2016-10-24 UTC
edsu_, sagerdearia, MeanderingCode, tantek, KevinMarks, KevinMarks_, nitot, Erkan_Yilmaz_, dkm_, bear, KartikPrabhu and willnorris joined the channel
sagerdearia, KevinMarks, gRegorLove and tantek joined the channel
KevinMarks, KevinMarks_, sagerdearia and tantek joined the channel
# 20:17 tantek the larger challenge is that even the terms like "area" and "place" are so heavily overloaded as to potentially be indistinguishable, and mean different things to different people
# 20:18 tantek by "just an area" do you mean a literal area? like a polygon?
# 20:19 gRegorLove data-value pattern, then the text can be something more human-friendly
# 20:21 gRegorLove You could potentially put the human-readable name in the one data tag and leave the other empty, though maybe that's considered a bad practice
# 20:21 tantek jonnybarnes, interesting wasy of phrasing the question "doesn’t warrent an h-card" - implying there's a hierarchy (or ordering) of different microformats, from simpler to more extensive
# 20:21 gRegorLove <data class="p-longitude"...>Salford, UK</data> <data class="p-latitude".../>
# 20:22 tantek jonnybarnes: fundamentally the "how should it be marked up" question requires a real world example of the location information in plain user visible text, in context, to help determine what a good (least intrusive) method of marking it up would be
# 20:23 tantek reasoning by abstract example is a poor substitute for that, and often results in abstract markup that works ok in isolation but not in context of real world content / pages
KevinMarks joined the channel
# 20:24 tantek jonnybarnes, so let's start with, what location information are you displaying on your page? (on what page?)
# 20:24 jonnybarnes okay, my notes that aren’t checkins but have geo-data simple end with “in {town/city}, {country}
”
# 20:25 tantek (aside: gRegorLove that h-geo FAQ likely needs updating with the nuances I just pointed out above about context / real world example driven design)
# 20:26 tantek jonnybarnes - is it plain text? is it a clickable map link? is there an embedding of a map, thumbnail or otherwise?
# 20:27 tantek do you have any UX/design-based thoughts about how you may want to evolve it in terms of interactivity?
# 20:27 tantek nevermind the markup thoughts for now - that's plumbing
# 20:28 tantek jonnybarnes - a plain text location (named or otherwise) may seem "abstract" to the average reader - they may have no way to relate to it unles they're already familiar with the city/state/country
# 20:28 tantek (certainly very few humans relate to lat long #s)
# 20:28 jonnybarnes the next step with them is to get all the notes also from that place
# 20:29 tantek would clicking on that location text then show a list of all the posts made at that location?
# 20:30 jonnybarnes but it shouldn’t be too hard to get all the notes also associated with that place
# 20:30 tantek I understand not currently - hence I'm asking if that's your future thinking
# 20:30 tantek thinking through design plans like this helps better figure out what kind of markup will best support current and soon design plans
# 20:33 jonnybarnes from the h-adr wiki: “For named addresses, e.g. people or venues, use h-card.”
KevinMarks joined the channel
# 20:34 tantek but as I noted to gRegorLove, I likely need to update the h-adr documentation to reflect more nuances
# 20:35 tantek in particular, we've learned a lot about this problem through the indieweb pursuit of figuring out venues in particular
# 20:35 tantek a venue has a more specific meaning than a "place"
# 20:35 tantek whether something is specific enough for an h-card is a very good question and I think we're still figuring that out case by case
# 20:36 tantek as you said / implied, if an h-adr is sufficient that's likely the way to go
# 20:36 tantek the reason I asked about the display is that when you show a map, you're essentially showing a visual representation of the lat long
# 20:37 tantek and could even use <object> that way to show the interactive map, with fallback plain text coordinates inside, marked up with microformats properties
# 20:38 tantek it should work for embedding anything you would use iframe for
# 20:39 tantek and if it doesn't, I want to know about it :)
# 20:40 tantek how does your site decide to show a map or not for a post? (those two example you gave - CB vs Bs)
# 20:42 jonnybarnes the one in Salford just has associated latitude/longitude values with the note, that just gets turned into text via nominatim
# 20:53 tantek is it an explicit design decision to not add a map of a city/state/country by itself? e.g. coarser? or just an accident of order of implementation?
# 20:54 jonnybarnes explicit, for a map I can center it to a specific point, Salford has 220,000 residents, where do I center the map?
KevinMarks_ and sagerdearia joined the channel
# 21:12 tantek jonnybarnes - if you zoom out far enough, a city looks like a point as well
KevinMarks and miklb joined the channel
# 22:27 tantek to indicate the user explicit intent of "I am/was at this location [at this time - autofilled, or past noted]"
# 22:28 tantek then a checkin could be represented by an h-entry with a "p-checkin" of that spacetime coordinate
# 22:28 tantek would h-entry p-checkin h-event be too much of a hack?
# 22:29 tantek where the check-in time is noted as the dt-start of the event?
# 22:29 tantek (events can already only have a dt-start, dt-end is optional of course)
# 22:30 aaronpk that would mean the location would be the location of the h-entry?
# 22:30 tantek is that overkill? the venue ends up being even deeper
# 22:30 tantek the location of the h-entry is where they were when they published the checkin
# 22:30 tantek the location of the (p-checkin) is the p-location of the h-event inside
# 22:31 tantek which itself is thus likely an h-card for a venue
# 22:31 aaronpk that makes sense, but i'm a little worried that buries it too deep
# 22:31 aaronpk deeply nested by itself isn't so much a problem as when there are multiple places the data could be
# 22:32 aaronpk if that's the *only* place a consumer should expect to find the location of the checkin that wouldn't be terrible
# 22:33 tantek an alternative would be relying on a collection of tags
# 22:33 tantek but that's once again imprecise and doesn't indicate user intent of a checkin
# 22:33 tantek e.g. a person-tag of themselves, a location-tag of the venue, a datetime-tag of time of checkin
# 22:34 tantek but lots of things could have that set of *-tags
sagerdearia, bear, edsu_, Garbee, aaronpk, JonathanNeal and iwaim joined the channel