#dev 2016-09-07

2016-09-07 UTC
KevinMarks joined the channel
#
miklb hmm. I wonder if cweiske's CLI micropub client could be wired to Alfred to post notes eventually
KevinMarks_, dkm, AngeloGladding and KevinMarks joined the channel
#
GWG I need to figure out how to map webmentions to the WordPress comment type.
KevinMarks_ and tantek joined the channel
#
tantek has anyone worked on synthesizing summaries from notes?
#
tantek i.e. like a synthetic title?
#
tantek just came across https://en.wikipedia.org/wiki/Do_not_go_gentle_into_that_good_night and saw this: "It has no title other than its first line ..."
#
GWG Classic poem?
#
tantek yes, but it was that sentence from wikipedia that caught my eye. I wonder if that's a pattern of other poems / notes
#
tantek where the name/title is implied from its first line/sentence
#
Zegnat I am not big on poetry, but I think most of e e Cummings' work is untitled. People refer to his poems by their first line as well.
#
Zegnat Referring to poems by their first line might just be common practice, the writer did not imply a title in those cases though
#
tantek Zegnat, that's my point. it's implied by people *other* than the author
#
tantek as a /reader might from a /note
#
Zegnat Ah, gotcha!
#
Zegnat I thought this was about people "synthesising" titles on their own notes. Which would have been a little silly
tantek, KevinMarks, cweiske and KevinMarks_ joined the channel
#
KevinMarks_ Songbooks often have an index by first line as well as by title
#
KevinMarks_ Tantek: previously http://www.kevinmarks.com/poemfragmentions.html
#
Loqi [Kevin Marks] Fragmentions for poets 2014-04-23
#
KevinMarks_ Which includes a link to Dylan Thomas reading that poem
#
tantek interesting that we found the same example
#
cweiske the main problem is that the paragraphs of most pages have no ids. this is unfortunately not going aways
#
tantek indeed, we are talking about different use-cases
#
cweiske and with text sources written in markdown or other plain text formats, there is no way to store paragraph ids in the source itself
#
tantek identification/reference vs. summary/display
#
www.jimtyhurst.com created /User:Www.jimtyhurst.com (+44) "Place holder" (view diff)
cmal and tantek joined the channel
#
KevinMarks_ That's exact what fragmentions are for,,
#
KevinMarks_ The problem is that ids are a bad bit of hidden metadata
#
www.jimtyhurst.com edited /Template:jimtyhurst (+112) "Thumbnail photo and link to my user page." (view diff)
#
cweiske but fragmentions are inherently ephemeral, while the IDs are always there
#
cweiske (on systems that support them of course)
tantek joined the channel
#
cweiske but that's the usual problem of ideal-world vs. reality
loicm_ and tantek joined the channel
#
cweiske aaronpk, I can't log into teacup.p3k.io with commentpara.de - "Unable to verify the sign-in attempt"
#
cweiske quill is more chatty on errors
#
cweiske might have to do with different me urls
#
cweiske.de edited /Micropub (+144) "/* IndieWeb Examples */ commentpara.de" (view diff)
tantek joined the channel
#
KevinMarks_ That makes no sense, cweiske
#
KevinMarks_ Fragmentions point to words
#
cweiske KevinMarks_, so when you want to link your quote to the source you use the full quote text as fragmention?
#
cweiske or just the first couple of words?
#
cweiske.de edited /Micropub (+68) "/* WordPress */" (view diff)
#
cweiske.de edited /Micropub (+122) "/* Clients */ add shpub" (view diff)
#
cweiske.de edited /User:Cweiske.de (+74) "add shpub" (view diff)
#
KevinMarks_ Full quote is better. That way the link intent is preserved even if the page changes
#
cweiske do you do that?
#
KevinMarks_ Some of the time.
