freenode :: #microformats

2 Jul 2015
00:07Loqi[[to-do]] * Tantek * (+305) start microformats2 updates with pages that need updating to microformats2
00:10Loqi[@lubiana] Dafür habe ich jetzt voll viel über microformats und semantische-tags gelernt. (
00:20Loqi[[Main Page]] * Tantek * (+71) update wiki home page for microformats2 since they've been stable for over a year, simplify some of the text before the specs, separate rel microformats to reduce confusion, classic microformats
00:22Loqi[[Main Page]] * Tantek * (-75) /* Get Started */ simplify wording
00:22Loqi[[Main Page]] * Tantek * (+7) /* Classic Microformats */ classic
00:24Loqi[[microformats2]] * Tantek * (+33) note lessons learned from other syntaxes
10:12Loqi[@mfhepp] @kevinmarks @danbri @denials Anyway, I do not think we can resolve the dispute via twitter, & personally I think #microformats are history (
16:37csarvenWhat is the expected value for p-author?
16:38csarvenCan it contain multiple values?
16:38csarven"Alice Bob Carl"
16:38csarvenOr will that be treated as one item (I presume this is the case)
16:39gRegorLoveIt would be treated as one author, yes. You would want multiple p-author for multiple authors
16:40aaronpkevery property can have multiple values which is why in the parsed JSON everything is an array
16:41gRegorLoveI embed my h-card when I publish p-author on my site. I think that's common, but it's not required
16:41aaronpkif you have multiple p-author elements, the parsed result will be like "author":["Alice","Bob","Carl"]
16:41ben_thatmustbememultiple h-cards
16:42ben_thatmustbemewith p-author, or would you do p-author and multiple h-cards beneath that?
16:42csarvenIn the case of h-cite, would it be appropriate to use p-author along with p-publication in the same class?
16:42gRegorLoveOne p-author, multiple child h-cards has a nice spot to plug in html and see what it parses as
16:44csarvenHmm, no, p-publication and p-author shouldn't go together.
16:44gRegorLoveI'm not familiar with p-publication
16:50csarvenCan mf2 properties chain its values to its descendent elements e.g., <span class=p-author><span class=h-card/></span> ? Or do they look for the values from their text nodes?
16:52aaronpki believe in that case the h-card appears as a child element rather than as the &quot;author&quot; property
16:52csarvenre: multiple authors, I think I&#39;m stuck with either using p-author incorrectly or drop its use. I&#39;m getting data from an API which returns random text for the author field. Can&#39;t be bothered to clean/normalize that
16:55csarvenSo, if this wasn&#39;t already discussed, if mf2 (or mf3?) allowed chaining, there&#39;d be an ever clear way to map mf<->RDFa. That&#39;s not a justification for it to exist in any way.
16:57aaronpkwhat do you mean chaining?
16:57csarvenSorry, I should be more clear here. I was thinking of things at once.
16:57csarven^two things at once.
16:57aaronpkthere&#39;s quite a bit of chaining going on here
16:59Loqi[@Shelia_Kay] @thatjohn #ExedeSucks (
17:00csarven1. possibility to complete attribute-value declarations through child elements 2. and the possibility to describe a new entity through the value of its parent.
17:02csarvenboth appear to be possible, with the exception that every entity is sort of like a bnode
17:03KevinMarks_every entity can have a url - in that example they all do I think
17:03csarvenAs I don&#39;t think something like this makes sense: <s class=&quot;u-uid&quot;><s class=&quot;hcard/></s>
17:04csarvenHaving a url is a property of that thing.
17:04csarvenUnless mf2 treats url as the way to identity things
17:04csarvenor uid or hwatever
17:05KevinMarks_yes, urls identify things in microfromats
17:06csarvenAnd if you don&#39;t have an url, the thing you&#39;ve described can&#39;t be identified? (URI is to &quot;identify&quot;, URL is to &quot;locate&quot;. Aside: In Tantek&#39;s beautiful words that I&#39;d like reuse: this is &quot;architecture astronomy talk&quot;)
17:07gRegorLoveI was about to ask what the use-case is, to better grasp what you&#39;re referring to. :)
17:07KevinMarks_you&#39;re free to append a # to the URLs to satisfy your pedantry
17:08gRegorLoveHere&#39;s a real-world example of my p-author with h-card within it, if it helps. h-* always parses as a child element.
17:10csarvenIrrelevant at this point. mf doesn&#39;t acknowledge URIs. In any case, the question remains re: &quot;urls identify things in microfromats&quot;, in the absence of an url, can a thing be identified?
