#indiewebcamp 2012-12-29

2012-12-29 UTC
friedcell, laurian, zztr, brennannovak, catsup and tantek joined the channel
#
tantek happy birthday aaronpk!
tilgovi joined the channel
#
zztr yay getting older! yay still living! yay time!
brennannovak and barnabywalters joined the channel
#
Loqi barnabywalters: tantek left you a message on 12/28 at 11:01am: FYI: when I tried to login with indieauth to http://waterpigs.co.uk/notes/740/ with http://tantek.com/ I got the error: "Unable to find the controller for path "/login". Maybe you forgot to add the matching route in your routing configuration?" - whereas it *worked* when I logged in via indieauth on your home page http://waterpigs.co.uk/. Then when I logged out and tried logging in via the form at the top of
#
barnabywalters tantek: Yep, that's a weird bug I haven’t resolved yet. Login fails the first time but succeeds the second
#
barnabywalters Not that logging in will make any difference right now unless youre me anyway :)
barnabywalters joined the channel
#
@BarnabyWalters Making a plugin for my #indieweb software to add #OSM Nominatim geocode data to my posts with location. Also almost… http://waterpigs.co.uk/notes/779/
barnabywalters and brennannovak joined the channel
#
tommorris barnabywalters: http://tommorris.org/posts/8003 ;-)
#
barnabywalters tommorris: upwards of 8000 posts? Impressive!
#
tommorris indeed.
#
tommorris there'll be more when Twitter give me my data.
#
barnabywalters I was actually surprised just how well Devon is mapped. I still manage to correct things and add new stuff everywhere, but it’s very denser
#
barnabywalters On a similar note, what happened to osmcheckin?
#
barnabywalters s/denser/denser
#
tommorris it's very, very slow and almost totally unoptimisable
#
barnabywalters stupid autocorrects
#
tommorris I did manage to accidentally leave the Amazon server running, which cost me $50.
#
barnabywalters wow — a pricey thing to run! Where along the chain is the slowness?
#
@tommorris On other #IndieWebCamp matters, @aral now seems to have rebooted his personal site after having previously reboot... http://tommorris.org/posts/8004
#
tommorris Amazon charge most of that for a MySQL server I wasn't using.
#
tommorris Which was my own silly fault.
#
tommorris The slowness is the Overpass/XAPI API
#
tommorris the eventual plan is to have a community hosted venues database.
#
tommorris which would have a "find my the (n: Int) nearest things to (lat, long)"
#
barnabywalters So who is organising that?
#
tommorris Me. Eventually.
#
tommorris I've been busy with work type work.
#
barnabywalters Sounds like a good plan if current APIs make things difficulty
#
tommorris and, you know, Christmas and stuff.
#
barnabywalters Yep, all that stuff :)
#
tommorris I have another OSM project that I'm working on soon
barnabywalters joined the channel
#
barnabywalters tommorris: Is there a way to easily add photo data to OSM?
#
tommorris I've been playing with Leaflet.js for WikiQueer Maps, a map service specifically for LGBT stuff. it's amazing how good OSM has got in terms of tool support
#
tommorris as in, mapping photos on to maps?
#
tommorris leaflet.js is probably the way to do it.
#
barnabywalters Yep
#
barnabywalters I need to play with that. Soon I am going to start adding checkin-like features inc. maps to my notes
#
barnabywalters I’ve been looking at the various tools build around OSM and have been amazed. It seems like a really solid community
barnabywalters joined the channel
#
tantek tommorris, so regarding a "community hosted venues database", I have mixed feelings about that.
#
tommorris tantek: well, it's really more just an API wrapper around the underlying OpenStreetMap data.
#
tantek So far my experience with Foursquare's "community generated/patrolled venues database" is that a few nitwits can pollute/damage the data quite badly, with deletions, overmergings etc.
#
tommorris there's nothing stopping people not using it if they don't want to.
#
tantek And sad to say, same thing happens with Wikipedia (excessive deletions, overmergings).
#
tantek So I hold little hope that a "community hosted venues database" wouldn't have all the same problems, if not worse.
