#AnnBat the same time, I think we need to look at ALLLLLL the various efforts as important contributions
#tantekAnnB - it's interesting, I don't actually believe in "colective defining" - I believe in distributed independent implementations organically converging
#tantekAnnB - "ALLLLL" sounds a bit too PC, e.g. plenty of efforts that were mere re-attempts of previous failures that didn't really inform much if at all
jasnell joined the channel
#AnnBit might be said that I'm much more social than you .. I'm all about the collective
#tantekAnnB - I'm all about the collective that works in practice rather than in theory
#AnnBI do think previous efforts are important .. many efforts must be tried usually, and most fail -- before something 'gels'
#AnnBthe word "reject" makes me nervous .. seems to harsh
#tantekI like taking an evolutionary approach to considering ideas, if the ideas are worthy (fit) enough, someone will build it and make it work, if not, then perhaps the ideas need iteration
#AnnBif I have an idea, and you say you reject it .. I feel as if you're rejecting me
#tantekAnnB - rejecting something is how you focus
#tantekand that's why we have a good split between WG and IG!
#tantekfor the IG to explore those other ways "in which people explore and communicate"
#melvstertantek: not *all* the sem web folks rejected micro formats, myself rhiaro tommorris_ and others embraced microformats where they have been useful, I would say the converse is less common
#melvsterat the same time, as you scale, e.g. to the enterprise, you need to start to focus on extensibility
#tommorris_I'm neither in the Social WG or IG but I should note that I'm not an "XML" person. SemWeb tech is best without XML.
#tommorris_Also "enterprise" gives me headaches. ðŸ˜
#melvstertommorris_: didnt say you were xml, but I thought you had used some sem web stuff in the past ... and also some indieweb stuff
#tantekmelvster: initially there was total rejection, however you are right that fairly quickly (within a year of introduction), a few semweb folks found ways to interop - there was a lot of deliberate outreach to make that happen
#AnnBtantek, why do you say that about scale and enterprise ... I don't get the joke
#melvstertantek: we discussed IG items on the call and how they relate to the WG ... e.g. user stories ... it would be nice to organize them ... the social architecture doc, we thought about finding common prerequisites to the user stories, and for the vocabs I think that's largely done already by james
#melvsteryes well, just different perspectives ... happy to disagree there, let's just say some of the work is done, and im personally quite optimistic
#tantekBTW to be fair, h-entry has the benefit of being more strictly based on prior shipping/interop work, like Atom 1.0
#tantekI too am optimistic, but more about h-entry since it has far more interop publishing/consuming implementations than AS
#AnnBhey, tantek ... when you say that's a "formal spec" .. what does that mean? (in terms of formality, is my question)
#tantek(not "done" yet, but certainly lots of it is "working")
#tantekAnnB - formality means: 1. with enough precision for multiple implementers to independently write code following the prose in the spec. 2. have those implementations interop.
#AnnBI guess I'm asking, in MF community, how does something get formalized
#tantekvs. informal specs which have prose that is too vague to get interop
#AnnBso .. in this case . you wrote that description in a structured manner .. and that gives implementers a consistent basis on which to proceed?
#melvsterwhat we talked about on the IG call was some kind of social architecture that would facilitate many of the hidden assumptions in the user stories ... for example for users following each other etc. which comes up time and again ... we dont have any documentation for this so that could be a nice to have
#tantekso far prior attempts at social architecture before implementation have all been failures
#tantekmelvster: e.g. given your for example, "how does a user follow another user?"
#melvsterwell that's a perspective, but let's try and be constructive
#tantekmelvster: not a perspective, but rather, historical evidence
#AnnByes, we talked about some aspect of documenting various approaches... seeking a factual description, and describing the results .. if we can get away from the emotions of personal biases
#melvsterone system we discussed that's arguably not a failure is facebook
#tantekothers are welcome to try a priori social architecting - it just has a very bad (100% failure) track record
#tantekthus I am pursuing an implementation-based incremental social architecture as it were
#tantekin the form of interoperable building blocks
#tantekmelvster - there is a lot to be learned from how facebook got so many websites to deploy: OGP, Like Button webaction iframes
#aaronpkfacebook is a silo, not an architecture, not a way to build a distributed social web
#aaronpkbut yeah like tantek said there is a lot to learn from that!