17:11csarvengRegorLove I&#39;m merely asking about what mf2 is doing. Don&#39;t need use-cases to explain myself :)
17:11gRegorLoveSorry, just having a hard time understanding. I&#39;m no mf pro, though
17:12gRegorLoveIdentified how?
17:12csarvengRegorLove That&#39;s a good example thanks
17:13gRegorLoveYou&#39;re welcome
17:13csarvenrhiaro url/uid/xyz are all properties. KevinMarks_ &#39;s statement sounds like if url&#39;s are present, that would be the way to identify that thing. I&#39;m asking what do you call &quot;that&quot; which has no url. Same applies for uid or any other property.
17:14KevinMarks_no, you need to explain *why* you want to do that
17:14csarvengRegorLove &quot;value&quot;: &quot;http:\/\/\/assets\/img\/profile.jpg\r\n\t\t\tby gRegor Morrill&quot; is not clear to me.
17:14csarvenDude, I&#39;m asking what mf2 is doing. I don&#39;t have to prove anything to you!
17:15csarvenIt is a simple question.
17:16ben_thatmustbemecsarven: the default in many of these groups is &quot;real world example&quot; otherwise its just doing theoretical exercises that may be pointless
17:17gRegorLovecsarven: I think the php-mf2 parser used in my example is not correct. It should be pulling the p-name from the h-card child, &quot;gRegor Morrill&quot; in this instance.
17:17gRegorLoveThe issue&#39;s been raised and hopefully php-mf2 will be patched soon for that example.
17:18KevinMarksoops, fell off
17:18ben_thatmustbemeh-entries in mf2 could have no u-url, I don&#39;t see any problem with that.
17:18* ben_thatmustbeme comes in to the conversation a bit late
17:18KevinMarksare you trying to link to the mf2 object?
17:18csarvenI&#39;m asking about how something works or is called. Do you or do you not have a name for something which has no url? (As it was stated that &quot;yes, urls identify things in microfromats&quot;. I&#39;m asking about the opposite.)
17:18KevinMarksare you trying to address part of the parsed data?
17:19csarvenben_thatmustbeme Okay, in that example, there is no way to identify that h-entry correct? Other than perhaps its location in the document?
17:19ben_thatmustbemeas far as I know there is no name for it in mf2 world, its just parsed data at that point.
17:19csarvenCool, thanks!
17:20ben_thatmustbemecsarven: nope, though you could possibly use an ID but thats outside of MF2 parsing
17:20csarvenIn the RDF world, something like this is called along the lines of a bnode. I was looking for a similar term or treatment from the mf side.
17:20ben_thatmustbememf2 parsers just parse &quot;a document&quot; that could be fetched by parsing the HTML for a fragment
17:20KevinMarksright, you could link to the h-entry if it has an id
17:20KevinMarksor you could use a fragmention link to its text
17:21ben_thatmustbemeyou can use any external method you wish actually,
17:21ben_thatmustbemethats entirely outside of MF2 land
17:21csarvenCool, got it!
17:21KevinMarksor you could parse the page and get the nth h-entry
17:24aaronpkinteresting that the concept doesn&#39;t have a name in mf2, most likely because it doesn&#39;t matter
17:25csarvenThat&#39;s totally fine. I was only trying to determine whether it exists or not.
17:25KevinMarksis there a CSS selector for nth-of-class ?
17:25ben_thatmustbemei think so, there is in jquery
17:26gRegorLoveBeat me to it
17:27csarven!tell tantek I&#39;ve dropped h-listings. Using h-cite now. Defaulting to &quot;broken&quot; way of using p-author in h-cite because I am lazy (as a developer) to clean/normalize its value where it is retrieved from an external resource, and it contains multiple authors along with some other random text.
17:27LoqiOk, I&#39;ll tell him that when I see him next
17:27KevinMarksno, that&#39;s by type
17:28aaronpkin RDF-land what is the opposite of a bnode?
17:28aaronpkweb page : email :: IRI : bnode
17:29csarvenIRI or bnode
17:29csarvenwebpage/email are IRIs
17:30aaronpkoh, i was trying to make a joke, but failed because I don&#39;t have a word for a bnode with an address
17:30csarvenNot sure why you are sarcastic
17:30aaronpknot sure where you&#39;re seeing the sarcasm there. normally i have to work pretty hard for my typed text to be read as sarcastic
17:31* csarven is confused
17:31KevinMarksoh, that does work gRegorLove
17:31KevinMarksnot clear from the docs
17:31csarven*[class] probably
17:32aaronpkI am writing a pretty lengthy email, which should probably live as a web page somewhere, and then realized that this email is like a web page with no URL
17:32csarvenIt has Message-ID
17:33KevinMarks!tell tantek shoudl show that you can address classes as well as types (or is that not generally true?)
