#microformats 2025-08-26

2025-08-26 UTC
voxpelli, barnabywalters, Zegnat and gRegor joined the channel
#
gotra so is the argument against schemas that they put machines first, and kind of sidefiles info in a sense? are they ever acceptable? do microformats ever use JSON? any good links of articles discussing why not schemas?
#
gRegor mf2 parses to JSON https://microformats.org/wiki/microformats2-parsing
#
Loqi [preview] Tantek Çelik
#
gRegor schema.org tends to be overdesigned, classic example being that volcanos can have fax numbers. More on https://indieweb.org/schema.org
#
gotra lol
#
gotra I don't like that it was a joint venture of Google, MS, Yahoo, Yandex, but DuckDuckGo is using schemas apparently
#
[tantek] not really. it was a Google venture, and MS & Yahoo (which was using MS search at the time) merely signed on. Yandex didn't join until later.
#
gotra Oh I see. Yes, Google shouldn't dictate what we need to put in our code
#
gotra Brave search using it also. So it seems hard to avoid if you care about search engines? Hard to know how much they figure in
#
gotra this looks good https://storyneedle.com/time-to-end-googles-domination-of-schema-org/
Saige_Leah, fluffy, barnaby, gRegor, Kaja and [KevinMarks] joined the channel
#
[KevinMarks] Dan Brickley, who was the developer relations lead on http://schema.org and a genuine standards believer, left google last year. I'm not sure who's running it now
GWG, barnaby and [lazcorp] joined the channel
#
[lazcorp] I did some (voluntary) work on http://schema.org while Dan Brickley was there back in ~2014/2015, and Kevin is right - Dan was very much a believer in open standards
#
[lazcorp] That doesn't stop many of the criticisms of Google's dominance of many aspects of the project being valid and correct, of course - I just wanted to second Kevin's remarks about Dan
barnaby joined the channel
#
[KevinMarks] Also, Guha is at Microsoft now.
barnaby, btrem, [Sophia_wood], gotra, gRegor, [schmarty] and [pfefferle] joined the channel
#
gRegor Yeah, schema.org may be important if you want things to show up a certain way in a big search engine, but not strictly required. They still crawl and index "just" HTML, so the old ways probably work fine for most :)
#
gRegor I'm thinking of maybe product data for a shop, or maybe a news article. I don't know what the latest is for news articles, maybe they're still doing AMP?
#
gRegor I think a newer concern, aside from schema or not, is the decrease in people clicking through to actual websites when they're shown the LLM overview in search. I've turned that overview off on Google myself, following https://tenbluelinks.org/
#
[KevinMarks] The subtext I get from them both leaving is that the explicit markup approach that schema exemplified for search and that inherits from CYC lost out to the "read everything into soup and use an LLM" approach
#
[tantek] agreed on that subtext [KevinMarks]. which is just fine by me. if that results in Google de-prioritizing schema markup I think that's fine.
#
btrem I was on the schema mailing list for a while. I liked it at first, but fairly quickly realized it was an attempt to categorize everything that was doomed to failure. A regular post on the list would be "I have such-and-such, there's no schema for that, let's make one." The result being, I dunno, thousands (tens of thousands? surely not!) schemas with absurd specificity.
#
[tantek] ^ that's a decent valid and accurate critique
#
btrem Like, restaurant, iceCreamRestaurant, gourmentIceCreamRestaurant, ad infinitum. Because someone said "my client's shop is not just an ice cream shop, it's a /gourmet/ ice cream shop. How can I document that with schema?"
#
btrem Mispelled that, should have been gourmetIceCreamShop. But you get the idea.
#
btrem I only knew Dan Brickley from that list. It's nice that he believed in standards, but his justification for the multiplying schemas was frustrating. I eventually gave up.
#
gRegor I remember hearing of some old microformats discussions along those lines, like lunar coordinates: https://microformats.org/wiki/luna
#
gRegor lunarIceCreamShop
#
[tantek] The hierarchical venue types was so stupid. It's like we already had flat tags and folksonomies to solve that problem
#
btrem lunarIceCreamShop :-D
#
btrem [tantek]: I think the idea behind hierarchical types was to keep things DRY. So iceCreamShop inherited all the properties of restaurant. And gourmetIceCreamShop inherited everything from iceCreamShop, and thus everything from restaurant. (Not saying that was good thing. Only that it was the justification. I think.)
#
[tantek] it trivially fell apart when with multitype venues
#
[tantek] which many cities have many variants thereof
#
btrem Heh. I think it fell apart not just with venues.
#
[tantek] and not just combination Pizza Hut and Taco Bell
#
[tantek] agreed, there were many more flaws like the canonical volcano with a fax number
#
Loqi inheritance has -1 karma over the last year
#
[tantek] inheritance--
#
btrem BTW, the inheritance thing was abused in both directions. There were absurdly general schemas, too. Like a schema for a webpage. That way, more specific pages could inherit the properties of webpage.
#
btrem I think I tried to point out the idiocy of marking up a webpage with schema to tell search engines. "hey, this is a web page." Like WTELF? :-D
#
[tantek] 😂
#
btrem I wish I were making that up.
#
btrem https://schema.org/WebPage
#
btrem Forgot about this one: https://schema.org/WebPageElement Why?
#
gRegor If only there were some kind of object model to know if it's a table XD
#
btrem I have to say, though I was frustrated at the time, now I find it all amusing. Here's an example of a generic schema: https://schema.org/Thing :-D
#
btrem gRegor: lol
#
gRegor Maybe a WebPage Object Model? :D
#
btrem :-D That is good!
#
gRegor *goes to schema mailing list*
#
btrem LMAO
#
gRegor Yeah I imagine that would be frustrating at the time. I'm glad I missed all that.
#
btrem Oh, you aren't exaggerating.
#
[tantek] unfortunately the SEO cult is still parroting ancient Google utterings so it's still around in various places and lore
#
btrem ...exaggerating about the "goes to mailing list" part, that is.
#
gRegor (feel free to take these ideas for an April Fools joke)
#
btrem Nah, I left that mailing list years ago. And probably the mailing list is dead, replaced by a git repo or who knows what.
#
[tantek] pretty sure I never joined the list because I was so frustrated with everything at the time (circa 2011?)
#
btrem I was on the list from 2012-17, apparently. (Looking at emails from the list that I saved.)
barnaby, gRegor, [mattl] and hidjy[d] joined the channel