#social 2018-02-26
2018-02-26 UTC
mattl joined the channel
# 
melody i guess my interest is mostly in trying to determine how to evaluate what is being said when i'm missing tons of context and whether it's important that i be able to do so
xmpp-social joined the channel
# saranix well, about the person saying it: "Team member of Technology and Society Domain at the World Wide Web Consortium, where I was W3C Team Contact of W3C Web Cryptography Working Group, W3C Web Authentication Working Group, and W3C Social Web Working Group."
# saranix ... "Previously, I worked at the World Wide Web Consortium, primarily on security, in the Technology and Society Domain, but resigned over their stanardization of DRM."
# saranix ... "My dissertation studied the impact of the Web over traditionally difficult questions of meaning and reference in philosophy of language, with applications to creating search engines over heterogeneous data and colloborative tagging, now available as the book "Social Semantics". "
# saranix "Social Semantics: The Search for Meaning on the Web discusses how the largest problem facing the Semantic Web is the problem of identity and reference, and how these are the results of a larger general theory of meaning. This book hypothesizes that statistical semantics can solve these problems, illustrated by case studies ranging from a pioneering study of tagging systems to using the Semantic Web to boost the results of commercial search engines."
# saranix ... "My research is focused on discovering, formalizing, and empirically testing the social and philosophical assumptions inherent in technology using the methods of data science: From the Web to blockchain technology, at the heart of every technical problem is a social problem waiting to get out."
# saranix interestingly, my impression of his work is that he treats every social problem as something that can be solved with technology
# saranix as a consultant, the first thing I say to small businesses is "It's not a technical problem"
# saranix "Currently, my work is centered on how protocol design can enable collective intelligence in adversarial environments." -- my interpretation "blockchain is god" ;-)
# saranix given the subject matter of his book, I find it odd that he advocates statistic inference to grab context (which can never work) but then poo-poos RDF which is explicit reference rather than inferrence
cwebber2 joined the channel
# 
rhiaro saranix: yeah, harry chartered and fought hard for the SWWG in the first place. I think he's disappointed it didn't turn out exactly as he imagined. But we did the best we could under the circumstances. His stories about the group are a little stuck on the first year when he was around, he missed how much things changed in the years after.
gobengo-irc-bot_, gobengo-irc-bot and bengo joined the channel
# melody i guess maybe i'm more interested in like, what were the other possibilities -- like i don't really know who the "factions" were trying to have their interests represented or like, even which specs are being vagued about in the tweeter thread, none of them were mentioned by name
# melody and also like, even after reading the linked blog post i still don't really understand what the decision was that was made which made diaspora just refuse to even sit down at the table to talk, it seems like there was something like, fundamentally divisive that happened which i'm not able to piece together after the fact
# saranix I can't say I'm exactly thrilled either, but I don't see how you can be disappointed about how something you started turned out if you didn't stick around to finish it
# saranix melody: re: diaspora there was no animosity, just bad timing, they had just put in 2 years of bikeshedding to update their own protocol, so they didn't want to implement another right away again
# saranix I actually don't know which post you are referring to diaspora so I may not be fully informed either
# saranix but just from the convos I've seen, that's how it went down
# melody while it didn't seem like there was animosity it did seem like more than just "don't want to implement another thing" based on https://schub.io/blog/2018/02/01/activitypub-one-protocol-to-rule-them-all.html which was the post that kicked off the whole twitter thing
# DenSchub we surely are not opposed to implementing something different. but as i pointed out in my post, it would create problems for us, so it's hard to convince people to invest a lot of time when we have something that, at least for us, works better
# DenSchub the twitter thing was just people getting mad for the sake of getting mad.
# melody i wish there was someplace i could go to find information about the risks, holes, limitations and poor use case fits for these different federation & decentralization protocols
# melody everyone only talks about the benefits and it makes it hard for me to find the corner cases that would impact me
# rhiaro melody: did you look at https://www.w3.org/TR/social-web-protocols? It may not be as super critical as you want, or as technical even, but just in case you missed it. Tried to go for a 'if you want to do A or you prefer B you need spec X' kind of thing where possible
# saranix as an inherent charactaristic, there isn't really 1 someplace to go... they are decentralized after all ;-) From my point of view, all of the protocols are very much lacking. Zot comes the closest to my own proto-protocol which is still in development for *looks at the date* holy crap 5 years now
# DenSchub and some "corner cases" might not even be interesting for others. for example, my whole point about private messaging is simply irrelevant if you are a network like mastodon, for example.
