silos as dead malls 2018-01-20
#
vikanezrimaya
[KevinMarks]: I'd personally say "silos can go do whatever they want and I'm not gonna cater to them" because I don't get a lot of engagement from silos anyway
#
vikanezrimaya
I guess I can thank myself for being extremely unpopular
#
[KevinMarks]
Quite - the point of that article is that you can tune your presentation on silos, but then you're adding DRY violations, so I didn't repeat myself and gave every silo a different preview
#
vikanezrimaya
I think image previews don't matter
#
vikanezrimaya
Mainly because I think text is the most important medium on the internet
#
vikanezrimaya
All these fixations on images and pretty pictures are detracting from the original goal - transfer a coherent (and sometimes not) thought from one brain to an arbitrary amount of other brains
#
vikanezrimaya
And thoughts are mostly represented in text, with pictures accompanying it, not replacing or substituting
#
vikanezrimaya
But that might be a bit too radical of a worldview to expect others to share it
#
vikanezrimaya
Besides, some people like sharing pictures
#
Murray[d]
It's not just a binary either. I think most information I consume or interact with these days is video πŸ€·β€β™‚οΈ
#
Murray[d]
That's certainly where I get the most "worth" out of the web
#
vikanezrimaya
I noticed I usually only watch videos when I have nothing else to do - for example, I find something interesting on YouTube from an array of channels I'm subscribed to, or open a Twitch page to see if any of my favorite streamers are live
#
vikanezrimaya
When I need to consume a piece of information and reuse it I usually prefer text - which probably impacts my views on information sharing on the Web a lot
#
vikanezrimaya
Additionally textual information just so happens to be the most accessible out there β™Ώ - since it can be read normally, converted to audio for those who can't read but can listen, displayed in Braille for those who don't see but have a Braille display and probably adapted in some other ways I didn't think of
#
vikanezrimaya
Current commercial speech recognition is flawed, resulting in numerous errors when processing audio data. Even I, having a good hearing, sometimes don't recognize words in an audio or video and have to rewind, sometimes several times
#
vikanezrimaya
I don't have such a problem with text (though some people do, but text-to-speech solutions of acceptable quality do exist and potentially solve the problem of someone not being able to read things quickly enough or at all)
#
vikanezrimaya
In other words, I believe text to be a superior form of communication thanks to its adaptability to any medium and needs of a user
#
Murray[d]
Text certainly has a lot going for it, but whether or not it's accessible is very much down to the individual. There are certainly a lot of tools available for text conversion: between languages, to audio, to other sensory input (such as touch). And, ultimately, accessibility for most other media results with text in some form or another: captions, annotations, alt text.
#
Murray[d]
That said, many people cannot read and to them, text is completely inaccessible. I'm talking about those with reading impairments that are behavioural, cognitive, physical, and more. To many, video (or images) become a lifeline to information, and for many more (myself included) video/images simply *stick better*, so as learning tools are very useful. But to others, those same videos/images would be accessible, or text is easier/more useful
tetov-irc joined the channel
#
capjamesg
vikanezrimaya I had no idea Braille displays existed. That's so interesting.
#
capjamesg
I prefer text for a lot of my consumption.
#
capjamesg
I almost never watch video tutorials, instead opting for text whenever possible.
#
capjamesg
As Murray[d] said, video has a lot going for it.
#
vikanezrimaya
capjamesg: https://en.wikipedia.org/wiki/Refreshable_braille_display <- I only know of them because I heard once that the Linux kernel supports these
#
capjamesg
I'm seeing more subtitles available, YouTube now has support -- although the extent to which this is implemented I am unsure -- for marking sections in video.
#
capjamesg
vikanezrimaya that's amazing.
