#indieweb 2017-08-10

2017-08-10 UTC
amz3` and snarfed joined the channel
#
GWG
Evening
snarfed, eli_oat, fzylogic, amz3` and wolftune joined the channel
#
Loqi
[superfeedr] "? Universal Sign-In Button" by Chris Aldrich on 2017-08-10 http://boffosocko.com/2017/08/09/universal-sign-in-button-unicyclic/
#
aaronpk
hwc pdx is rolling
DanC_ joined the channel
#
GWG
Rolling?
#
Loqi
[superfeedr] "But just like Facebook has a HUGE cadre of geeky engineers who've been working for more than 10 years to make it work easily enough" by Chris Aldrich on 2017-08-10 http://stream.boffosocko.com/2017/but-just-like-facebook-has-a-huge-cadre-of-geeky
fzylogic, biquillo, snarfed, amz3`, KartikPrabhu, [miklb], waynr and leg joined the channel
#
Loqi
[superfeedr] "Tomorrow is the first of many DreamHost Website Club meetings! Inspired by Homebrew Website Club, but for our employees to spend company time working on" by Jonathan LaCour on 2017-08-10 https://cleverdevil.io/2017/tomorrow-is-the-first-of-many-dreamhost-website-club-meetings
eli_oat, wolftune, KartikPrabhu, sh4l_, [tantek], jjuran, [chrisaldrich], snarfed, loicm, amz3`, jihaisse, j12t, friedcell, barpthewire and Pierre-O joined the channel
#
[tantek]
Quiet night #indieweb.
#
[tantek]
I concur with GWG. Rolling?
adactio and jeremycherfas joined the channel
#
jeremycherfas
Good morning IndieWeb
#
Zegnat
Changing the topic again, assuming we are no longer live ;)
#
Zegnat
Good morning jeremycherfas, [tantek] :)
#
jeremycherfas
Morning zegnat
j12t joined the channel
#
[tantek]
Good morning Zegnat
webhat and biquillo joined the channel
#
Zegnat
[tantek], do you remember any dublincore-microformats talk? Thought I would ask before going on a search spree.
#
petermolnar
I was considering putting together examples of h-entry + hentry + schema + dc documents but it's a can of worms
#
petermolnar
schema, mostly
#
petermolnar
dc is straightforward, but weirdly documented
Pierre-O joined the channel
#
Zegnat
Asking because sl007 was at vHWC yesterday and according to him journalists use DC a lot
#
petermolnar
the existance of DC is why schema.org shouldn't be a thing
#
jeremycherfas
Zegnat: I’m not sure that journalists use DC a lot. I think some of the organisations that publish hjournalism use it. There’s a difference.
davidmead joined the channel
#
Zegnat
jeremycherfas, aah, could be.
#
Zegnat
I am also not sure why you specifically need mf to handle DC, when DC by itself is perfectly usable. But I am not super into DC, so, *shrug*
biquillo, j12t, gigitux, jjuran and [kevinmarks] joined the channel
#
[kevinmarks]
We looked at DC when designing some microformats
#
Loqi
[kevinmarks]: snarfed left you a message 2 days, 12 hours ago: ^^^ i'd love thoughts on this if you have any! (https://snarfed.org/indieweb-activitypub-bridge )
#
sebsel
what is delete-of?
#
Loqi
It looks like we don't have a page for "delete-of" yet. Would you like to create it?
#
sebsel
what is untag?
#
Loqi
untag is sending a request to delete a tag-of reply post on someones post https://indieweb.org/untag
#
sebsel
!tell tantek We discussed delete-of in the context of untag, so it is mentioned there. https://indieweb.org/untag
#
Loqi
Ok, I'll tell them that when I see them next
#
Zegnat
[kevinmarks] do you know if there is anything written down about it?
sl007 joined the channel
#
sl007
Hello IndieWeb
#
Zegnat
Hi sl007!
#
Zegnat
I was just asking about DublinCore for you :)
#
sl007
Hi Zegnat! Yes thankyou ! "I am also not sure why you specifically need mf to handle DC, when DC by itself is perfectly usable. But I am not super into DC, so, *shrug*" because I want do it with class and neither with RDF nor per "page" as in http://wiki.selfhtml.org/wiki/Websemantics/Dublin_Core
#
sl007
It should be per article and combinable with MF2
#
jjuran
what is DC?
