#microformats 2015-06-02

2015-06-02 UTC
#
tantek also, KevinMarks, there is nothing ambigous about it, you explicitly listed rel-alternate in the proposal for rel-urls, therefore it got included. I understand now that is not exactly as you intended, with the additional impact on "rels" etc. - but that was not what you wrote down. hence the need to write these things down.
#
KevinMarks it is clearly ambiguous as you and I interpreted it differently
#
tantek I'm not sure what's ambiguous about the explicit inclusion of alternate
#
tantek I agree that's apparently not what you meant
#
tantek btw using alternate as an example helped make the case for rel-urls, FWIW
#
KevinMarks the ambiguity is you omitting it from rels but including in rel-urls
#
tantek there was no change to "rels" specified in your brainstorm
#
tantek whereas alternate was listed as one of the examples to include in rel-urls
#
tantek again, no ambiguity. inconsistency, sure, but neither I nor anyone else noticed it til now
#
tantek edited /microformats2-parsing-issues (+260) "/* drop alternates collection and include them in rels */ add discussion with a +1 to start it off" (view diff)
#
tantek KevinMarks, assuming this issue captures your intent, could you comment? http://microformats.org/wiki/microformats2-parsing-issues#drop_alternates_collection_and_include_them_in_rels
#
tantek plus you should feel free to implement and cite that proposed issue
#
tantek !tell tommorris, kylewm, aaronpk your comments would be welcome here too: http://microformats.org/wiki/microformats2-parsing-issues#drop_alternates_collection_and_include_them_in_rels
#
Loqi Ok, I'll tell them that when I see them next
#
tantek !tell glennjones note that we have a further fix to rels based on adding rel-urls and the clarification you requested this morning, please comment: http://microformats.org/wiki/microformats2-parsing-issues#drop_alternates_collection_and_include_them_in_rels
#
Loqi Ok, I'll tell them that when I see them next
#
kevin marks edited /microformats2-parsing-issues (+278) "/* drop alternates collection and include them in rels */ +1" (view diff)
#
tantek cool - assuming we get more consensus that just the editor (myself) and an implementer (go ahead and patch mf2py KevinMarks), then I'll add it to the spec
#
tantek s/that just/than just
#
Loqi tantek meant to say: cool - assuming we get more consensus than just the editor (myself) and an implementer (go ahead and patch mf2py KevinMarks), then I'll add it to the spec
#
kylewm edited /microformats2-parsing-issues (+138) "/* drop alternates collection and include them in rels */ +1" (view diff)
#
kylewm edited /microformats2-parsing-issues (+48) "/* drop alternates collection and include them in rels */ would help if i signed it" (view diff)
#
KevinMarks kyle, do you want to do the testcase parsing thing that I have in my fork in the main mf2py?
#
KevinMarks only pointing to the mainline tests instead of my fork?
#
kylewm KevinMarks: yes, but I don't know whether it's a good idea to run them as part of the current test suite, or to invoke them separately
#
Loqi kylewm: tantek left you a message 17 minutes ago: your comments would be welcome here too: http://microformats.org/wiki/microformats2-parsing-issues#drop_alternates_collection_and_include_them_in_rels
#
kylewm and also I'd like to fix the directory names so they are not "test" and "tests"
#
KevinMarks heh
#
KevinMarks that is the name of the repository in microformats
#
KevinMarks I'm getting whitespace mismatch errors in test_embedded_parsing and test_hoisting_nested_hcard
#
kylewm also I tend to agree with ben_thatmustbeme that submodules are a pain
#
KevinMarks is there a better way to depend on the other repository?
#
KevinMarks being able to do "git submodule update --remote; nosetests" is handy
#
kylewm something feels slightly off i guess... should the version of the test suite be dependent on the version of mf2py, or vice versa
#
kylewm probably it should... at least on a macro level.