#
KevinMarks_ A suggestion that a fragmention highlight the exact text and not the containing element has been made
#
petermolnar voxpelli, cweiske I'm taking the search conversation over here: I honestly thought of a dead simple regex search field that just does a regex over text files
#
Loqi petermolnar: GWG left you a message 2 weeks ago: I agree
#
petermolnar instead of fancy indexed things like Solr
#
cweiske you can do that of course
#
cweiske I wanted something better with features listed there: http://p.cweiske.de/118
#
cweiske especially excerpts were on my list, and relevance-ordered results
#
cweiske which elasticsearch/solr bring with them
#
cweiske regex wouldn't work for me since I don't have plain-text files but html only
#
cweiske and I have php-generated pages, too
#
cweiske for relevancy: i have different weights for title, headlines, domain and content text - see https://github.com/cweiske/phinde/blob/master/data/elasticsearch-mapping.json ("boost" field)
#
cweiske and it was a fun project
#
cweiske petermolnar, if you want i'll add your domain to my search index. then you can use my search engine, too. it supports site-specific search
#
petermolnar thank you very much for the offer, but I'm still just trying to figure out what ways to take with my site
#
voxpelli http://www.blevesearch.com/ is built with Hugo and uses itself for search
#
voxpelli DuckDuckGo doesn't do good site search?
#
cweiske I don't care. I don't want to depend on anyone
#
voxpelli you will always depend on your internet provider
#
voxpelli so there will always be dependence, there's just different levels of it
#
cweiske are we really back at this point?
#
voxpelli well, I'm rather having a constructive discussions about benefits and drawbacks of different solutions than just someone disregarding it for the single fact that it's not independent enough for them
#
petermolnar voxpelli yes, and we all depend on the DNS system
#
petermolnar but that is infrastructure, which we are doomed to depend on
#
petermolnar and I'm refusing to view Google as infrastructure
#
petermolnar because if I do, I'll need to accept the fact that we're in deep shit, where private companies are the actual infrastructure providers and that William Gibson's future is indeed about to happen
#
voxpelli not Google specifically, but that kind of search-engines are in their nature something that's suitable to not self-host
#
voxpelli s/suitable to not self-host/not suitable to self-host/
#
petermolnar I'm not convinced X search provider will show the "proper" results
#
voxpelli there is http://commoncrawl.org/, but it lacks coverage and frequent recrawls and is mainly meant to try out new algorithms with
#
petermolnar when I want explicit word in Google results, I'm having a hard time
#
petermolnar so my reasoning against depending on services like this is that you cannot rely on them
#
petermolnar I could say Google = non-deterministic results
#
petermolnar which, for me, is unacceptable
#
voxpelli at some point we will probably get Twitter-like search within the IndieWeb as well and that will also have to be more or less centralized (with more actors than in the current search space though, but still not something every hosts themselves)
#
voxpelli s/every/everyone/
#
cweiske voxpelli, you're talking about full web search while petermolnar and I only talk about site search
#
cweiske you're reasoning that full web search engines are better at site search, which we don't
#
voxpelli well, I'm going to head for lunch now
#
cweiske what? no, /I'm/ heading for lunch now.
#
petermolnar eh, 11:00 here, not lunch for me yet
ben_thatmustbeme joined the channel
#
GWG Good morning all.
#
cweiske GWG, i'd be grateful if you'd take action on https://github.com/snarfed/wordpress-micropub/issues/25#issuecomment-230833187
#
GWG cweiske, I am in a way.
#
GWG I am fiddling with the WordPress REST API infrastructure which could be used for Micropub.
cmal joined the channel
#
petermolnar GWG in theory, the REST api could be used ever for a static generator, but as far as I got with it, it's a bit twisted
#
GWG I implemented webmentions using it.
#
cweiske using the REST API to have the MP endpoint on a different server, or internally using the REST API?
#
GWG They implemented REST API infrastructure. So a bunch of functions to create custom APIs.
#
cweiske ah. I didn't read the "infrastructure" part in the first comment
#
KevinMarks_ A web search engine can give better results for a given site as it has the link graph of the rest of the web to inform it about that site
#
KevinMarks_ Also the linguistic analysis derived from the whole web
singpolyma, KevinMarks and miklb joined the channel
#
voxpelli thinks about the possibility of connecting his node.js micropub endpoint with a wordpress rest api publisher - hmm
#
miklb voxpelli interesting. I looked at a WP plugin that syncs WP with a jekyll github repo
#
voxpelli miklb: I'm thinking one could "just" replace the Jekyll formatting and GitHub pushing with a formatter/pusher for the WP APi or any other similar API – mostly a case of mapping everything from A to B (and in the future the reverse as well)
#
voxpelli could be fun to maybe do a proof of concept for that against some other API, like Twitter
#
miklb I showed someone from GitHub the jekyll-admin gem last week at a patchwork event and his first reaction was "eww, looks like WordPress" :)
#
voxpelli miklb: then show them Quill + Micropub and they're like: "yay, looks like Medium" ;)
#
Loqi does a happy dance!