#
Loqi
IndieWebCampDC 2016 is a 2016 IndieWebCamp being planned for Washington DC https://indieweb.org/DC
#
sebsel
jjuran in this context, it is DublinCore
#
sebsel
what is DublinCore?
#
Loqi
It looks like we don't have a page for "DublinCore" yet. Would you like to create it?
#
sl007
sebsel - the CORE vocabulary / convention between standard media metadata like IPTC (photos), XMP (photo, video, audio), MP3 tags
#
sebsel
I was asking to trigger Loqi, so if you repeat that starting with "DublinCore is ...", then Loqi picks it up and stores it in the wiki :)
#
Zegnat
http://dublincore.org/documents/dc-html/ is the document that applies here. But it is indeed al scoped to a resource and not to part of it.
mblaney joined the channel
#
sl007
Yes - exactly - that is my problem.
#
sebsel
I still don't understand what it is used for. What is your use case for putting it into Microformats?
#
sebsel
If the argument is "this is what journalists use", then great, let them use it. But why does it have to be put in Microformats2?
#
Zegnat
That conversation should probably move to either -dev or even #microformats :)
barpthewire joined the channel
#
Loqi
microformats process FAQ
j12t joined the channel
#
Zegnat
sl007 do you know any DC-HTML parsers? I am coming up empty. Which suggests to me that it isn’t being consumed currently anyway :/
#
petermolnar
sebsel DC is not just web, it's for anything, incl. library catalogues
#
Zegnat
Not sure what you want me to read there, sl007.
#
Zegnat
By no DC-HTML parsers I just mean that DC-HTML has been around for a long time, and I expect people have been using it, yet nobody ever started consuming it? So why worry about implementing a new standard for it at all, when the previous one never got consumed?
#
sl007
The most basic example in real world - 1) A photographer is at G20 summit. He fills in all the IPTC informations in the picture (much work).
#
sl007
2) Who is the author? Who are the persons on the picture? Who wrote this caption? What happens? When? Where? What is the transmission reference?
#
sl007
What strong typed category applies and what tags? Often ISO info but much info ...
#
sl007
Realy much.
#
sl007
Now the editor takes the picture and sometimes put the authorname below the picture and sometimes forgets this step. One sentence about the content appears.
#
sl007
This sentence is different in the context of the publication. Always.
#
sl007
There is no more information now because meta was removed from the picture.
#
sl007
There is no more information for the WEB ...
#
Zegnat
The meta data is only removed if someone makes the conscious decision to remove it. I think petermolnar keeps all his meta data in the photos he publishes, for example.
#
sl007
But CNN does not ...
#
petermolnar
Zegnat that part of the truth
#
petermolnar
most of the old, default libs - GD, PIL, etc - doesn't keep metadata when you, for example, resize
#
sl007
(not that petermolnar does not matter) - but the media companies are ignorant if there is no force to use it ...
#
Zegnat
They don’t, but that’s still a choice of tools
#
petermolnar
anything that uses imagemagick does
#
Zegnat
But that would be CNNs decision, and a decision a photographer agrees to when they supply the photos? If CNN decides to remove all the metadata, do you think you can get them to publish all that data in their HTML?
#
petermolnar
sl007 actually I've been planning to make a tool that would use various available deep learning apis to detect the source of an image on the web
#
sl007
After websub was specified and became a standard, CNN used it …
#
petermolnar
but it's harder than I anticipated at first
#
Zegnat
DC-HTML is a standard, do they use it?
#
petermolnar
sl007 the in-image metadata is a rathole to be honest
#
petermolnar
I spent a lot if time with it, IPTC should never have been created (come on, ASCII only for real????) and XMP was never properly tooled, as in there is no generic tool that can handle (read/write) it well apart from exiftool
#
Zegnat
There is no reason you can’t start documenting and working on a new microformat for e.g. photographs. You could definitely borrow from DC/IPTC for the property names. I am just trying to see if anyone, anywhere, is already marking up information like this in HTML. And am having a hard time finding it.
#
sl007
petermolnar agreed - Zegnat how could they per resource ?