17:33LoqiOk, I&#39;ll tell him that when I see him next
17:34csarven*[class]:nth-of-type(2) should work
17:35Loqi[@kevinmarks] @anshublog twitter is email with no to: field (I think that is Sergey Brin&#39;s description) (
17:42KevinMarksit works for me in chrome and safari:
17:42KevinMarksand firefox
17:43KevinMarksthough stack overflow has several answers saying it doesn&#39;t dating from 2013
17:44KevinMarksso csarven, a natural way of addressing microformats within a page is by css class selector
17:44KevinMarksif you want to point to them structurally
17:48csarvenCould just as well be XPath
17:50ben_thatmustbemethough generally if you are pulling something within a feed i&#39;d recommend against using nth-of-type as thats will change as soon as anything is added
17:50* ben_thatmustbeme wonders how much browsers respect the &quot;only one of an ID&quot; and if [id]:nth-of-type would work
17:51csarvenIn a way XPath is actually a lot more powerful to refer/address/name/point to things in a document as it can traverse or look for certain conditions.
17:51KevinMarksright, tantek uses files with long collections of posts queried with xpath as the backend to his site
17:52ben_thatmustbemebi-monthly storage files
17:52KevinMarkswith month defined using his personal modified calendar
17:59csarvenHow would one use u-uid for the ISBN?
18:00KevinMarksfor u-uid it would need to be URL
18:00KevinMarksp-uid if it&#39;s an inline value
18:00csarvenDidn&#39;t know p-uid existed
18:01KevinMarksu- and p- are parsing directives
18:01aaronpkin mf2, the p- u- e- dt- prefixes are *only* a parsing indicator, not part of the vocabulary
18:01aaronpkthat&#39;s why they don&#39;t appear in the parsed result
18:01ben_thatmustbemestill not clear on exactly what e- is for actually
18:01KevinMarksthough h- does
18:01ben_thatmustbemekeep forgetting to look at that
18:01KevinMarkse- means &#39;the html contents of this element&#39;
18:01KevinMarksuseful for h-entry, for example
18:02ben_thatmustbemee-content is the only place i ever saw it
18:02aaronpk&#39;e-*&#39; for embedded markup properties
18:03KevinMarkslosing the prefixes on parsing does mean it&#39;s harder to round-trip
18:03KevinMarksif something could a url or text, your consuming code has to sniff it
18:03ben_thatmustbemei don&#39;t think it makes sense to round-trip though, its not 1-to-1 so it doesn&#39;t really matter
18:04KevinMarkswell, also if I&#39;m trying to construct a page from the parsed output, I have to assume a type
18:04aaronpkchances are you also have to know about the vocabulary, so that&#39;s fine
18:04ben_thatmustbemeso long as you can generate some html that will parse back to the same results, there are infinitely many options
18:04aaronpkfor example, comments-presentation
18:05csarvenI think it might be confusing to tell in situations where an URL is used as p-url
18:05aaronpk<span class=&quot;p-url&quot;></span>
18:05ben_thatmustbemeunlinked url but still contains url as text
18:06csarvennot sure what the difference is in the end.
18:06csarvenor how it will/should be treated
18:06aaronpkno difference in the end, shows up as &quot;url&quot;:&quot;;
18:06csarvenp- is more like for displaying (for the human)
18:07KevinMarksurl is clear enough
18:07KevinMarkswe&#39;ve defined that as a general property like name
18:08csarvenaaronpk Right. I was giving an example to KevinMarks&#39; point. p-url will give you url, but going back to p-url or u-url is not clear. Actually this holds true for all properties.
18:08aaronpkright. and it doesn&#39;t matter
18:08KevinMarkswell, it can matter - I got this when trying to make the indiecards
18:08csarvenThe preference/default for some properties may be that &quot;okay, if you see url, go with u-url&quot;
18:08aaronpkthe way we&#39;ve been using it, we go from parsed result back to HTML when doing things like displaying reply context or comments
18:09aaronpkin which case we know abotu the specific vocabulary of h-entry, so we can make assumptions about how to use the properties
18:09ben_thatmustbemethe only difference between <span class=&quot;p-url&quot;></span> and <span class=&quot;u-url&quot; href=&quot;;></span> will be the value of Name
18:09ben_thatmustbemeso really u-url has more fidelity
18:10csarvenRight. It just illustrates that it is a case by case or whatever the preference may be for the consumer/publisher.