#
tantek I'd rather pursue an alternative of a distributed indieweb hosted venues "database".
#
tommorris well, OpenStreetMap is sort of a community hosted venue database, it's just a gigantic pain to query
#
tommorris all I want to do is expose it in such a way that people can start using it.
#
tommorris incidentally, it's also a giant pain to update. having a way for normal people to report back inaccuracies would help with that.
#
tantek I'd rather post a check-in on my own site, and have it use a venue URL that is *also* on my own site. Then at that indieweb venue URL, it can link to equivalents on Foursquare, OSM, FB, and whatever other centralized venue databases are created in the future.
#
tommorris OpenStreetBugs already exists, but nobody really puts venue related stuff in
#
tommorris so, that's very sensible
#
tantek because frankly, I don't trust any centralized venue database, no matter how well intentioned or how "open"
#
tantek have had too many experiences with too few bad actors screwing things up
#
tommorris and indieweb reusers of OSM data would be highly advised to store the data about that venue as both a value and a reference.
#
tommorris and… hCard kinda solves that.
#
tantek yeah, hCard seems good enough for this
#
tantek I personally would add some extensions
#
tantek beyond simple "tags" in as "category" property values
#
tantek things like: hours open
#
tantek that seems to be a hard one to solve from a format perspective
#
tommorris so, OSM have come up with a rather strange string format for opening hours
#
tantek even more so, hours *foo*
#
tantek happy hours
#
tantek hours that food is served
#
tommorris for churches, there's a service times property
#
tommorris I can't remember what it's called
#
tantek hours that espressos is served (some cafes start shutting down their espresso machines early, or bars too)
#
tantek s/is served/are served
#
Loqi tantek meant to say: hours that espressos are served (some cafes start shutting down their espresso machines early, or bars too)
brennannovak joined the channel
#
tommorris for nightclubs and bars, you have some which are free to enter before time T but then have a cover charge after time T
#
tantek these are all from real world cases I've encountered, and made a personal note of in my address book for future reference, but which I'd be more than willing to share publicly for others to use
#
tantek typically I share these today as Foursquare tips
#
tantek so at least they're somewhat discoverable
#
tantek and potentially I can re-import them semi-automatically into my own indieweb venue URLs in the future
#
tommorris well, part of having an open source venues db is that people could post that stuff on their own sites, and then pingback them to such a db
#
tantek liking the new aralbalkan.com btw - thanks for linking to it
#
tantek except it needs an hCard :)
#
tantek wonders how we get Aral to show up in IRC instead of just Twitter.
#
tommorris seduce him with good UX
#
tantek wonders the same about barnabywalters sometimes ;)
#
tantek hah - IRC and good UX?
#
tantek.com edited /checkin (+2279) "aspirations and indieweb venues" (view diff)
#
tantek tommorris - I amended happy hours to include the free to enter before time T case ("discounted cover")
#
tantek feel free to add +1s or add more items: http://indiewebcamp.com/checkin#indieweb_venues
#
tommorris operating hours is possibly something one can steal from OSM
#
tommorris for microformats that is
#
tommorris https://wiki.openstreetmap.org/wiki/Opening_hours https://wiki.openstreetmap.org/wiki/Key:service_times https://wiki.openstreetmap.org/wiki/Conditional_restrictions https://wiki.openstreetmap.org/wiki/Key:collection_times and smoking_hours
#
tommorris opening hours gets really difficult when you take into account countries which have extended holiday periods
#
tommorris Ramadan for instance
#
tommorris (when I was in Israel, my hotel was fortunately right opposite an Arab-owned shop that remained open during Shabbat.)
#
tantek or really easy, in that every day gets it's own custom opening hours "event" - maps much better to hCalendar
#
tantek s/it's/its
#
Loqi tantek meant to say: or really easy, in that every day gets its own custom opening hours "event" - maps much better to hCalendar
#
tantek huh, I thought I had already added some OSM references to this discussion on microformats.org
#
tantek hmm
#
tantek wait a minute. darn forking discussions. http://microformats.org/wiki/opening-hours
#
tantek vs. http://microformats.org/wiki/operating-hours
#
tantek sigh, time for some merging.