#aaronpkbut you can't possibly argue that facebook is a successful example of a distributed web
#AnnB(I'm a pretty good 'shill' for you guys ... needing to often ask what the heck you mean)
#tantekit's still hub-and-spoke, however I'm presuming that when we (in #social) talk "social architecture" we are implying *distributed* or even *federated* social architecture
#AnnBFB is certainly a silo, not open, etc .. but you have to agree it's huge, vast, and has worked -- in it's own way -- across bazillion users and sites
#tantekthus I wouldn't count FB (nor it's use of OGP, Like button iframes etc.) as an example of a "social architecture in the context of Sociwl Web WG
#aaronpkwhich means more than "here's how to build stuff so facebook can consume it and show it in the feed"
#tantekAnnB - and frankly, the fact that blogs / RSS / SemWeb etc. have all failed to learn from FB's UX is one of the major reasons none of them have built anything vaguely even potentially compeititive
#AnnBwe might not go down some path, or repeat some approach .. but surely we can learn and incorporate
#tantekAnnB - in order to build something, you can't spend all your time thinking, you need to spend some time building. you have limited time to think, thus you must focus it on what is most productive to think about. focus = reject things that are less productive to think about.
#tantekit's a good way that asking people to build things forces such focus
#AnnBthe balance between human interactions and the need for tangible progress is very hard to achieve
#AnnBwe probably neutralize each other in these ways
#AnnBI worry about how people get along; you worry about tangible implementation
#AnnBand now I'm worried about how I'll get along with my boss if I don't get in my car and go to the office!
#tantekAnnB - it's not "out-of-hand" rejection, but rather, "great, nice idea, let me know when you have it working!"
#tantekhowever when you hear someone come up with ideas over and over again, and never actually implement/build/ship something, you tend to treat it ias a "crying wolf" type situation
#melvstertantek: facebook is indeed a silo, but that's a policy, ie same origin policy, for connections in its graph, the graph itself architecturally is a valid system, even if the policy isnt
#tanteknope, it's single implementation, not an architecture
#tantekasserting architecture that way is just as flawed as pointing to an open source project and asserting that it's a spec
#melvstertantek: imho indieweb itself would be monoculture by that definition, according to the bespoke "you homepage is your identity" dictum, but you'd probably disagree on that ... everyone will have different definition of different terms, it's more constructive talk about what's common ... in the case of the user stories there's are common assumptions that underly them, I think it would be good to try and capture those
#aaronpkmelvster: "Monoculture refers to the antipattern of one piece of software..."
#tantekmelvster: nope. indieweb is a plurality of implementations
#aaronpkthe key is the "one piece of software" part. if indieweb were saying "hey everybody install ___ on your website and we can all interoperate" then that is monoculture
#melvsteryes people will have different definitions here, it's a hornets nest, diaspora will argue each instance is different, so will indie web ... from my POV it doesnt matter too much, what matters is interop
#aaronpkthere's not really anything to argue about with that definition
#aaronpkdiaspora is a single implementation. if there were other implementations of the diaspora protocol that's a different story.
#aaronpkif so, we should update that wiki page accordingly
#aaronpkmodding is different, that's like installing wordpress plugins, but the core is still the same code
#ben_thatmustbemeif the API is reasonable one would have expected other implementations to exist already
#tantekmelvster, citation needed for real world deployments of such "can mod diaspora"
#tantekfor a monoculture, wordpress is surprisingly flexible with plugins, it's likely what's help it stay alive in contrast to other monocultures.