#
vikanezrimaya
capjamesg: as far as I know, timecodes on the timeline work only when marked up by the video uploader
#
Murray[d]
At an old job we needed to provide fully accessible documents, which meant having braille displays, braille printers, and haptic feedback screens in our test device library; some very interesting technology all around 😊
nertzy and rockorager joined the channel
#
vikanezrimaya
I think I just managed to assemble a prototype web crawler that uses MF2 content! I could probably connect a search indexer to it and create an IndieWeb search engine
#
vikanezrimaya
this kind of thing probably exists but this one I built myself
hendursaga joined the channel
#
capjamesg
vikanezrimaya I think there have been talks about an IndieWeb search engine.
#
capjamesg
My web crawler actually processes microformats too.
#
vikanezrimaya
There was a link dropped to whoosh a week or so ago
#
vikanezrimaya
I plan to integrate that to get an index database
#
vikanezrimaya
Then I could expose the database as a service or potentially offer database dumps
#
vikanezrimaya
Depending on bandwidth requirements
#
vikanezrimaya
The crawler thingie is written in Python and I will probably upload the sources once I go have a break and fetch a burger and a coffee for myself as a treat for coding useful stuff
#
vikanezrimaya
I feel like hypoglycemia is coming soon... 😡
#
capjamesg
My crawler is written in Python too!
#
capjamesg
Well, it's a crawler and spider.
#
vikanezrimaya
Python is so good for quickly making something cool
#
capjamesg
It crawls but also logs any potential issues on my site (missing alt text, very long meta descriptions, etc.)
#
capjamesg
For sure.
#
capjamesg
Have you heard of pytesseract?
#
vikanezrimaya
capjamesg: Wait what's the difference between a crawler and spider
#
capjamesg
It's an open source package that lets you extract text from an image.
#
capjamesg
And you can integrate it into Python in a few lines of code.
#
vikanezrimaya
capjamesg: Nope
#
capjamesg
Optical image processing in a few lines of code!!
#
capjamesg
(Sorry, that was unrelated :D)
#
vikanezrimaya
It's very related because text on images needs to be accessible
#
capjamesg
They are basically the same but crawlers typically are used for search indexing but spiders are used for architecture analysis / etc.)
#
capjamesg
*architecture analysis and finding issues on a site, etc.)
#
vikanezrimaya
Then mine is more of a crawler than a spider
#
capjamesg
pytesseract has its limits but I'm going to use it to extract text from coffee bag images into a dictionary.
#
capjamesg
So I can snap a pic of a coffee bag and get all the meta data in a dictionary.
#
vikanezrimaya
Currently my code jumps from page to page using common mf2 properties for reactions (e.g. comments and likes), authors and reply contexts. If I insert an indexer in-between jumps then I could build a search database
#
vikanezrimaya
I think I'll upload it now
#
vikanezrimaya
So I could drop the link
#
capjamesg
I'm interested!
#
capjamesg
I use sqlite3 for building my search db.
#
capjamesg
And sqlite3's full text search for retrieving information.
#
@tef_ebooks
a microservice is what you call it when, three years in, the original monolith programmer is tired of dealing with their own legacy code, tired of code review, and begins a greenfield project to recapture the feeling of working on their own once more https://twitter.com/ChrisAnnODell/status/1423618814390964224
(twitter.com/_/status/1423628574561652738)
#
capjamesg
I think the thing about an IndieWeb search engine is that building a search index for multiple sites is a big job beacuse you can't rely on everyone's markup.
#
[KevinMarks]
Oh yes. One of the motivations for microformats originally was that, as we were building a search engine at the time.
#
sknebel
phinde, the search engine behind the chat logs (and other things) also parses microformats
#
vikanezrimaya
capjamesg: Re: can't rely on everyone's markup: I think we need to encourage correct markup if we build a search engine, for example, by making it somehow affect search rankings
#
vikanezrimaya
By the way... I will need to research search ranking algorithms too
#
capjamesg
vikanezrimaya BM25.
#
capjamesg
That's what I use. It's built in to sqlite's full text search so it is super easy to implement.
#
capjamesg
But there's also TF-IDF and a bunch of others you can look at.
