2013-08-11 UTC
tantek, xtof, aaronpk, tobyink, eschnou, bnvk and barnabywalters joined the channel
bnvk joined the channel
danielfilho, bnvk, nonge, barnabywalters, adactio, tobyink and bnvk_ joined the channel
# 20:06 neuro` bret: it's going well, thank you. Next version will provide hcard and h-entry out of the box (plus POSSE for Twitter) # 20:09 bret good, demos went well at osfw3c but i think I need to find a better way to describe why what is happening is interesting # 20:09 neuro` I followed osfw3c from what tantek and kevinmarks twitted, it sounded really interested. # 20:10 Loqi neuro` meant to say: I following osfw3c from what tantek and kevinmarks twitting, it sounding really interesting. # 20:10 bret cool! Then we can use publify content in reply contexts ^_^ # 20:11 bret Yeah, business/enterprise people really thing differently about this stuff # 20:11 neuro` I didn't tweet much, mostly read :) And I need to implement "in reply to" and retweet there. shaners joined the channel
# 20:14 shaners bret: let's talk about this over here, not in #indiewebcamp # 20:15 shaners so, as I understand it, JSON-LD is vocabulary agnostic. which means that if one wanted to use h-card's vocabulary in JSON-LD, they could. in theory. # 20:16 shaners but you can see from their simple example on their homepage, they haven't # 20:17 shaners at the end of the day, we've got something that A: already works, 2: is being widely used and D: is optimized for published first, not developer (which has lead to greater adoption of mf) # 20:17 bret i was getting the impression that the vocabulary and its representation in json was what they were trying to standardize # 20:17 bret uF2 isn't hugely adopted yet though, no? mostly classic # 20:18 bret uF2 does vocab inside html which maps to json # 20:20 bret tommorris: do you see much crossover with uF2 and JSON-LD # 20:20 shaners tommorris: i understand your point in your post. but i think it's the same logic as "everyone is using xml" from 10 years ago. # 20:21 shaners my point is that mime types are fashionable. they come and go. but html has stuck around. # 20:22 bret It seems like JSON-LD recognizes "html + RDF" solved data inside html # 20:22 bret but they want this standard to do the same for data represented in JSON # 20:23 tommorris so, here's a tougher use case than putting-contact-details-online # 20:24 tommorris you have a chemical database that describes research data into the interactions of genes in the progress of complex diseases and medical conditions - cancer, HIV/AIDS etc. # 20:25 tommorris it links between specific highly-complicated scientific areas - genetics, diseases, drug interactions etc. # 20:26 tommorris and it links into existing complex databases, some of which are public and some of which are behind academic firewalls/paywalls # 20:26 tommorris it's very important that when two scientists are talking about a particular gene, they are talking about the same thing # 20:27 shaners tommorris: so let's unpack this into small pieces. # 20:27 tommorris the systems that use this data are diverse. they might be RDBMSes, they might talk XML, they might talk JSON # 20:27 bret it seems like given the argument shaners is making about mime formats failing, and uF not, that any kind of data vocab should be tied to the non-mime representation, and map to the mime style representation # 20:28 tommorris the point is that you have RDF which is a Resource Description *Framework* # 20:29 tommorris whether that's in HTML (with RDFa), in XML (RDF/XML), in JSON (RDF/JSON originally, now the much, much simpler JSON-LD) # 20:29 bret can you have such a framework, and still have the simplicity of uF2? barnabywalters joined the channel
# 20:30 tommorris one of the reasons I'm writing my own mf2 parser is that I can sneak on a make-sane-RDF option, so you can merge data contained in microformats2 with RDF data barnabywalters joined the channel
# 20:31 shaners tommorris: i guess, it's like this. every time i hear these examples, they're either # 20:31 shaners A: completely made up not real world use cases (not saying you're making this one up) or # 20:31 shaners 2: have yet to see a case where if someone wanted to model a "complex disease" (or whatever, anything more complex than h-cards) in mf2 that they couldn't # 20:32 shaners i haven't seen any effort from the LD set to work on an mf for one of these straw man examples # 20:32 tommorris that's because Linked Data people already have an existing data-in-HTML solution, RDFa. # 20:33 shaners i mean, right? is there some history of someone trying to make a mf for diseases or genes or whatever complex science thing? # 20:36 shaners tommorris: when did people start using RDFa in HTML? # 20:36 tommorris What put a lot of people off in the early days was that it didn't validate. # 20:37 tommorris Microformats had that as a significant win: because it was just classnames, it was valid HTML 4.x/XHTML 1.0 # 20:37 shaners from my perspective (and this is certainly) outside looking in, was that it wasn't until the past couple years did people start using RDFa. # 20:38 tommorris it's become a lot more usable in the last few years as they've made the effort to spec out how it works with HTML5 # 20:38 tommorris and also people stopped giving a shit about validation so much. ;) # 20:38 barnabywalters I don't think it’s very helpful to look at when something was published as opposed to consumed # 20:39 shaners