#tantekmelvster: your assertion of different definitions is irrelevant here in a W3C context, as for example, W3C requires multiple interoperable implementations
#tantekthus it makes no sense to pretend that a single implementation is sufficient or desirable
#ben_thatmustbemewell, honestly i'd argue that wordpress is basically just a CMS, of which there are tons, and thus, not monoculture
#tantekben_thatmustbeme: that's like saying "it's a website"
#tantekit's HTML and HTTP that have multiple implementations, not "WordPress"
#ben_thatmustbemewell, yes, it is, unless there is some interop, why bother defining monoculture or not for that case
#tantekto identify anti-patterns vs desired design patterns
#melvstertantek: i see your point, just everyone has different anti patterns, I dont see much point going down that rabbit hole, I can see I see your point of view, but also respect others, that's about all
#melvsterfor some saying your homepage is your identity is an anti pattern
#melvsterwhat should be distinguished is adoption ... ie HTTP and HTML are very well adopted, things like wordpress well adopted but not universal, things like diaspora relatively little adoption etc.
#tantekit's not a rabbit hole - multiple implementations is a W3C process requirement
#tantekif you're not ok with that as a given you should appeal to the W3C Process Interest Group
#melvsterim not arguing against multiple implementations, I would like to see multiple implementations, what I'm saying is that each person has their own perspective on what is an anti pattern and what isnt
#tantek"each person has" sounds like weasel wording again
#melvstertantek: then it's up to you to prove that every member of the WG and IG have the same anti patterns as you, and no others, in that event you would be correct, but I doubt you could do that
#melvsterand id suggest it's not a good use of time
#tantekno such proof is necessary when alternative assertions are trivially debunked
#melvsterthen you have to prove every diaspora pod runs the same software
#melvsterthere's no a priori reason to think they do
#tanteknah, statistical sampling sufficient for scientific purpose
#tantekforks of open source projects are kinda irrelevant from a deployment perspective, heck, even just the presence of open source projects are irrelevant from a deployment perspective given the poor adoption rates
#tantekwell documented %s of abandoned open source projects etc. on Google Code and others
#tanteknope, that was on diaspora's own wiki page that aaronpk provided above
#melvsterwell if you can show they are all the same software, you'll have convinced me, until then it's just a dubious claim without evidence on a wiki page
#aaronpkgreat link. that page actually demonstrates they are all the same software
#aaronpkthat page also demonstrates the inherent monoculture aspect of diaspora, since it's written to show which versions of *the diaspora software* are running on various pods, rather than any sort of mention of protocol versions
#melvsteri see what you are saying, but it's a subjective viewpoint, not fact, and I dont think it's a productive use of time to argue about
#melvsterthe software is clearly not the same, line for line, it may be similar enough to be called the same, but that's a judgement call
#aaronpkI doubt that would pass w3c process to be considered different implementations
#tantekas far as how much does software have to fork / diverge for it to be considered different - that IS a difficult question and without specific ways to evaluate it, tends to be subjective (debated politically)
#tantekreal world example: Chrome (Blink engine) vs. Safari (Webkit engine)
#tantekin terms of considering interop of CSS features
#melvstertantek: 'same software running on servers run by different people' is the claim on the wiki page ... there's zero evidence for this linked from that page ... you cant compare that to climate change, where there is a lot of evidence its getting warmer, but people argue about the rate and effect
#tantekaaronpk: I didn't catch the reasoning on the podupti link - could you add it and explain how it supports the assertion that Diaspora is a monoculture?
#aaronpkyeah again, totally not relevant to what i'msaying
#aaronpki was pointing out that matrix, in contrast to Diaspora, is consciously promoting multiple implementations of their protocol
#tantekadoption has shown to be important for social silos, however it has not yet shown to be (more) important to a social (distributed) web (than say, multiple implementations)
#tantekeven the social silos, with all their "adoption" have very poor (if any) interop
#tantekvery little "web-like" about the adoption in/of social silos
#tantekadoption is also something that can be done incrementally generation by generation, secondarily to more important progress such as a plurality of inteorperable implementations. http://indiewebcamp.com/generations
#tantekpossibly, depends if it expands costs more than revenue
#tantekmay also trigger the human proclivity for premature excessive growth and subsequent failure (profitable at scale N, unprofitable at scale 10N - as happened with numerous sites/silos in the dotcom boom/bust, e.g. webvan etc.)