#
capjamesg
Google's search centre docs on structured data basically says if you use structured data you're more likely to get some interesting search feature (like being featured in an article box, a recipe box, a book box, whatever) on Google.
#
capjamesg
They advocate for JSON-LD but they still process microformats.
#
capjamesg
(If you can't tell, I have researched this a lot for my search engine)
#
capjamesg
KevinMarks Really?
#
capjamesg
I had no idea.
#
capjamesg
Oh, Technorati. Yes. I remember now!
#
vikanezrimaya
capjamesg: This could be a nice incentive if user is searching for a specific thing, e.g. a recipe
#
vikanezrimaya
For example, automatically rank everything containing h-recipe higher
#
capjamesg
For sure.
chenghiz_, rockorager and push-f joined the channel
#
GWG
I am thinking about navigation at the post level
#
GWG
What is navigation?
#
Loqi
Navigation refers to the UI, UX and markup allowing people to move around a website https://indieweb.org/navigation
#
GWG
Needs a refresh...last touched in 2018
#
GWG
Options to consider:
#
GWG
all posts
#
GWG
"root" posts only (everything but responses)
#
GWG
sequential nav only among posts of a specific type (notes, photos, articles, etc.)
#
GWG
I just was fiddling with the idea of switching from all posts to the last option, posts of a specific type
#
GWG
Wondering what others are thinking
rockorager and [manton] joined the channel
#
[manton]
@GWG Sequential nav within a post type makes sense, although I wonder in practice how often people do that vs. just going back to the main list and clicking what they want. I personally like keeping replies off the home page because they are so different and sometimes lack context.
neceve joined the channel
#
Murray[d]
I have sequential navigation within posts of a given type (for me: notes, articles, journals) as well as dedicated paginated feeds for each. My homepage feed is then a mixture (effectively "latest").
#
Murray[d]
I'm not sure how many people use the sequential nav, though the per-type feeds definitely get some use. But then when I'm catching up on other people's posts, I often use sequential navigation (next/last), so I wanted to have it on my site πŸ˜„
#
sknebel
that'd be interesting to measure on a somewhat-popular site
#
GWG
I'm switching from all post navigation I think on single post pages to per type
#
GWG
I have per type archive pages, and those are paginated
#
GWG
I don't keep replies off the homepage yet, but am considering it, or at least, separating them
#
GWG
Did anyone catch [chrisaldrich] talking about slow and fast streams?
#
GWG
I thought I might add an article feed as well as recent posts
neceve joined the channel
#
Murray[d]
I guess I've already made that split (to an extent), because my bookmarks and notes never hit the homepage any more πŸ˜„ Main reason was to not lose articles too quickly
#
Murray[d]
reviews too, now I think about it; and a lot of the content that used to go into the journal is now hidden entirely as private check ins (only private until I build a front end)
rockorager joined the channel
#
GWG
I don't have review posts yet
#
GWG
I have checkins...
#
GWG
They are rare enough I might want them on the front page
#
GWG
Lots of things to think about. I am renovating, but I don't have to figure out everything now
#
Murray[d]
Very true 😊
#
Murray[d]
I'm unsure what to do with check-ins (when I get around to making them public), but I'd been toying with the idea of a single item "latest update" style section on the homepage. Reviews, check-ins, and a few other post types might go in there, so that it's ephemeral but at least highlights what I'm currently up to (like a mini /now page, I guess).
#
Murray[d]
If it goes stale, I could always hide it conditionally, or replace with latest note instead
#
Murray[d]
I've also been wanting to change my homepage article feed to multiple smaller feeds based on top-level categories that I use, which will make some articles a bit stickier (as some categories get less use), but that's all very far in the future ideas right now
rockorager and [snarfed] joined the channel
#
[snarfed]
could => #indieweb
rockorager, capjamesg and [jeremycherfas] joined the channel
#
GWG
[tantek]: Since you last worked on the navigation wiki page, has your thinking evolved since 2018?
rockorager, [tantek]1, shoesNsocks and tetov-irc joined the channel; rockorager left the channel