meanwhile, were people just waiting until RDFa validated before doing anything in HTML. why weren't they getting into mfs? # 20:39 shaners tommorris: either way, debating history doesn't change what's what today. # 20:40 tommorris the social model of microformats is: come to microformats.org, show evidence, work on a spec collaboratively # 20:40 shaners as opposed to, just make up a new vocab for an already existing thing? # 20:40 tommorris social model of RDF is: do whatever the fuck you want. publish it yourself, get together with your friends, build a gigantic committee. whatever you want. # 20:42 tommorris so if you think schema.org is right, use that. if you think facebook graph is right, use that. # 20:42 tommorris and you can declare equivalencies, you can extend existing specs # 20:42 shaners in your example about _making sure that two scientists are looking at the exact same gene_, this "use whichever thing you think is right" does't jibe # 20:42 Loqi shaners meant to say: in your example about _making sure that two scientists are looking at the exact same gene_, this "use whichever thing you think is right" doesn't jibe # 20:42 tommorris so if you are aforementioned scientists you can use complex gene ontology put together by beardy experts # 20:43 tommorris and you can also use microformats to publish your contact details # 20:44 tommorris on the one hand, everyone bitches about how the RDF world is filled with committees spending years on specifications. # 20:44 shaners i'm still waiting for a genuine real world use case, where an mf couldn't be used or created # 20:44 tommorris on the other, most RDF specs are put together by a programmer in a text editor. often while drunk. # 20:45 tommorris Yes, microformats2 has brought microformats a lot closer towards an RDF-like model. # 20:46 shaners i don't believe that an mf couldn't be specced out for chemicals, for example # 20:46 shaners and that there isn't already some (a lot of) prior art on vocabulary # 20:46 tommorris sure, but why bother? the community of people who do that already have semweb technologies which they are using # 20:47 shaners i also don't believe that it's Hard™ to consume html+mf2 # 20:47 tommorris well, I'm having to run a node.js daemon on my server to read microformats2 # 20:48 shaners and i don't believe that there is any technology that has more available tools and potential consumers for it than HTML # 20:48 shaners publish HTML + (MF2 or RDFa) on the web and literally everyone with a browser can use it # 20:49 shaners publish JSON (in any "format") and only programmers or snowflake apps can use it # 20:50 tommorris the problem I have with data-in-HTML approaches is that actually solving the problem of consistent rules for working out the relationship between the page and the objects contained therein is difficult # 20:51 shaners i would love to see the day where we're only debating HTML + ?, rather than HTML vs JSON vs XML, etc # 20:51 tommorris well, for me it's not a matter of debating. it's a matter of people actually using JSON, so what's the best way to do that in a non-sucky way # 20:52 tommorris same as people used to use XML, there's ways to do that in a sucky way and a non-sucky way # 20:53 shaners they'll have to put coding time into working with this new flavor of json. that time could be used instead to code against parsing mf2 (or mf2 to json) # 20:54 tommorris like, a "debate" on the merits of data-in-HTML vs. JSON is pointless because people aren't convinced of technology choices on the basis of reason and argument. they are convinced by fashion and engage in a lot of cargo cult science # 20:55 tommorris XML became popular because it was pushed by industry, was intimately connected with lots of other sexy things they liked # 20:55 tommorris it had a certain sheen from the W3C, it had the approval of the Java enterprise crew # 20:56 tommorris someone thought a programming language in XML was a good idea. # 20:57 bret The thing that appeals to me about data in HTML is that HTML never goes away, where your API backend will eventually die # 20:57 tommorris indeed, one of the worst things that happened to RDF was the fact that the first implementation was RDF/XML # 20:57 bret maybe I am an being to simple about this # 20:57 shaners and to me, all of this JSON stuff just feels like that all over again # 20:57 tommorris most people in the RDF community would be very happy if RDF/XML would go away and we'd never see it again. # 20:58 bret the other problem is nobody is flat out wrong about any of this stuff :p # 20:58 tommorris the problem with JSON is that it's the building block of the webapp.js world which is slowly eating away at the HTML world # 20:59 tommorris instead of HTML on the web, we're getting websites that are just JavaScript that load more JavaScript # 20:59 tommorris if you open them up with JS turned off, you get nothing at all # 20:59 bret I'm not making claims, just trying to undertstand # 20:59 shaners tommorris: agreed. it's a shit show. and i think it's a passing fad. # 20:59 bret javascript doesnt seem like an issue to me either… it never disappears, like html, its the API backend that goes silent # 21:00 bret if your API backend is data from some other HTML source called in with JS, its should be fairly resilient # 21:01 shaners bret: it IS the scripting language of the web. it's not going anywhere anytime soon. anymore than HTML or CSS are going away. # 21:01 