#microformats 2020-12-14
2020-12-14 UTC
[tantek] joined the channel
# 
btrem edited /menu-examples (+245) "/* T.S. McHugh's Irish Pub */ Adds details about draughts menu." (view diff)
btrem joined the channel
# btrem My idea for an include-pattern:
# btrem I did all this work on it, and realized that, when all's said and done, the easiest thing is to just use the itemref attribute, which is already in html5. (sigh) I don't know why that didn't occur to me before I did all the work.
[prtksxna] joined the channel
# btrem But, in case you want some idea of how it might work, check out the link. And if you don't want to use itemref, maybe my idea could be used instead.
ethanyoo, [chrisaldrich], gRegorLove_, Loqi, sebbu, [tantek], [KevinMarks], [Ana_Rodrigues], [jgarber], [jgmac1106], KartikPrabhu, [snarfed] and btrem joined the channel; btrem left the channel
# btrem Ok. I don't know what the protocol is about doing work outside the wiki. I created the pages because it seemed too large to include in their entirety there.
[chrisaldrich] joined the channel
# btrem I wanted to create a plausible case where an include pattern would match existing practice and would make authoring easier.
# btrem Similar situations exist for the proposed h-menu, where some menus -- notably beer lists -- have a price in a header that applies to all beers in that menu section.
# btrem I don't see a page or section for include pattern brainstorming. (My search fu is weak today.) Can someone point me in the right direction?
# btrem [tantek] or did you man https://microformats.org/wiki/include-pattern-strawman ?
# tantek moved /include-pattern-strawman to /include-pattern-brainstorming "it's not really a strawman"
ben_thatmustbeme joined the channel
# btrem edited /include-pattern-brainstorming (+1736) "Adds proposal for custom element with custom id" (view diff)
# btrem edited /include-pattern-brainstorming (+407) "Adds proposal for itemref attribute." (view diff)
btrem joined the channel
# btrem edited /menu-brainstorming (+437) "Proposes using p-nutrition for recipe items that have calories added to it, with caveat that it is not part of h-product." (view diff)
ben_thatmustbeme, [jgmac1106], Phae, craftyphotons__, [barryf] and [tantek] joined the channel
[jeremycherfas] joined the channel
# tantek edited /include-pattern-brainstorming (+125) "/* Use html5 itemref attribute */ note mf2 brainstorming for use of itemref" (view diff)
[schmarty] joined the channel
# 
[schmarty] hehehe, "[...[ if we believe the schema.org implementation announcements[...]"
btrem joined the channel
# btrem edited /microdata (+112) "Corrects status of microdata at W3C (editors for microdata were found, so the standard survived. Rumours of its death were greatly exagerated.)" (view diff)
# btrem Yes, I think the part that's relevant for us is that itemref is, I think, viable. If we decided to adopt it, that is.
# tantek edited /microdata (+139) "Undo revision 70305 by [[Special:Contributions/Btrem|Btrem]] ([[User talk:Btrem|talk]]), actually no, it was abandoned and dead for nearly 4 years, then briefly re-adopted in 2017 and once again abandoned subsequently in less than a year afterwards" (view diff)
# btrem FWIW, I actually think there was real potential for microdata, but for something bigger than microformats. I think it should have been developed with something bigger in mind. IMHO and all that.
# [schmarty] https://caniuse.com/mdn-html_global_attributes_itemref itemref as an attribute is supported across the board.
# btrem That's a strange assertion. What *should* a browser do with itemref?
# [schmarty] caniuse points to MDN docs which appear (at first glance) to be a dupe of the whatwg docs for microdata.
# btrem The point of itemref, afaict, is for microdata parsing. Which is not really something browser do, at least not natively.