18:11ben_thatmustbemeyes. well, publisher, the consumer parses it by the parser
18:14KevinMarksaaronpk: I had to do
18:14csarvenSo, when an h- type doesn&#39;t have a property does that mean that it is disallowed or can one use any property as they like?
18:15csarvenh-cite for example mentions u-uid but not p-uid.
18:16csarvenIn my case (and I imagine others do something similar here), I use p-uid for the ISBN of an h-cite
18:16KevinMarksbecause benwerd&#39;s h-card has an html note
18:16KevinMarks(this is the &#39;writing code in templates&#39; problem I was talking about last night
18:18KevinMarksi should probably move that logic to the python side
18:26ben_thatmustbemep-uid of an isbn would make sense i think
18:37csarvenben_thatmustbeme Yea, I think so too. ISBN is a string any way
18:39ben_thatmustbemei believe there was an isbn class for MF1
18:40csarvenp-isbn would be a sub-property of p-uid
18:44csarvenSo, p-isbn p-issn .. would be a more appropriate than p-uid, especially in the case of h-cite
18:44csarvenPlenty of examples listed here
18:47KevinMarksright, the isbn is a more specific identifier than uid
18:50csarvenI am going to go ahead and use p-isbn. See you all on the other side!
21:00Loqi[@Shelia_Kay] @exede @NHRA Read before you sign a contract. Worst ISP ever. (
21:14jgaytantek, any chance you could make rel-jslicense to the existing-rel-values list? More sites have started to use it in the past year and more sites and frameworsk ahve complained to us about it not passing w3c validation
21:15tantekjgay - is there a spec for it? even just a wiki page would be a good start
21:15Loqitantek: csarven left you a message 3 hours, 47 minutes ago: I&#39;ve dropped h-listings. Using h-cite now. Defaulting to &quot;broken&quot; way of using p-author in h-cite because I am lazy (as a developer) to clean/normalize its value where it is retrieved from an external resource, and it contains multiple authors along with some other random text.
21:15Loqitantek: KevinMarks left you a message 3 hours, 42 minutes ago: shoudl show that you can address classes as well as types (or is that not generally true?)
21:15jgaytantek, you mean <>?
21:16tantekhah! thanks :)
21:31tantekdrat - was just trying to help with rel-jslicense edits
21:36Loqi[[rel-jslicense]] * Tantek * (+1058) note no current issues, verify EFF live example in wild, note research of previous examples / other techniques used for indicating js license, mv Consumers to Implementations
21:37tantek!tell jgay I tried to improve the rel-jslicense spec a bit, preserving all the intent/meaning I think you put in, with additional explicit prose / description for publishers / consuming code.
21:37LoqiOk, I&#39;ll tell them that when I see them next
21:37tantek!tell jgay you mentioned &quot;More sites have started to use it in the past year&quot; re: rel-jslicense - could you links to their pages that use rel=jslicense to: ? Thanks!
21:37LoqiOk, I&#39;ll tell them that when I see them next
21:38tantekbased on there being a written spec, real world usage of rel=jslicense by someone other than the author, and presence of real world consuming code, I definitely think its worthy of being added to existing-rel-values for HTML5 additions
21:39tantekanyone object to rel-jslicense? if you could document your issues at that would be appreciated. seeing no other input I&#39;m going to add it to existing-rel-values HTML5 values - though jgay could have done that himself too!
21:43Loqi[[rel-jslicense]] * Tantek * (+81) /* Examples in the wild */ add, note use on a href
21:49Loqi[[existing-rel-values]] * Tantek * (-102) move js-license from brainstorming to HTML5 proposed additions, could have been here before, but certainly now with real world publishing examples and consuming code!
22:08Loqi[@microformats] Thanks @joshuagay for making rel=jslicense happen! use it to link from HTML to license for scripts (
22:11Loqi[@wbwdt] RT @microformats: Thanks @joshuagay for making rel=jslicense happen! use it to link from HTML to license for scripts (
23:00Loqi[@pinceladasdaweb] RT @microformats: Thanks @joshuagay for making rel=jslicense happen! use it to link from HTML to license for scripts (
2 Jul 2015
Last message: 6 minutes and 3 seconds ago