# saranix wonder why loqi didn't preview schub.io
# DenSchub i have new lines in my <title>, maybe it does not like that
# saranix is probably going to need a couple sittings to make it through the article
# melody yeah but there's no real way for me to figure out whether the corner cases are relevant to me if i can't find out what they are, and i don't have a perfectly formed mental model of every protocol i read
# melody to identify them myself
# saranix melody I think your best bet is just discussing with folks. especially when it comes to corner cases most is not even well documented
# melody yeah, but finding the folks to discuss with is hard
# melody and so is knowing what questions to ask
# saranix ah, well I am a consultant in this space so I may be able to help
# saranix I'm super curious what your end game is
# DenSchub okay, it wasn't the white space. sorry!
# melody saranix: complicated, and technically still not particularly well-defined, i'm pretty early on right now (for some definitions of early) trying to do some feasibility analysis for a feature set that is pretty well still in flux, to some extent the protocol(s) will shape the communication model though so i want to know like, where each falls apart so i know whether i care or not
# saranix melody: "to some extent the protocol(s) will hape the communications model" - I think if you allow that to happen it would be a mistake. That is where most existing projects failed IMO. Many of the protocols involved end up being least-common-denominator (making basically everything an edge case)... with AP, it is loosely defined so you probably wouldn't even want to implement 100% of the protocol in any case...
# saranix ... so to summarize the best approach (I think) is to define your end result, then try to work towards compatibility with the most open protocol you can
# melody so the problem is that there are actual real limits to what you can support and remain interoperable, there are types of things you just cannot do in some models, some of which may be important, and some of the things you *can* do but would break interaction with people not doing it if you did
# melody that's when knowing the boundaries becomes important
# saranix yeah. If your goal is maximum federation that can be a problem... from my POV "there are types of things you just cannot do in some models, some of which may be important" is an understatement. As I said before I find all the protocols severely lacking in functionality, the only singular exception being Zot, and my protocol which was inspired by Zot
# saranix Still trying to make AP work. So pending on that one...
# saranix I think with the advent of LD-OCAP, it will finally complete the lineup of features available in my own protocol
# saranix or...possible, not available. much hacking is still necessary unfortunately
# saranix gah, still muddled the word order on that. What is available with my protocol, will then become theoretically possible with AP
# saranix I sometimes wish I could just zap ideas from brain to brain instead of trying to translate to words first. Kind of worse than translating to federation protocols ;-)
# melody maximum federation isn't a goal for me, but some level of mastodon compatibility is likely going to be a must -- but an open question for me is whether it makes sense to build with AP in mind, look elsewhere, or reinvent my own data model + communication model with a federation layer very distinctly separate (and what challenges doing that will pose)
# saranix What I have observed to be most successful, is the latter
# melody yeah, that poses a lot of issues for me though
# saranix like?
# melody idk, i think i reached a point of babbling, maybe it just raises more questions i don't have answers to yet
# saranix I'd like to help you get there
# saranix Is there a feature list you have in progress? Is there an audience defined yet? Are there other stakeholders that you have to consider?
# melody potential stakeholders and governance and legal structure are a big piece of this but being worked on largely in parallel and separately from the completely raw tech feasibility questions i'm trying to answer
# melody very long feature list in progress but a lot of it is really focused on stuff that could technically be done in the client but would often be much better if supported at a deeper layer of the stack, audience has a lot of demographic overlap with mastodon but i'm hoping to capture a segment of the group that shares some of their values but were scared off/found the anti-abuse, moderation, and anti-harassment features insufficient -- lots of
# melody those questions are going to impact the tech needs but i have enough of a vague outline of what the stakeholders might look like and what their concerns might be that i'm confident in my ability not to bury myself in dead-end tech without waiting on a finalized structure
# melody being vague because it's hard not to be, there aren't like, NDAs involved, but a lot is still up in the air
# saranix hmm... yeah I feel like I'd have more to say if I knew some juicy deetz. We can talk in private if you like...
# melody that would be easier for me
gobengo-irc-bot_ and gobengo-irc-bot joined the channel
# saranix melody "Check out this cool thing: libre.net/foo"
# saranix oops
cwebber2 joined the channel