# [schmarty] there's
# btrem [tantek] what did Firefox do with microdata before it "unshipped" it?
# [schmarty] a note above all the item* attributes in the MDN HTML global attributes list: https://developer.mozilla.org/en-US/docs/Web/HTML/Global_attributes#attr-itemid
# [schmarty] just says "*Note:* The `item*` attributes are part of the WHATWG HTML Microdata feature."
# 
KartikPrabhu I am confused. You can literally put any attribute in your HTML like foojangle = "beepboop"
# KartikPrabhu so unless someone consumes it it seems useless
# [schmarty] so yeah i MDN just pulled that in because the whatwg spec exists
btrem joined the channel
# 
aaronpk "Microdata does not generally interact with browsers, being a static document format that lacks any DOM interface." https://www.w3.org/TR/microdata/#security-considerations
# btrem arrgghh meta complaint: webchat.freenode.net dropped me from the channel. Again. :(
# [schmarty] whatwg version also dropped the section on how to convert microdata to rdfa (not really relevant here but 😬 )
# [schmarty] this is a bit confusing but it looks like there was agreement to remove microdata from the whatwg spec in mid-2016 https://github.com/whatwg/html/issues/208
# [schmarty] that was blocked by web-platform-tests still having microdata tests in them, but that was merged in 2019
# [schmarty] oops sorry it was merged in 2016
# [schmarty] yeah i am confused by that. i think it pre-dates the move from w3c to whatwg so maybe this issue was from back when this repo was under w3c control??
# [schmarty] wait sorry that doesn't make sense since the move happened in 2013 according to https://www.w3.org/Bugs/Public/show_bug.cgi?id=22783
# [schmarty] oops, did i link to some junk?
btrem joined the channel
# [schmarty] OH okay. that makes a lot more sense.
# btrem I'm curious: unless there was a bug or whatever, why did FF remove the implementation? IOW, why should they care whether other browsers implemented it or not? I've never worked on a project at even a small scale, so it's hard to understand one as large as a browser.
# btrem Another question: what is the point of the microdata DOM api? Microdata and microformats are, to my eyes, something for communities to use. E.g., cooking sites/blogs etc. use h-recipe, and so search engines show recipes in search result rich snippets. But that's for Google or DDG to figure out, isn't it? And for microformats or schema.org to create
# btrem or change as needed.
# btrem To take another example, one of the motivations for my idea of h-menu is so that restaurants can publish menus on their web site, and any services they subscribe to -- Open Table, Grub Hub, etc. -- can read.
# btrem If a menu were created in microdata instead of microformats, I'd think it would be up to restaurant publishers to agree on a format, and for services to agree to read them. This is sort of why I don't quite get what browser are supposed to do with microdata.
# [schmarty] btrem: if i understand right, w3c and whatwg agree with your reading of it - the API specification was for browsers but folks eventually agreed that browsers were not intended to be consumers of microdata.
# btrem [schmarty]: but it that's the case, then what does it mean that Firefox dropped the md dom api because of lack of implementation?
# 
[tantek] btrem, re: "microdata ... it would be up to restaurant publishers to agree on a format", two big problems with that: 1. there is no microdata community to actually make that happen. and 2. restaurant publishers are not vocabulary experts, so they don't actually have the necessary infoscience skills to come up with a "format".
# [schmarty] the big open question in my mind is: who is taking care of this standard? who are the groups or individuals that are consuming and publishing microdata?
# btrem See, that why I think it's a shame that microdata died. I think of microdata as a possible way to make html extensible without the rat's nest of namespaces.
# btrem Ok. Also, I note that many new elements in html5 came about because of common class names.
# btrem What's missing is a middle step between a class name and a new element.
# btrem If anyone is familiar with wine classifications, there's an analogue: in Italy, for example, they had DOC and DOCG for classified wines, iow, wines deemed high quality. And "table wines" for those not deemed high quality. There was too wide a gulf between them.
# btrem The solution was a middle step: IGT, for wines that are not ready for DOCG, but are better than table wine.
# btrem aaronpk: an obvious way for web authors to imagine how new elements are created. By sharing existing conventions. Microformats offers that, but authors have to know about them before they know about them. That is, the source code alone does not tell a novice that class=h-card means something more than class=blue or class=bottom.