#
tantek since OSM uses the term "opening hours" I'll prefer that and redirect from the other.
#
tommorris will write a brief description of how OSM handles them.
#
tommorris they've formally specced out how to parse the OSM opening_hours tag
#
tommorris and, incidentally, deprecated various complex things based on real world usage
#
tommorris and there's a JS implementation. ;-)
#
tantek.com edited /checkin (-2) "/* indieweb venues */ opening-hours" (view diff)
#
tantek thanks tommorris - feel free to add said brief description etc. here: http://microformats.org/wiki/opening-hours-formats#OpenStreetMap
#
tommorris yep, just expanding it now.
#
tommorris tantek: OSM opening_hours handles weeks.
#
tommorris week 4-16 Sa, week 50-53 Tu 0:00-24:00
zztr joined the channel
#
tommorris that's a venue that's open on Saturdays on weeks 4-16 of the year and Tuesdays on weeks 50-53. ;-)
#
tantek interesting. how "international" is the OSM community? (I mean, as opposed to "English-speaking-world-centric")
#
tommorris well, it's mostly Western European
#
tommorris it's heavily used in Germany and NL for instance
#
tommorris it's heavily UK based because Steve Coast is British (albeit he is now a Seattle-based Microsoftie)
brennannovak joined the channel
#
tommorris it's quite heavily used in EU countries, both English speaking and not.
#
tommorris US is weird because there's not so much demand for it
#
tommorris OpenStreetMap usage correlates pretty well with the popularity of cycling.
#
tantek fascinating (re: cycling)
#
tommorris or rather, the need for a map that's not heavily car-centric
#
tantek the US is quite car-centric
#
tantek ok, and there's been no big drama over using English abbreviations of dotw like Sa Tu etc. ?
#
tommorris so, you get cyclists, hikers, motorcyclists, skateboarders/inline skaters etc.
#
tommorris not particularly
#
tantek cool
#
tommorris in my little rural hamlet, I've added almost all the footpaths. the cycling people tend to like adding all the obscure cycling paths. ;-)
#
Loqi ALMOST ALL THE FOOTPATHS http://loqi.me/63d
#
tantek hahaha
#
Loqi hehe
#
tommorris almosts some verbs.
#
tantek I don't know of any other well established opening-hours like format, so that might a reasonable, "just pick the winner" and make the equivalent microformat approach that I used with hCard and hCalendar.
#
tommorris OpenStreetMap == long-tail maps.
#
tommorris Steve Coast "I stopped keeping track when they started adding base jumping sites"
#
tantek in the same way, Foursquare's user-created venue database is much better (more thorough, more up to date) than the venue databases that companies buy/sell, and for example, all the corporate maps tools use.
barnabywalters joined the channel
#
tantek Google is sort of trying with letting businesses create their own venues for themselves on maps, but it doesn't really help much in practice - nearly all Foursquare venues are created by someone who is *not* an authority for that venue, just someone local to it>
#
tantek s/>/.
#
Loqi tantek meant to say: Google is sort of trying with letting businesses create their own venues for themselves on maps, but it doesn't really help much in practice - nearly all Foursquare venues are created by someone who is *not* an authority for that venue, just someone local to it.
#
tantek will be interesting to see if/when OSM venues surpasses Foursquare venues in comprehensiveness, recency, etc. A better venue creation/editing UI would probably help.
#
tantek hey barnabywalters - we're talking about indieweb checkins and venues
#
tantek http://indiewebcamp.com/checkin#indieweb_venues
#
barnabywalters evening tantek
#
barnabywalters reads the logs
#
tommorris tantek: bulk import/merge is where it gets amazing. have I told you about the UK's release of the Food Standards Agency data?
#
tantek um no
#
tommorris every place that is inspected by the government to serve food or drink. in big XML files.
#
tommorris pubs, bars, restaurants, off-licenses (liquor stores), church/community groups/homeless shelters etc.
#
tommorris Sport England also have some data I'm trying to wangle that contains every amateur sports venue.