bret tommorris: I saw that post, and it was definitely an appropriate response to the troll who provoked it # 21:02 tommorris what provoked it was the redesign of Google Groups and the utter frustration I felt using it # 21:02 barnabywalters I didn’t live through the xml wars, not do I know much about RDF in any of it’s various forms. Not do I suspect that trying to actively persuade anyone that one technology is better than the other is going to get us anywhere. What will is what we're doing at the moment, solving real problems with mf and proving the community model works by iterating on them, and most of all building things which actually consume the data and make it # 21:03 tommorris has his theological preferences but is a big fan of anything that works # 21:04 bret I guess I'm still worked up over this weekend listening to all of this JSON-LD propaganda, but not seeing anything tangible from it. http://www.fatberg.org # 21:05 tommorris pragmatism: the realisation that the pain caused by shit not working is larger than the pain caused by the solution matching up exactly with your dogmatic preferences. # 21:05 shaners tommorris i have ideological prefs too. they just happen to align with what works. …mf2 # 21:06 tommorris (Although the last time I did .NET programming, it wasn't awful.) # 21:06 tommorris Googles: "Shibboleth is among the world's most widely deployed federated identity solutions" # 21:07 shaners bret: so back to your original message about mf2 + JSON LD compatibility, i have very low expectations of any interop. # 21:07 tommorris oh, it's something made by academia with the support of Internet2 people. # 21:08 tommorris shaners: I've looked into it, and microformats2-to-rdf is fairly straightforward. # 21:09 bret what do people think of the analogy that uF2 is like the tardigrade of webAPIs? # 21:10 bret it provides an api that just sits there, in the html, dead. In the right environment though, it comes alive as a data API # 21:10 shaners back when it was kicking off, i lived in Bloomington, Indiana (where IU is) and was all excited about this New Shiny Fast Internet™. I was 19. # 21:11 tommorris Internet2 in simple terms: the UK has an very fast inter-university network called JANET. Now make it bigger, brasher and American. that's Internet2. # 21:13 shaners alright friends, i'll be AFK for a while. doing an airport pickup. laters! # 21:13 bret barnabywalters: unlike json-ld, it isn't assumed that it relies upon some complex backend (which seems implied) # 21:15 tommorris JSON-LD doesn't require a complex backend. trust me, I've seen some crazy RDF specs. JSON-LD is pretty tame. # 21:15 tommorris it's just JSON with a very thin layer of "here's how it maps to RDF", almost like a stylesheet # 21:17 tommorris in fact, it wouldn't necessarily be that hard to layer JSON-LD on top of microformats2-JSON # 21:17 bret ok gotta run, thanks for increasing my understanding on all of this barnabywalters shaners tommorris # 21:17 tommorris just wishes he had a few extra days each week to work on this stuff # 21:18 bret uF2 still seems the most minimal spec of them all # 21:29 tommorris my favourite spam was entitled something like "For the gay boys out there" and then the body was promoting books on how to pick up women. tantek joined the channel
# 22:27 tantek bret - I see shaners already discussed a bunch of the microformats2 JSON representation stuff krijnh joined the channel
# 22:30 tantek retry: bret - I see shaners already discussed a bunch of the microformats2 JSON representation stuff # 22:46 bret tantek: yeah, we had a bit of a discussion # 22:46 tantek bret - the short answer, which I didn't see in the discussion, is that the microformats2 canonical JSON is based on microdata canonical JSON # 22:47 bret I was curious if json-LD could share common ground with MD canonical json # 22:48 bret but then it came down to the fact it probably wanst important # 22:48 bret at least thats the impression i was left with # 22:49 bret i need to stop distracting myself with things that don't involve what I am focused on # 22:50 tantek at the time when I was figuring out canonical JSON for microformats2, I chose the simplest other such generic use of JSON # 22:50 tantek so I deliberately chose for microformats2 JSON to share common ground with microdata canonical JSON # 22:51 tantek when something more complex comes along (like JSON-LD), it's up to the more complex thing to share common ground with the simpler thing, not the other way around # 22:51 tantek so basically, someone in whatever channel/forum JSON-LD is discussed should be asking if they can share common ground with microformats2 JSON, not vice versa # 22:55 tantek always pursue the simplest solutions, and ignore the complex ones, despite their pleas for common ground or compromise # 22:55 tantek leave the responsibility to those who pursue more complex solutions to educate themselves and simplify (or use the simpler solutions as building blocks for their own complex use cases)
qihe5848 edited /User:qihe5848 (+3337) "/* google google The tooth did make it out 4794aOqvjg22822 */ new section" (view diff)
tommorris shaners: http://tommorris.org/posts/8487
barnabywalters I don't think it’s very helpful to look at when something was published as opposed to consumed
bret I guess I'm still worked up over this weekend listening to all of this JSON-LD propaganda, but not seeing anything tangible from it. http://www.fatberg.org