#melvsteryes also true, i think we agree, adoption is one part, normally a plus, implementations are another, also a plus ... everyone will assign their own weightings to those
#melvsterso my original point was that your use of the term 'monoculture' really just means 'similar software' ... there's room for interpretation in the defintion and what defines similar ... to me a mono culture is relevant to the extent that it enables people to work, e.g. by helping or hindering interop
#tantekthere's been some brainstorming about how to do IRC via POSSE from your indieweb site, however I can't seem to find them. aaronpk may know where to look
#tantekThere's even a Known plugin for doing POSSE to IRC from your indieweb Known site!
#melvsternow that we have slack, gitter, zulip etc.
#tantekIRC sets quite a high bar for a low latency open protocol with multiple implementations
#tantekmelvster: it's more than just "similar software" - what makes a community a "monoculture" is the *assumption* of that community that everyone will use the same (nearly the same) software
#tantekwhich tends to happen when communities form around singular open source projects
#melvsterit's still a niche tho and in the last 1-2 years is really getting killed by web based offerings, id guess that's a trend that will accelerate now some open source systems are available
#tantekthat is, communities based on single open source projects (e.g. with one central repo) tend become monocultures
#tantekmy anecdotal evidence re: Slack is that it's displacing more email than IRC
#tantek(of course there's more email to displace ;) )
#melvstertantek: I get what you are saying, and if I were to make the call I could probably agree they are all running the 'same software' (words from your wiki) ... but I can also imagine others might not agree ... but Id say it doesnt matter much, because diaspora only interoperates with diaspora which for me is the real monoculture, until that changes ...
#tantekand I'm all for displacing of email. whether to Slack or Github issues etc. which is happening even here at W3C.
#tanteknew W3C WGs are tending to use Github issues *instead of* email, e.g. for spec discussion
#melvstermy point is that 'same software' even tho that's not defined, and if you take it literally clearly false, is a red herring ... what is more important is promoting interop
#melvsteri dont find monoculture a particularly helpful term in this respect
#aaronpkhey so I think I have a definition of "multiple implementation" that helps here
#aaronpkthe reason multiple implementations is useful is because it means no single organization or person can control things
#aaronpkwith that in mind, if a project like diaspora has 2000 forks, they are all the same implementation if the intent of the fork was to contribute back to the original
#aaronpkbut if a group decides to fork the software and maintain it separately, that seems sufficient to count as a different implementation
#aaronpksince at that point there are two different groups maintaining the software
#melvsterim actually more interested in the design decisions and architecture, multiple implementations of a useless spec is still useless
#melvstermultiple implementations of a great architecture will produce great systems
#melvsteri could write the best spec in the world for users called 'john smith' and all the john smiths out there could run it ... it still would be a poor spec
#aaronpkIRC = multiple implementations of a terrible spec. still it has, and will likely continue to outlast everything else
#tantekmelvster - I think we have different opinions on even the very assumption / framing of "great architecture" in the space of the "social web" or even just "web"
#tantekI think "great architecture" is a fallacy in this context
#melvstertantek: well i think the web itself has a great architecture, im inspired by that
#tantekWhereas I believe the more likely path to success (sustainable reliable interop) is through incremental adoption of modular building blocks, each of which is simple enough for solo webdevs to implement in < 24hrs.
#tantekmelvster: hah - the web itself has an accidental architecture at best ;)
#tantekand frankly, started with incremental adoption of modular building blocks - HTML, HTTP, URL all did not require each other, and yet worked well together
#tantekanyway different approaches - certainly different opinions as I have no way to provide a scientific explanation of preferring "incremental building blocks" over "great architecture"
#tantekwe can let multiple such attempts evolve and see which succeeds more (per sustainable reliable interop)
#melvstertantek: im a believer in incremental building blocks, so that's fine, but I *also* like great architecture, I dont agree it was accidental, I think it was well thought out
#melvsterThe way the Web spread was a piece at a time. So you could take html without taking http. So the failure of NEXT was a lesson, don’t try to sell it all at one time. Sell each piece on its own merits. Never insist that everybody take all. They will take all the pieces once they see how it fits together.
