#microformats 2015-05-27
2015-05-27 UTC
Tigt joined the channel
# 
Tigt pardon me folks, I was wondering if I could ask a few questions about h-feed
# Tigt okay, first one is
# Tigt Is there some de facto size for its u-photo
# Tigt or at least some sort of minimum/maximum I should be mindful of
# Tigt I'm probably going to repurpose one of my shortcut icons in the <head> so it could be anywhere from 16x16 to 192x192
# Tigt ideally I could point it at an SVG but I doubt things are ready for that yet
# 
KartikPrabhu Tigt: I don't understand the use-case for u-photo of a h-feed? Are you thinking that readers will use the u-photo?
# Tigt well, I didn't know myself, but it's on there: http://microformats.org/wiki/h-feed
# Tigt I figured it would be used like a favicon mostly
# Tigt like how feedly marks each feed provider with its own favicon
# KartikPrabhu Tigt: since I doubt any h-feed readers use it that way yet you don't really need to worry about it
# Tigt that works
# KartikPrabhu Tigt: also photo parsing at the moment does not use the new srcset and source stuff because again, no explicit use-cases
# KartikPrabhu but once they do, I'm sure your problem will be addresseed
# Tigt oh it's just a regular old link[rel="shortcut icon"]
# Tigt so they should be able to follow the href
# KartikPrabhu Tigt: yeah I doubt if current h-feed readers are doing that. But even if they are, I really think they are not using the h-feed>u-photo
# Tigt my other question is, if I put an h-feed on all pages of my site, with the u-url pointing to the actual location of the feed, is that okay
# Tigt sort of like how the RSS <link> allows autodiscovery no matter what URL of a site you paste in
# KartikPrabhu Tigt: see https://indiewebcamp.com/h-feed#partial_feeds for a possible answer though again I doubt if readers are using such an algorithm. Mainly they'll look for rel=feed links on the page
# KartikPrabhu Tigt: on my site I have put the rel=feeds in the nav bar at the bottom which is common to all post pages: https://kartikprabhu.com/
# Tigt huh, interesting
# Tigt do mf parsers obey the <noscript> block?
# KartikPrabhu Tigt: good point. I think it varies between parsers unfortunately
# KartikPrabhu hopefully will be standardized soon
# Tigt yeah that's what I was worried about
# KartikPrabhu Tigt: mf parsers are not supposed to run javascript
# Tigt sounds about right
# Tigt okay, if I'm reading this correctly, if I don't have a page with full h-entries on each, I should use a separate u-uid h-feed page?
# KartikPrabhu Tigt: that was the suggestion yes. I don't know if people use it and h-feed readers don't use that I think
# Tigt I can't say I'm big on making another separate feed file
# Tigt I'm marking up a webcomic site, so I really don't want to make the home page load each image for UX reasons
# Tigt so having a full h-entry for each update isn't ideal
# KartikPrabhu Tigt: that is also how RSS works though
# Tigt but if it's what's got to be done, I'll do it
# Tigt yeah it's even worse
# KartikPrabhu Tigt: do you expect your readers to use h-feed readers?
# Tigt oh, no
# KartikPrabhu don't worry about pre-mature optimisation. Even current h-feed readers don't use that algorithm
# Tigt I just figured for those out there using/experimenting with h-feed readers, what good practices I should be following
# KartikPrabhu does Woodwind or monocle do this?
# KartikPrabhu Tigt: best practices haven't been fully-eshtablished yet because all h-feed readers are still experimental so I would recommend not worrying too much at the moment