#
kylewm but like if the tests have something slightly wrong in them, and it gets fixed, it's a pain to have to make a new commit to mf2py to update the submodule
#
kylewm strawman alternative: have the test harness download the test files before running them
#
ben_thatmustbeme re submodules, the biggest problem is how other people mismanage them
#
ben_thatmustbeme forgetting to push submodules, etc
#
ben_thatmustbeme or rather not realizing they editted a submodule
#
KevinMarks I'm avoiding thta by not editing the submodule
#
ben_thatmustbeme that is one option, haha, but then if you want to update submodules its annoying. 'git submodule udpate --init --recursive' only does a checkout to whatever version of the submodule it was committed with. to pull all the latest libs is a bit annoying
#
ben_thatmustbeme worse, i have had people commit and push the submodule ONLY not realizing they had changed directory to inside the submodule
#
ben_thatmustbeme so their changes never actually go anywhere
#
ben_thatmustbeme on the main code anyway
#
KevinMarks hm
KartikPrabhu joined the channel
#
KevinMarks kyle, did you do some whitespace changes?
#
kylewm no?
#
kylewm wait, if you mean change the whitespace in the source code, i did slightly https://github.com/tommorris/mf2py/commit/c1dd21cce6573fd497adb9b471a7ac90b8619b52
#
KevinMarks no, I'm seeing whitespace missing in test_embedded_parsing
#
KevinMarks and .test_hoisting_nested_hcard
#
KevinMarks it's stripping the spaces after the newline
#
kylewm I did not intentionally change any whitespace parsing
#
KevinMarks is this another 'which html parser do we have this time' problem?
#
kylewm very likely
#
KevinMarks OK, updated to put the alternaets in the rels and rel-urls
#
KevinMarks http://www.unmung.com/?html=%3Ca+rel%3D%22author%22+href%3D%22http%3A%2F%2Fexample.com%2Fa%22%3Eauthor+a%3C%2Fa%3E%0D%0A%3Ca+rel%3D%22author%22+href%3D%22http%3A%2F%2Fexample.com%2Fb%22%3Eauthor+b%3C%2Fa%3E%0D%0A%3Ca+rel%3D%22in-reply-to%22+href%3D%22http%3A%2F%2Fexample.com%2F1%22%3Epost+1%3C%2Fa%3E%0D%0A%3Ca+rel%3D%22in-reply-to%22+href%3D%22http%3A%2F%2Fexample.com%2F2%22%3Epost+2%3C%2Fa%3E%0D%0A%3Ca+rel%3D%22alternate+ho
#
kevin marks edited /microformats2-parsing-issues (+791) "/* drop alternates collection and include them in rels */ implemented in mf2py" (view diff)
voxpelli, Phae, benward, Garbee, bret, rknLA, KevinMarks___ and philipashlock joined the channel
#
kylewm KevinMarks: what would should happen if two rels point to the same URL with different properties?
#
kylewm http://www.unmung.com/?html=%3Clink+rel%3D%22alternate%22+href%3D%22http%3A%2F%2Fexample.com%2Ffr%22+type%3D%22application%2Fxml%22%3E+and+%3Clink+rel%3D%22alternate%22+href%3D%22http%3A%2F%2Fexample.com%2Ffr%22+type%3D%22text%2Fhtml%22%3E&pretty=on
#
kylewm rel-urls is lossy compared to alternates in that case, is that bad?
KevinMarks_, KevinMarks__, andicascadesf and tantek joined the channel
#
@flylinux 最近看到了一個流行的詞彙『微格(Microformats)』，就到微格 http://microformats.org/ 的網站看了一下，布丁在microformats & semantic web... http://job.achi.idv.tw/2010/04/13/%e9%97%9c%e6%96%bc%e5%be%ae%e6%a0%bc/ (twitter.com/_/status/605573743277142016)
fuzzyhorns joined the channel
#
tantek !tell kylewm re: "if two rels point to the same URL with different properties" - it's a bit odd, but the union of the rels is what applies so it's ok to do so. as far as which text label or type applies, we could pick either one.
#
Loqi Ok, I'll tell them that when I see them next
andicascadesf, KevinMarks___, dariusdunlap, eschnou, gRegorLove, KevinMarks_, Left_Turn, pfefferle, kez, glennjones, adactio, pfefferle_ and benborges joined the channel
#
KevinMarks_ Glennjones do you think we can replace the magic alternates with rel-urls?