# btrem [tantek] only to those who already know that it's a mf2 prefix.
# btrem By contrast, an itemtype has a url, where one could (presumably) learn what the type is, and how to use it.
# btrem I've thought about this a lot. Sorry you don't like analogies, [tantek], but I think it is appropriate. Something like microdata, where authors could have a semi element, if you will, would be useful imho.
# btrem And if a microformat gains enough adoption, it could then be elevated to an actual element. Or not, maybe that step isn't even necessary. Still it'd be useful imho.
KartikPrabhu joined the channel
# btrem I guess that raises the question: is there any democratic way for new elements to be added to html?
# btrem Joining a w3c working group is simply beyond the reach of most web authors.
# btrem science <> techology Not by a long shot.
# btrem technology (where's spell check when I need it?) :)
# btrem I missed that. (freenode irc dumped my off. Again.)
# btrem Well, that last comment sums up why that's not really going to work.
# btrem As to your remark about web publishers and restaurants, yeah, that's true, but I'm using restaurants loosely, more like, "people who publish websites for restaurants".
# btrem I think there are ideas worth exploring and considering from people who are not part of w3c working groups is my point.
# btrem I wasn't arguing that microformats are done at the w3c.
# btrem But I'm "people who publish websites for restaurants". So that means I shouldn't participate in h-menu.
# btrem I'm not an expert in vocabulary. That means people who are familiar with the needs of restaurants don't know what's needed for restaurants. This is a sort of weird anti-tautology.
# btrem I'm not sure we actually disagree about this. :)
btrem joined the channel
# btrem Meta question: any recommendations for an irc client? This web thing is not working out for me. I tried Colloquy but couldn't get it to anything.
# btrem ...do anything.
# [tantek] subject matter experts can be expected to design bad technology because they're not technology experts. the technology experts can be expected to design a theoretical technology that doesn't actually match real world needs because they're not subject matter experts in those real world needs.
# btrem aaronpk: do you use the web interface, then?
# btrem :-D
# btrem As to the point that I missed when I got dumped: my point is that, it seems to me, a method for subject matter experts to participate is needed. The microformats community is good for that, but it's not something that subject matter experts would likely know about.
# btrem The microdata idea -- or something like it -- might have been a way for subject experts to learn what's out there simply by looking at source code. Which is how a lot of learning on the web happens.
# btrem And a better alternative is...?
# btrem Yes, another frustration is learning bad habits from bad authors.
# btrem Agreed. I'd wager that would happen much more quickly if a bunch of websites were using microformats.org/h-entry, because the link is right there.
benterry joined the channel
# btrem I agree. But seeing it in lots of places is enough. I think a url is a much more recognizable pattern, though.
# btrem As I said earlier, if you don't know about mf2, you would have no way of knowing that h-xxxx is a root microformats name.
# btrem I do recognize that the added complexity of microdata attributes is a draw back.
# btrem I guess I'm saying that microdata had a potential usefulness that was not really explored. It was limited to an alternative to microformats when an alternative was not really needed.
benterry joined the channel
# btrem edited /microformats2-brainstorming (+625) "/* adopt itemref */ Adds quote about itemref from whatwg html5 spec." (view diff)
# btrem I added a note about itemref, just in case we decide it's a good idea. It is easy, and the attribute already exists, and won't trigger errors in a validator. At the end of the day, while my custom element is nifty, it doesn't offer any advantage over itemref, and it does add lots of complexity.
# btrem Lots of effort on my part, but a failed effort. I should have realized that much earlier. :-o
# btrem For the sake of ease, in case you want to see my idea that I think is not really that great, here's the link: https://btrem.github.io/microformats/include-pattern/
# aaronpk simple example: https://pin13.net/mf2/?id=20201214235106730
# aaronpk if you paste that HTML into XRay, XRay parses the microformats and then pulls out the full author info into each entry https://xray.p3k.app/
# aaronpk (note you have to include a URL like https://example.com when checking that on XRay)