# Tigt good point
# KartikPrabhu Tigt: but your use-case is a good one
# Tigt oh, for future reference, is there some sort of validator/babelmark equivalent I could test my pages on
# KartikPrabhu for all microformats?
# Tigt or just some, I'm not picky
# KartikPrabhu this does some sort of validation: http://indiewebify.me/ and you can test your markup with parsers from http://microformats.org/wiki/microformats2#Parsers
# Tigt wonderful, thank you very much
# KartikPrabhu note not all parsers in that list are the same
# KartikPrabhu i mean same level of microformat support
# KartikPrabhu php-mf2 and mf2py are more widely used, though the Go parser also seems good
# Tigt well, I guess it wouldn't be the web if consumers didn't differ from each other
# KartikPrabhu yup :)
# KartikPrabhu but people have been working on a standard suite of tests that parsers will test against so hopefully in the future all of them will be similarly compliant
# Tigt I'm assuming some of these don't understand mf2
# KartikPrabhu I think most of them do understand mf2 but maybe not all kinds of properties
# KartikPrabhu specially newly added/changed parsing rules
# KartikPrabhu hence I recommended php-mf2 and mf2py and Go , afaik those have been on top of new rules and features
# Tigt excellent
# Tigt you've been extraordinarily helpful, thank you so much
# KartikPrabhu sure thing Tigt
# KartikPrabhu Tigt: also if you want to see how your feeds appear in current h-feed readers here are 2, http://reader.kylewm.com/ https://monocle.p3k.io/
# Tigt ah, even better
# Tigt noted
# Tigt these sign-ins are the rel=me stuff, right
# KartikPrabhu Tigt: see: http://indiewebcamp.com/indieauth
# Tigt I feel like a traitor using Tumblr for this but baby steps, I guess
# KartikPrabhu Tigt: yes. baby steps
# KartikPrabhu babysteps++
# Tigt yeah I snagged a pretty good one, just need to make sure I can hold onto it
# Tigt Tumblr and Cloudflare get into spats every so often
# Tigt ah, hm
# Tigt it looks like IndieAuth does not honor HTTP redirects
# Tigt at least the version Woodwind is using
# KartikPrabhu hmm I think it does honour permanent redirects 301 or something
# KartikPrabhu kylewm: ^^
# Tigt Tumblr's "link pages" functionality might do 302s to prevent caching issues for when the user changes them, then
# Tigt I'll have to hardcode it in
# Tigt there we go
# Tigt it's likely it's not a 301 at all
# Tigt I changed it, but let me set up another
# Tigt hm
# Tigt http://ti.gt/test
# Tigt I can set up one that points to an actual rel=me endpoint if you prefer
# Tigt it also appends this weird PHP session indicator fragment of #_=_
KartikPrabhu joined the channel
# Tigt yet another reason for rel=canonical I guess
KevinMarks_, fuzzyhorns, KevinMarks__, tantek, KartikPrabhu and dariusdunlap joined the channel
# 
@Root013 æ§‹é€ åŒ–ãƒ‡ãƒ¼ã‚¿ï¼ˆhttp://microformats.org hAtom）をテンプレートã«çµ„ã¿è¾¼ã‚“ã§ãƒžãƒ¼ã‚¯ã‚¢ãƒƒãƒ—ã—ãŸãŠè©±ã€‚ http://website-homepage.com/wordpress/customize/microformats-hatom #wordpress #google #SEO (twitter.com/_/status/603440545147428866)
dariusdunlap_, KartikPrabhu, Left_Turn, KevinMarks, pfefferle, kez, pfefferle_, csarven and glennjones joined the channel
# 
KevinMarks hi glennjones
# KevinMarks I'm looking through the tests from mf2py pov
# 
glennjones Morning Kevin
# KevinMarks did we work out the white space issue?
# KevinMarks there is also a relative path issue
# KevinMarks eg https://github.com/microformats/tests/tree/master/tests/microformats-v2/rel%3Dalternate is getting http://example.com from somewhere magic
# KevinMarks s/example/tantek/
# 
Loqi KevinMarks meant to say: eg https://github.com/microformats/tests/tree/master/tests/microformats-v2/rel%3Dalternate is getting http://tantek.com from somewhere magic
# glennjones Yes I saw the white space that andy pointed out on http://microformats.org/wiki/microformats2-parsing-issues#When_to_collapse_whitespace_in_properties