cweiske, tantek and gRegorLove joined the channel
#
loqi.me created /sloph (+139) "prompted by aaronpk and dfn added by rhiaro" (view diff)
#
loqi.me created /slogd (+39) "prompted by tantek and dfn added by rhiaro" (view diff)
AngeloGladding joined the channel
#
vanderven.se martijn moved /2016/Dusseldorf/Introductions to /2016/Düsseldorf/Introductions
#
loqi.me created /XAMPP (+91) "prompted by tantek and dfn added by voxpelli" (view diff)
#
vanderven.se martijn edited /2016/Düsseldorf/Introductions (+168) "Add category, fix Düsseldorf spelling, add link to the IRC logs." (view diff)
#
loqi.me edited /XAMPP (+50) "/* See Also */ new section" (view diff)
KevinMarks_ and Zegnat joined the channel
#
cweiske the micropub page is too long. now that the official spec draft exists, should we remove the MP implementation docs from the wiki?
#
ben_thatmustbeme i think its fine, people don't always want to read the spec, better to have more information than less
#
ben_thatmustbeme and yes that page is long, but most of it is examples of posting and implementations list
#
voxpelli I think its fine as well and if anything it's probably the implementation that could be removed and refered to the spec instead, no?
#
ben_thatmustbeme on another note... ugh, css can only set fill on svgs if they are embedded in the html
#
voxpelli ben_thatmustbeme: yeah :/ there's a way to do it through <use> or something as well, but the support in older browsers is poor
#
ben_thatmustbeme trying to figure out the design for my new MP client
#
ben_thatmustbeme and was messing with some bits
#
voxpelli at work I decided to wrap all our SVG embedding up into a template and for now make that template just embed them inline, and then later maybe we could move to some more DRY mehcnaism when supports for that gets improved
#
cweiske now that was fast. phar is working
#
www.boffosocko.com uploaded /File:remote_participants_indieweb_saturday.PNG "Remote participants of IndieWebSummit 2016"
KevinMarks joined the channel
#
www.boffosocko.com edited /2016 (+84) "/* Photos */ remote participants" (view diff)
#
KevinMarks If you inline them, gzip should make the repetition less of a problem.
#
voxpelli KevinMarks: more the repetition across pages that I don't really like :/
#
voxpelli icons like the css would be easily cacheable
#
KevinMarks True, and as we Move to http2 the fetch overhead goes down
#
voxpelli svg use xlink:href would be ideal then :/
#
voxpelli another issue apart from compatibility though is that xlink:href doesn't respect cors :(
#
voxpelli if I remember correctly, had some weird experiences with that as well when I tried to use it
cmal joined the channel
#
ben_thatmustbeme successful posting from the new system \o/
#
ben_thatmustbeme s/system/app
#
cweiske it's so sad to have to report a but just after extracting known-0.9.2.tar.gz
#
cweiske s/but/bug/
#
cweiske Loqi, replace!
#
cweiske oooooooooh
#
cweiske X-Clacks-Overhead!
#
cweiske benwerd++
#
Loqi benwerd has 122 karma (1 in this channel)
#
gRegorLove What is shpub?
#
Loqi It looks like we don't have a page for "shpub" yet. Would you like to create it?
#
cweiske shpub is a command line micropub client
#
Loqi ok
#
loqi.me created /shpub (+71) "prompted by gRegorLove and dfn added by cweiske" (view diff)
#
cweiske.de edited /shpub (+114) (view diff)
#
gRegorLove cweiske++ very cool
#
Loqi cweiske has 53 karma (3 in this channel)
#
aaronpk cweiske: did you notice that this site has your PuSH hub listed? https://angular.jsnews.io/laravel-user-registration-with-email-activationlaravel-php-angularjs-html-jquery-js-2/feed/
#
cweiske hah. no.
#
aaronpk i just saw it in my switchboard logs
KevinMarks_ and cmal joined the channel
#
aaronpk lol https://varvy.com/pagespeed/wicked-fast.html
KevinMarks_, tantek, AngeloGladding and miklb joined the channel