#
barnabywalters RE OSM editing UIs, I have found Pushpin for iOS to be very nice for POIs, and the latest version allows for editing of (but not creation of) areas too
#
barnabywalters it even manages to make all the tags fairly straightforward
#
tommorris Vespucci on Android is getting better… but it's no Foursquare. ;-)
#
barnabywalters and yes, IRC seems to be neglected by a lot of indiewebcamp folks
#
barnabywalters tommorris: so vespucci is an editor or a checkin app? Does foursquare do venue editing as well?
#
tommorris Vespucci is an OSM editor for Android
#
tommorris but in terms of the use case of "add the place I'm currently sitting in", it's hardly as friendly as Foursquare
#
tantek.com edited /checkin (+131) "hours that a customers-only bathroom is open" (view diff)
#
barnabywalters what about it is less friendly? less auto-filling of data?
#
tommorris well, you have to have an OpenStreetMap username. typing it in is fiddly
#
tommorris and you have to understand the tag formats and stuff
#
tantek tommorris - how about hours that a customers-only bathroom is open? https://foursquare.com/item/50df576ae4b0073ea8eaa235
#
barnabywalters oh, right. Well, with pushpin you obviously have to login, but IIRC it provides a UI for that, and it has a readable, searchable list of common tags
#
tommorris definitely useful. not something we've added to openstreetmap
#
tantek (at a gas station where the pumps themselves are self-serve and 24hr)
#
tantek well in case it helps - there's a real world example for you
#
tommorris yep, will add to the OSM wiki
#
tantek (have no idea about the "process" for adding things to the OSM format system)
#
tantek (but I figure that URLs to real world examples are likely to help)
#
tommorris mostly you just start tagging it in the most sensible way you can, then document on the wiki
#
tommorris toilets:opening_hours ;-)
#
tommorris goes to add it to the map.
#
barnabywalters foursquare’s 7.2/10 people like this place is weird, if statistically accurate
#
barnabywalters usually when referring to people I try to keep the numbers whole ;)
#
tommorris tantek: http://www.openstreetmap.org/browse/changeset/14456355 ;-)
#
tantek tommorris - so you just organically add stuff and people sort it out later?
#
tommorris well, in that case, the garage already exists
#
tantek are there naming conventions, how do divergent tags for the same thing get converged?
#
tommorris well, people use taginfo.openstreetmap.org to see the usage of tags
#
tantek and how do collisions get sorted out? the same tag being used for two different things.
#
tommorris well, ultimately, tags that are used to draw the map get sort of formalised by the mapnik developers
#
tommorris collisons sort of get sorted out eventually by documenting on-wiki, finding things that don't quite fit
#
tommorris so, a real world example. we have a tag to mark that a bar is a gay bar. gay=yes
#
tommorris so you do amenity=bar & gay=yes
#
tommorris but that's not very descriptive. on the to-do list is to separate it out and have a separate tag for lesbian bars. until that happens, that just goes in the "note" tag.
#
tantek why gay=yes rather than category:gay ?
#
tommorris cos too much namespacery requires people to remember them
#
tantek sorry, I meant
#
tantek why gay=yes rather than category=gay
#
tommorris for future extensibility
#
tantek where you can obviously have multiple categories
#
tantek uh, really? so every category is a boolean?
#
tommorris so, in general, you can tag something as 'yes' and also as more specific cases
#
tantek weird
#
tommorris no, it's a bit of a clusterfuck
#
tommorris so, there's amenity=[type]
#
tommorris shopping=[type]
#
tommorris like amenity=restaurant
#
tommorris but if, in a slightly weird circumstance, you wanted to say "well, this is an amenity, but I don't know what", you could do amenity=yes
#
tommorris it's one of those "it all sort of works" things rather than something that' sbeen well thought out.
#
tommorris worse is better.
#
tantek also I noticed that you provided the URL / attribution to me in the comment on http://www.openstreetmap.org/browse/changeset/14456355 but not the citation of the Foursquare Tip URL (which had photo etc.) - does OSM not do/need as much "citation" thoroughness as Wikipedia? Or is there a tacit policy of not linking to silo data for fear of polluting OSM?