#
Zegnat
sl007, I believe there are RDF and Schema mappings for DC. Saw a repo for that floating around on a WG. If they want to make the meta data available they can.
#
sl007
petermolnar - same here - see my contributions for exiftool
#
Loqi
sl007 has 6 karma
#
sl007
RDF is hated …
#
petermolnar
the only thing exiftool is missing is a cross-platform gui, with that, it could win a lot of photographers who are most probably a little shaky from command line utils
#
petermolnar
or a lightroom plugin at lest
#
sl007
petermolnar - Agreed. Need to work on this fork https://github.com/redaktor/exiftool.js to crack more encrypted makernotes ;)
#
Loqi
[redaktor] exiftool.js: A pure javascript implementation of exiftool
#
Loqi
[petermolnar] wp-extraexif
#
petermolnar
no, don't fork exiftool
#
petermolnar
don't reimplement it
#
petermolnar
Perl is fine, it's cross platform, and exiftool is HUGE
#
petermolnar
just use it
#
Zegnat
sl007, I think JSON-LD allows for DublinCore by default per spec, but I might be remembering wrong.
#
Loqi
The microformats process
j12t joined the channel
#
sl007
petermolnar - there was a reason for this fork (but I'll recontribute makernote changes) - redaktor, my IW-project is only JS and I wanted to decrypt some makernote infos like the physical temperature for Canon.
#
petermolnar
this is usually the reason for projects like this, but you won't be able to keep up with the source, see all the readability reimplementations vs the js one from mozilla
#
sl007
well yes JSON-LD is fine but not in the html or is there a way to rel it to an element ?
#
petermolnar
exiftool would make a brilliant shared lib and it's unfortunate it's not like that
#
jeremycherfas
zegnat, s1007, petermolanar Are you guys familiar with controlledvocabulary.com?
#
sl007
I see … The thing with the JS is that you only need a browser ...
#
petermolnar
and with perl, you only need the perl interpreter ;)
#
Zegnat
jeremycherfas, I am not, will have a look.
#
Zegnat
sl007, IIRC JSON-LD can be scoped to only describe a specific resource, so yes, it could describe specific images within an article.
#
Zegnat
sknebel did you read up on LinkedData? Comment?
#
jeremycherfas
They used to maintain a spreadsheet of, for example, which online storage removed exif (never mind IPTC!) and the people there are very knowledgeable on metadata for images.
#
petermolnar
I'm not familiar with that jeremycherfas
#
jeremycherfas
You might find it interesting. When I used to work with a lot of images progessionally I tried to implement some of the ideas about nested hierarchies for tagging etc.
#
jeremycherfas
And to make sure basic image creator data were in place and correct.
#
jeremycherfas
Glad to be of service :)
#
sknebel
Zegnat: JSON-LD afaik in practice has to be a blob, unconnected. microdata can be using similarly to mf2, but I'm not sure what the consumer side looks like (afaik Google is currently recommending JSON-LD primarily)
#
sknebel
I think I've seen DublinCore RDFa a few times? again, not sure about consumers
#
Zegnat
Hmm, I seemed to recall a JSON-LD island could be scoped to a specific resource, e.g. the URL for a photograph.
#
Zegnat
Yeah, but RDF apparently is “hated”
#
sknebel
(it's even wikipedias example for RDFa https://en.wikipedia.org/wiki/RDFa )
#
Loqi
RDF in Attributes
#
sknebel
Zegnat: yes, JSON-LD islands have scopes etc, so you can describe multiple objects with properties
#
petermolnar
the real trouble is that images are resources on their own, often link on their own, out of context
#
Zegnat
sl007 there you go ^^^ you could put DC in JSON-LD on a page to describe photos. So you have (at least) RDF and JSON-LD available within HTML.
#
petermolnar
which makes it necessary to include the information _in_ the image
#
petermolnar
though I fear someone will come up with the idea of putting json-ld on the iptc description field...
#
Zegnat
petermolnar this is about putting that information in the HTML though, as publishers (sl007 mentioned CNN) will remove data in images
#
jeremycherfas
Petermolnar: That’s precisely the problem, and why online services that strip metadata are such a bad thing.