#
glennjones We could do, but what about backwards compatibility, not sure who if anyone is using “alternates” data from the output of our parsers?
#
Loqi glennjones: tantek left you a message 8 hours, 38 minutes ago: note that we have a further fix to rels based on adding rel-urls and the clarification you requested this morning, please comment: http://microformats.org/wiki/microformats2-parsing-issues#drop_alternates_collection_and_include_them_in_rels
#
glennjones KevinMarks_: just reading through last nights chat to get context
pfefferle, KevinMarks___ and elf-pavlik joined the channel
#
glennjones KevinMarks_: yes we should drop “alternates” if we are happy no ones using it, but techically it is a backwards compatibility break
#
glennjones KevinMarks_: did you research if URL strings can contain chars that are not valid in JSON keys while coming up with structure?
Mark87_, gRegorLove, KartikPrabhu, pfefferle, glennjones, csarven and Garbee joined the channel
#
dating edited /gender-examples (+85) (view diff)
#
dating edited /gender-examples (-2) (view diff)
pfefferle, TallTed, fuzzyhor_, glennjones, KevinMarks_, eschnou and kez_ joined the channel
#
glennjones !tell tantek, KevinMarks is the “text” property of rel-urls a string or a array of string like the other properties? I think it should be an array for consistance
#
Loqi Ok, I'll tell them that when I see them next
tantek, glennjones, gRegorLove, eschnou, andicascadesf and KevinMarks joined the channel
#
KevinMarks hm, what chars are not valid in JSON keys?
#
Loqi KevinMarks: glennjones left you a message 3 hours, 12 minutes ago: is the “text” property of rel-urls a string or a array of string like the other properties? I think it should be an array for consistance
tantek joined the channel
#
KevinMarks !tell glennjones, tantek said text was OK as a single string, but kyle's issue of pointing to the same URL twice may suggest we need an array. Are there in the wild exampels fo linking to the same URL wiht multiple types? I think that is contrived, but linking to the same URL with multiple texts is quite plausible
#
Loqi Ok, I'll tell them that when I see them next
#
tantek KevinMarks: until we see real world examples and use-cases for that, I'd rather treat it either as first wins or last wins
#
Loqi tantek: glennjones left you a message 3 hours, 14 minutes ago: is the “text” property of rel-urls a string or a array of string like the other properties? I think it should be an array for consistance
#
Loqi tantek: KevinMarks left you a message 42 seconds ago: said text was OK as a single string, but kyle's issue of pointing to the same URL twice may suggest we need an array. Are there in the wild exampels fo linking to the same URL wiht multiple types? I think that is contrived, but linking to the same URL with multiple texts is quite plausible
#
tantek simpler for parsers and code that uses results of a parse
#
tantek I disagree with array for these because there's no use-case for it, and multiple values there would likely indicate an authoring error instead
#
tantek KevinMarks: I'm leaning towards first wins
#
KevinMarks hm, looks like deduping would eb good too:
#
KevinMarks http://www.unmung.com/mf2?url=http://epeus.wordpress.com&html=&pretty=on
#
KevinMarks so many rels in a default wp install
#
@ngeenen Structured Data – Your Key to Search Success in 2014 http://www.seoandy.net/biz/microformats/ via @andykinsey @cleodancer - microformatting (twitter.com/_/status/605829613621645312)
#
tantek many rels is fine. same URL multiple times in multiple <link> elements is likely indicator of bugs
#
tantek js;dr http://www.seoandy.net/biz/microformats/ "Javascript is required. Please enable it before you are allowed to see this page. " seriously?
#
tantek hey SEOAndy - you might want to serve some non-JS fallback content, you know, for SEO.
#
gRegorLove Hah
#
tantek really tempted to @-reply him accordingly
#
@andykinsey RT @ngeenen: Structured Data – Your Key to Search Success in 2014 http://www.seoandy.net/biz/microformats/ via @andykinsey @cleodancer - microformatting (twitter.com/_/status/605831496348233728)
#
KevinMarks his images 404 too
#
KevinMarks o_O "Schema.org is currently the best way to implement microformats."
#
tantek KevinMarks: come now, that kind of SEO nonsense is not even worth eye-raising for, it's no surprise that the SEO hamsterwheel churns out such things.