#
tommorris yeah, linking to Foursquare would be a very good way to get people going "WTF?"
#
tantek I'm not sure "worse is better" is true in all axes of format development. If worse means a ton of crap gets added, then worse is worse. Simpler (but less expressive and thus "worse") does tend to be better in that regard.
#
tantek ok, so if I had posted that on my own indieweb venue URL for Chevron, then you could have cited that
#
tommorris yeah. ;-)
#
tantek a-ha, now we're getting somewhere
#
tommorris so, there's a source tag, which a lot of people use. I ought to use it.
#
tommorris and a source:url tag for pointing directly to a URL
#
tantek because then it's up to *me* whether I want to cite (or syndicate out to) Foursquare for the tip about the bathroom hours from my indieweb venue URL for Chevron
#
tantek then any OSM user can cite my indieweb venue URL without worrying that they're citing a silo URL.
#
tantek so OSM content needs a level of silo-source-deniability
#
tommorris well, sure, the problem is one of copyright.
#
tantek sort of - I mean, I created the content, so I own it.
#
tantek and obv I'm ok with contributing it to OSM, or having you do so
#
tommorris EU though. database copyright.
#
tommorris Foursquare would have database copyright for everything in Foursquare even if it's user contributed
#
tantek weird
#
tommorris but if you posted it on tantek.com, it wouldn't be.
#
tantek so that's another reason to POSSE rather than PESOS then
#
tommorris it's one of the major problems with licensing.
#
tommorris Creative Commons never really thought about database copyright
#
tantek anything you PESOS, you may be violating the database copyright of the silo you PESOS'd from.
#
tantek even if *you* contributed it to the silo in the first place.
#
tommorris the reason the British government had to write the Open Government License is because database copyright isn't done properly by CC
#
tommorris it'd probably be okay if you did it.
#
tantek I thought CC0 did handle it
#
tommorris CC0 can't force someone to not put it in a database which they then lock up under database copyright.
#
tantek CC0 is not about forcing negatives
#
tommorris twitter is an interesting example: it's quite possible that individual tweets are not copyrightable.
#
tantek CC0 is about explicitly disclaiming rights
#
tommorris but the collection of tweets would probably have a database copyright.
#
tommorris this is one of the ways that Europe has more fucked-up copyrights than the US
#
tommorris you stack together enough of other people's stuff and you get to magically own the collection so long as you put some creativity into building the stack.
#
tommorris I believe the timezone info database had this problem.
#
tantek.com edited /POSSE (+236) "/* Advantages */ ownership" (view diff)
#
tommorris +1 on that. ;-)
#
tommorris knows far too much about how copyright is broken.
#
tommorris so, an example of a kind of database copyright: charts. as in Top 40 albums, that kind of thing.
#
tommorris not copyrightable in the US. but copyrightable in a lot of other countries.
#
tantek.com edited /PESOS (+662) "/* Disadvantages */ Possible loss or restriction of ownership" (view diff)
#
tantek that http://indiewebcamp.com/PESOS#Disadvantages sound about right tommorris? tried to paraphrase what you said, feel free to improve.
brennannovak joined the channel
#
tommorris yeah, database copyright doesn't affect individual entries, but it deals with large collections of individual works. it's arguable whether, say, all the tweets posted by the same user would be affected by DB copyright
#
tantek hence I used the word "may"
#
tantek ;)
#
tommorris yep, that looks okay
#
tommorris that reminds me, next time I see Geoff Brigham, Wikimedia's General Counsel, I must ask him whether there would be some arrangement to have a stock of lawyers from non-profit free culture type organisations like Wikimedia, Creative Commons etc. to help answer these kinds of things authoritatively.
#
tommorris tantek: also, you asked whether you need sourcing and stuff for OSM. the answer is "broadly no". I try to provide photos for most of the things I add, but I frequently add stuff from local knowledge.
#
tommorris certainly for roads and other things like that, a publicly available GPX file is kind of expected
#
tommorris but for adding venues like bars and restaurants, doing it from memory or whatever is usually fine
#
tantek ok, just a habit for me I guess. I like to have citations for things.