#
sl007
I am in contact with the IPTC people, for example when I was represented by Getty Europe, Andreas Gebhard was my boss. He is in the IPTC board of directors. We can work together but what is the preferred way to express it in HTML - link e.g. the resource to the JSON …
#
petermolnar
jeremycherfas I'm seriously against stripping metadata
#
jeremycherfas
zegnat: So this is just about making the data about an image machine-readable, not about preventing unauthrised use?
#
jeremycherfas
Because as soon as you save the image from a page, the HTML on the page is gone.
#
petermolnar
if it's on the internet, it's near impossible to prevent the latter :)
#
petermolnar
unless you go drm
#
sl007
stripping : well it depends if the resource is intended for public or private use
#
sl007
and if public if something of meta has to be kept private
#
Zegnat
This is all about presenting and preserving information when something is published, if I understood sl007 correctly. Else microformats are moot by default.
#
jeremycherfas
Petermolnar: Agreed.
j12t, eli_oat, John_Duh, [kevinmarks] and sl007 joined the channel
#
sl007
Has anybody tried https://solid.mit.edu so far ?
#
waynr
solid looks interesting
#
waynr
that web page isn't very concrete about what exactly solid is as far as i can tell
#
waynr
but this is a little more descriptive: https://github.com/solid/solid-spec
#
Loqi
[solid] solid-spec: The Solid spec and architecture
#
Zegnat
Prefers Turtle? So I guess, once again, plain HTML is going to left out? :(
#
Zegnat
“A WebID Profile Document is a Web resource that MUST be available as text/turtle”, I am sad now
#
sl007
Me too.
#
sl007
Zegnat - maybe we should invite Sir TimBL to an IWC. We could send it handwritten on golden paper in a unicorn envelope …
#
petermolnar
what is text/turtle?
#
sl007
What is turtle?
#
Loqi
Turtle is nothing like XML https://indieweb.org/Turtle
j12t joined the channel
#
petermolnar
this is one of the ugliest data formats I've ever seen
#
petermolnar
this is like xml written in rst
#
Zegnat
I think it is a parsed RDF tree written in Notepad.
webhat and grimspound joined the channel
#
waynr
scratches head confusedly
#
aaronpk
That's how I feel every time I look at Turtle. I can't get my head around the syntax.
[miklb] joined the channel
#
[miklb]
reading scroll back, had to do a double take to know what channel I was in ?
#
aaronpk
sticks to microformats in HTML instead
#
aaronpk
yeh this convo took a weird turn lol
wolftune, chimo, jmelesky and webhat joined the channel
#
sl007
Can Loqi answer directly (e.g. in direct conversations) ?
j12t joined the channel
#
Loqi
[o2Labs] etherpad-stream: Redirect STDOUT to Etherpad
#
sl007
I mean can Loqi reply to e.g. what is without logging it to a discussion ...
#
aaronpk
no I don't think that hook is configured for private messages
eli_oat and [kevinmarks] joined the channel
#
[kevinmarks]
Didn't Timbl come to one of the Cambridge iwc's?
#
[kevinmarks]
Or was it just Socialwg?
#
aaronpk
i think he was there for the socialwg meeting
wolftune, dgold, eli_oat, j12t, friedcell, arush, jensimmons, gRegorLove, snarfed, KartikPrabhu, sl007 and hs0ucy joined the channel
#
Loqi
[superfeedr] "We had over 30 people across three offices and several remote locations attend the first @DreamHost Website Club! #IndieWeb" by Jonathan LaCour on 2017-08-10 https://cleverdevil.io/2017/we-had-over-30-people-across-three-offices-and-several
[cleverdevil] joined the channel
#
[cleverdevil]
Pretty encouraging ?
snarfed and wolftune joined the channel
#
j12t
The IndieTech.rocks site now also has a T-Shirt. (I couldn't resist.) http://indietech.rocks/ -- preferably to running around wearing somebody's corporate logo IMHO.
snarfed, amz3`, wolftune and hs0ucy joined the channel
#
Loqi
[indienews] New post: "Syndicating Audio Posts with WNYC's Audiogram Generator" https://martymcgui.re/2017/08/10/180614/
biquillo, [miklb], KartikPrabhu, fzylogic, gRegorLove, snarfed, mblaney and sebi joined the channel