# KevinMarks should I fix the examples?
# glennjones Not sure I agree that this is the right answer, but I will update the tests so they dont collapse whitespace and add to the page why I think it wrong
# glennjones If you want to changes the v2 tests great, I working on the v1 at the moment and will change them with my next push to repo
glennjones joined the channel
# glennjones Thats why paser APIs like https://pin13.net/mf2/ has a baseURL input when parsing HTML fragments
# KevinMarks right, but if so we need to add that to the tests somehow
# KevinMarks I sent a pull for the xfn and other rels if you want to review
joskar joined the channel
# glennjones So I need an interface mf2py that allows me to send a baseURL to make the test work. https://kylewm.com/services/mf2 does not, so I cannot send it the URL context to make the test work correctly
# KevinMarks you could change the test to say <base href="http://tantek.com"><link rel="updates alternate" type="application/atom+xml" href="updates.atom" />
# KevinMarks then kyle's thing works
# KevinMarks that's better than adding an extra way to pass the base URL in
# glennjones Thats an interesting idea, happy to add that to top of tests. Would like to stick to <base href="http://example.com"> if thats OK
# glennjones I known my parsers pick up <base> wonder if they all do!
# KevinMarks well, it will test them ;)
# KevinMarks just make sure the base matches what the json is expecting
# glennjones I will try it out on the v1 tests I am changing at the moment
joskar and chiui joined the channel
# KevinMarks this whitespace is ugly
# glennjones Not collapsing whitespace makes any value taken from a block of HTML unusable in the real world use case. Label in the http://testrunner-47055.onmodulus.net/test/microformats-v2/h-adr/geo/ test is an good example of something I have used in the past and will no longer be of any use to me.
# KevinMarks though actually label is supposed to have newlines in , per vcard
# KevinMarks "The property can also include a "LABEL" parameter to present a delivery address label for the address. Its value is a plain-text string representing the formatted address. Newlines are encoded as \n, as they are for property values."
# KevinMarks I have a pile of changes in a pull request for you, working backwards though my test cases for mf2py
# glennjones Great the more input the better the test will be for all us
# glennjones The issue I have with returning whitespace and returns from parsed HTML is that it was added to make the code understandable and was not meant to be used/read outside of that context.
# KevinMarks yes, especially in "name"
pfefferle and csarven joined the channel
# glennjones KevinMarks I have merged in the pull request. Got a little complex as I had changed the directory structure. Hopfully its all there now. I will push changes to test server in a minute
dariusdunlap_ and joskar_ joined the channel
# KevinMarks__ Glenn, have you seen http://johnmacfarlane.net/babelmark2/
# KevinMarks__ I think your test harness is close to the same thing for microformats
# glennjones Babelmark looks like a nice tool, interpreting a spec is hard, often comparing implementations helps find outstanding issues and allows you to coalesce around one understanding more quickly. Thats why I built it
pfefferle joined the channel
benborges and tantek joined the channel
# glennjones Maybe adding <base> to all the tests is overkill. We could keep it down just to 3 or 4 tests where relative URL resolving is been tested?
csarven joined the channel
# 
@souiszak "WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ : Root-013" http://website-homepage.com/wordpress/customize/microformats-hatom (twitter.com/_/status/603550591403597825)
# glennjones Maybe Kevin can look at providing an API with baseURL
elf-pavlik joined the channel
# 
@SEO_news2015 WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ | Root-013 http://website-homepage.com/wordpress/customize/microformats-hatom (twitter.com/_/status/603553997694308352)
# 
@taaboolove WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ | Root-013 http://website-homepage.com/wordpress/customize/microformats-hatom?utm_source=dlvr.it&utm_medium=twitter (twitter.com/_/status/603556434693861376)
# 
@nandemopr ã€æƒ…å ±ã€‘WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ | Root-013　http://xcontents.x.fc2.com/jump2.html?q=http%3A%2F%2Fwebsite-homepage.com%2Fwordpress%2Fcustomize%2Fmicroformats-hatom (twitter.com/_/status/603562564736299009)
# 
@bichiko_seo WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ | Root-013 http://website-homepage.com/wordpress/customize/microformats-hatom?utm_source=dlvr.it&utm_medium=twitter (twitter.com/_/status/603563585977987072)
# 
@akio0911_news WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ | Root-013 http://website-homepage.com/wordpress/customize/microformats-hatom?utm_source=dlvr.it&utm_medium=twitter (twitter.com/_/status/603563605921910784)
elf-pavlik and pfefferle joined the channel
KartikPrabhu joined the channel
# glennjones kylewm: Thanks a lot that would help us test all the parsers equal baseline
# 
kylewm ok I think it should work now via POST params https://kylewm.com/services/mf2
dariusdunlap and pfefferle joined the channel
# glennjones kylewm: that work fine added two new test just for relative URLs - will look at adding a testrunner mode that does not flag "rel-urls": {}, as an error, so you can get more out of the results