#
tantek (kind of obsessed about it actually, though despite that, not a very good/frequent Wikipedia contributor).
#
tommorris indeed.
#
tommorris it's sort of a joke amongst wikipedians. before xmas, a few wikipedia admins got together and exchanged reliable sources.
#
tantek lol
#
tommorris your 4sq tips are useful. mine are sarcastic.
#
tommorris tip for London Bridge station "buy a jetpack to escape station during rush hours."
#
tantek I've written a few Foursquare "performance art" tips, usually only when they seem quite necessary.
#
tommorris also, for opening-hours: it almost feels like definition lists might be the answer.
#
tommorris <dl class="opening-hours"><dt>Monday</dt><dd>09:00 - 17:00</dd> ...
#
tantek e.g.: https://foursquare.com/item/4ffcb3c7e4b023bb62801ccc
#
tantek if you ever visit that venue, it will immediately make sense upon exiting.
#
tantek tommorris - so the problem with that for opening hours is that "Monday" is to English-centric
#
tantek that markup doesn't work for non-english sites
#
tantek we're likely going to need some approximation of the shorthand syntax from OSM plus how to wrap the "user visible text" on the page
#
tommorris it's something I'm going to have to deal with quite soon. I'm thinking of proposing that the RDF Core group completely rewrite the way they deal with datatypes.
#
tantek oh, just because you want to torture them with moving piles of sand?
#
tantek regardless, don't let me discourage you regarding the markup
#
tommorris no, just because datatypes were the most back-of-a-cigarette-packet thing.
#
tommorris and aren't fitting well with real world usage.
#
tantek that's perfectly reasonable POSH markup for opening hours, and if you were to publish examples like that, we'd use it to inform a potential opening hours property or format or other markup pattern for microformats.
#
tantek if it works for your site, use it. then we'll use that as data :)
#
tommorris yep. I know how-to-play. ;-)
#
tantek haha - right, *that* problem with datatypes.
#
tommorris so, an example:
#
tantek it was a misapplication of a CS-world modeling concept to a real-world-modeling format.
#
tommorris <#eiffelTower> <#cost> "650000"^^xsd:int .
#
tantek there's an even simpler answer
#
tommorris The Eiffel Tower cost 6.5 million francs in 1880.
#
tantek no data types
#
tommorris This is a real world problem that projects like Wikidata face.
#
tantek sure, but some CS-based type system is not the answer for any "real world problem" of data expression
#
tommorris The datatype should be "1880-era French francs" not "integer"
#
tantek well, we did finally come up with one datatype that was worth explicitly expressing - time
#
tantek hence we now have a <time> element in HTML5
#
tantek a pretty darn good and expressive one
#
tantek and one completely disconnected from the RDF/XMLSchema notions of time.
#
tantek because <time> is based on real world research into publishing
#
tantek not CS / databases
#
tantek I don't trust anyone who is starting with CS/databases to get anything right about data types for the real world.
#
tantek so yeah, time, strings. good enough for now. the rest we leave to the "application layer" :)
#
tommorris indeed, so my plan is to basically propose we pretty much deprecate XSD datatypes in RDF and replace them with an actual sensible solution for actual use cases which need some kind of typing, which is basically expressing units of measure.
#
tommorris things like temperature
#
tantek physics is not a bad start
#
tommorris saying "30" is an integer is not telling me anything useful.
#
tantek since that's based in real world examples ;)
#
tantek as opposed to math/compsci
#
tommorris telling me that "30" is in USD or EUR is telling me something useful.
#
tantek which is based in imaginary worlds
#
tantek anyway - sounds like a painful set of interactions. good luck.
#
tommorris well, fortunately, I don't have to convince the W3C to standardise it. I just need to make a compelling case to RDF users that they should ignore the W3C until they implement it. ;-)
#
tommorris and forge some URIs for actually useful things rather than badly-specified XSD nonsense.
#
tantek.com edited /PESOS (+55) "/* Disadvantages */" (view diff)
josephboyle, barnabywalters_ and tilgovi joined the channel