#
KevinMarks http://www.unmung.com/mf2?url=http://epeus.wordpress.com&html=&pretty=on the algorithm we have is very repetitive for tags and authors
#
KevinMarks is there any value in having the same url in the rels multiple times?
#
KevinMarks and conversely having the same rels in the rel-urls multiple times?
#
KevinMarks if we'er doing that, having a list for text and title makes more sense
#
KevinMarks or should we dedupe the rels?
KartikPrabhu joined the channel
#
tantek we don't dedupe rels, we dedupe URLs
#
tantek since multiple <link href="url" rel="something"> on the same URL, mean *union* of the somethings
#
tantek and from the perspective of consuming code, that makes it simpler
#
tantek rather than having to worry that that same URL might be elsewhere in rels
#
tantek also, since URLs are the keys in rel-urls, they have to be unique within rel-urls
#
tantek thus the parser already has to do a union of the rels on multiple links to the same href
#
tantek and yes at that point, if the multiple links href="sameurl" have any overlap in rel values, then those rel values get collapsed as well, since they're just a set
#
KevinMarks right, so we need to say that
#
KevinMarks we are presenting them as a list, not a dictionary
#
KevinMarks so deduping needs to be explicit
#
tantek currently rel-urls is a dictionary
#
tantek not a list
#
tantek *rels* is a list
#
tantek rel-urls is a dictionary where URLs are the keys
#
KevinMarks yes
#
KevinMarks I meant the ones within rels
#
KevinMarks http://www.unmung.com/mf2?url=http%3A%2F%2Fma.tt&html=&pretty=on
#
KevinMarks "rels": { "profile": [ "http://gmpg.org/xfn/11" ], "category": [ "http://ma.tt/category/asides/", "http://ma.tt/category/asides/", "http://ma.tt/category/asides/", "http://ma.tt/category/asides/" ], "author": [ "http://ma.tt/author/saxmatt/", "http://ma.tt/author/saxmatt/", "http:
#
KevinMarks is preserving order of rels important
#
KevinMarks "http://ma.tt/category/asides/": { "rels": [ "category", "tag", "category", "tag", "category", "tag", "category", "tag" ], "text": "Asides" },
#
KevinMarks is probably unhelpful
#
KevinMarks but is knowing that it was rel="category tag" important vs rel="tag category"
#
tantek something got truncated on my end
#
tantek no re: rel="category tag" vs rel="tag category" - they are equivalent
#
tantek just like class names in class attribute - order is irrelevant, undependable
#
tantek they are sets
#
KevinMarks OK. so using a python set for them then serialising to a list is OK
#
KevinMarks python syntax is nice there list(set(["tag","category","tag"]))
#
@ruthburr Schema/microformats aren't as important as you might think for Quick Answers. @cshel #smx (twitter.com/_/status/605841872691396608)
#
tantek what is smx?
#
KevinMarks danny's conference
#
KevinMarks http://searchmarketingexpo.com/
#
tantek is it like a big snakeoil swap meet? or do they actually provide citations for all their "recommendations" and "best practices"?
#
KevinMarks a bit of both - Danny Sullivan is pretty knowledgeable, and he will have Google speakers etc
#
aaronpk tweetable phrases are more fun
#
Loqi aaronpk: tantek left you a message on 6/1 at 5:16pm: your comments would be welcome here too: http://microformats.org/wiki/microformats2-parsing-issues#drop_alternates_collection_and_include_them_in_rels
KartikPrabhu joined the channel
#
kevin marks edited /microformats2-parsing-issues (+3879) "/* deduping of rels */" (view diff)
#
KevinMarks !tell tantek I documented the rels deduping issue
#
Loqi Ok, I'll tell him that when I see him next
warehouse13, elux, KartikPrabhu, tantek, elf-pavlik, JonathanNeal, twisted`, Phae, tommorris and bret joined the channel
#
kevin marks edited /microformats2-parsing-issues (+627) "/* deduping of rels */ implementation to try" (view diff)
#
kevin marks edited /microformats2-parsing-issues (+7) "/* deduping of rels */ close div" (view diff)