#tanteksure, I'd even say that all the failures subsequently (replacing web with semweb etc.) are a result of attempting "great architecture" at the cost of "each piece on its own merits"
#melvsterbut there was an architecture and design around the whole thing, of which the web of documents was only the first phase, indeed the web itself was an incremental release, that's where our thinking is not aligned
#tantek"great architecture" is precisely an anti-pattern because it distracts you from "each piece on its own merits"
#tantekbut as I said, that's my opinion. with "each piece on its own merits" there is no need for "great architecture"
#melvster'The web is more a social creation than a technical one. I designed it for a social effect — to help people work together — and not as a technical toy. The ultimate goal of the Web is to support and improve our weblike existence in the world. We clump into families, associations, and companies. We develop trust across the miles and distrust around the corner.'
#melvster'The web is more a social creation than a technical one. I designed it for a social effect — to help people work together — and not as a technical toy. The ultimate goal of the Web is to support and improve our weblike existence in the world. We clump into families, associations, and companies. We develop trust across the miles and distrust around the corner.'
#melvsterthat's what the WG is missing ... connections
#melvsterhence we should write up a social architecture because there's so much confustion
#tantekhence why the first such "microformat" I developed was XFN - for the connections. http://gmpg.org/xfn
#KevinMarksthe otehr key thing about the early web work was that si was building on what went before - it had ways to connect to ftp, gopher, wais, email as part of the structure
#tantekthat's assuming that a write up of a social architecture would help with the confusion - which I would assert it would not
#tantekwriting up architecture usually leads to more confusion
#melvsteryeah the layer cake is kind of a guide, i wouldnt take it too seriously, more try and grok the ideas behind it, such as using uris to connect things, that we still need
#melvsterthe most important part is 'applications'
Arnaud joined the channel
#melvsteronce you start playing with applications you can see what this can do, and most people will be surprised (in a good way)
#melvsterbut the problem is everyone built libraries and frameworks, and no one built apps
#melvsterim not an academic so i cant tell you why ... its been sad that relatively few apps were built in the sem web ... as one of the few people that have tried it, ive been amazed, hopefully others will see this too as they see it with their own eyes, i just dont know why it took SO long
#tantekthus today, it is right for us to reject them, because they never ship anything anyway
#tantekfew apps were built because they don't selfdogfood
#tantekso once the funding runs out for any particular science project experiment of an app, the app dies
bblfish joined the channel
#tantekrhiaro has plenty of experience (knows of) this
Arnaud joined the channel
#melvsteri think it depends on professor, ive noticed some places like liepzig and MIT are very productive, even with relatively few people, but you need people to get things done ... the best way to convince someone is to SHOW them
#melvsterthe next generation of the web, with social, will make you wonder why we ever used the old stuff
#melvsterwe spent 10 years largely building infrastructure
#tanteksad part is that to most people (e.g. personal interactions with friends etc.) that "next generation of the web" is just "use Facebook"
#melvsteri dont see any more infrastructure needed to be built, there's libraries, tools, parsers, frameworks a plenty
#tantekmelvster: nah, spending 10 years largely building infrastructure (not apps) means you built the wrong infrastructure, since it wasn't designed app-first, user-first
#melvsterwell i do like facebook's use of a social graph, its just a pity they have a same origin policy, but then if you were facebook, wouldnt you?
#tantekI think you'll find that as you actually try to build apps, there will be lots of things missing and just broken in your "infrastructure" that doesn't represent actual human / user needs
#tantekmelvster, Facebook can't even federate with Instagram - a wholly owned silo of theirs
#melvstertantek: ive spent the last 5 years building apps, because i need to use them, im just sad there's only a few people that do that, i self dogfood all the time, the *only* reason I use the semantic web is because it solves all the use cases other technologies cant
#tantekthey're just as stuck with their single-site-implementation
#tantekmelvster++ for building app for himself and selfdogfooding!