gRegorLove joined the channel
# KevinMarks__ Kylewm did you see the test running I added to my mf2py fork?
# KevinMarks__ Could do that, but with the main tests repository as the submodule
# 
@4SEO1 WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ | Root-013: æ§‹é€ åŒ–ãƒ‡ãƒ¼ã‚¿ï¼ˆhttp://microformats.org hAtom）を… http://4-seo.net/news/story/6902/ (twitter.com/_/status/603603516612079616)
KartikPrabhu joined the channel
# KevinMarks__ I fixed a lot of tests, but it is showing things that aren't implemented too
# KevinMarks__ See issues
# KevinMarks__ Tantek: should we revisit whitespace preservation?
# glennjones kylewm: half way through adding some to testrunner, but it would be much clearner if you added flag to your side to strip rel-urls
KevinMarks_ joined the channel
# KevinMarks_ Adding parameters is not a good idea
KevinMarks__ and Erkan_Yilmaz joined the channel
# KevinMarks__ If you want examples, look at the diffs I put in th tests last ni
# 
tantek KevinMarks,kylewm can you add the whitespace issue to http://microformats.org/wiki/microformats2-parsing-issues ? my suggested fix is going to be to "add trimming from start end" - pretty sure that will fix all current use-cases (and back compat)
# glennjones Kylewm: that’s done your now, I think Vendan: idea of using a “strict” parameter is more flexible and would allow you try out other changes without creating errors.
Erkan_Yilmaz joined the channel
# KevinMarks the rel-urls thing si that the tests reflect the parsing page, not the updated stuff
# KevinMarks so my test harness on mf2py currently deletes the rel-urls object before comparing
# KevinMarks wiht glenn's changes too, down to 37 failures in mf2py (from 80 last night)
glennjones_, eschnou and chiui joined the channel
# kevin marks edited /microformats2-parsing-issues (+2836) "/* whitespace collapsing revisited */" (view diff)
# KevinMarks tantek, issue posted
# KevinMarks__ Glennjones did I capture your issue too?
zappl joined the channel
# glennjones KevinMarks: I am going to pick up adding issues and ideas to wiki first thing tomorrow
# KevinMarks OK. I'll have another look at what mf2py is doing later on
# tantek edited /microformats2-parsing-issues (+951) "/* whitespace collapsing revisited */ add option to trim leading/trailing, separate choose from vs. orthogonal choices, add -1/0/+1" (view diff)
# KevinMarks leading/trailing isn't the issue I've seen
# KevinMarks it's the stuff inbetween spanning properties
# KevinMarks right, so they're invisible metadata
# tantek edited /microformats2-parsing-issues (+304) "/* implied properties when an explicit class is provided */ note old consensus, new proposal based on new use-case" (view diff)
# KevinMarks 80:20 they're invisible
# KevinMarks you could use <pre>
tantek joined the channel
# KartikPrabhu collapsing whitespace in name is going to completely mess up the classification of notes. I thought name = content.value implied that the post was a note
# tantek the reason to preserve whitespace in notes - e.g. p-content is for other reasons: http://indiewebcamp.com/note#Whitespace
# KartikPrabhu yeah saw that
# KartikPrabhu but if mf parsers remove whitespace from the name property then this https://indiewebcamp.com/notes#Note_Type_Algorithm will get a whole lot messier
# KevinMarks add comments to the wiki, KartikPrabhu
# KevinMarks value also seems to accrete whitespace
# KevinMarks I proposed removing it from value and name
# KevinMarks at least I thought I did
# KevinMarks looking at the places where the tests differed from the process
elf-pavlik, dariusdunlap, KevinMarks_, KartikPrabhu and KevinMarks__ joined the channel
# 
@kimura_macoto 2件ã®ã‚³ãƒ¡ãƒ³ãƒˆ http://b.hatena.ne.jp/entry/website-homepage.com/wordpress/customize/microformats-hatom#tw?u=kimu12 “WordPressã®æ§‹é€ 化データ（http://microformats.org hAtom）マークアップ | Root-013” http://website-homepage.com/wordpress/customize/microformats-hatom (twitter.com/_/status/603702115790258176)