2014-04-06 UTC
paulcp and snarfed joined the channel
# 00:08 snarfed i don't think i'm going to make it to portland, but i'll see what i can do # 00:08 snarfed i'd happily support someone else giving a talk about backfeed and including bridgy! tilgovi, scor and pfenwick joined the channel
mydjey, squeakytoy, krendil, squeakytoy2, snarfed and ben_thatmustbeme joined the channel
# 02:10 kbs speaking for myself at least, just refining some of the crawling stuff # 02:10 kbs and starting to think through some ideas for private messaging # 02:11 kbs how 'bout you? Anything exciting brewing? tantek, squeakytoy, squeakytoy2 and pfenwick joined the channel
# 02:54 ben_thatmustbeme also trying to convince a friend of mine to take a look at marking up his site with mf2 at least and get him doing some indie web stuff scor joined the channel
# 02:59 kbs could I ask [out of curiosity] how you're planning to set up webmention? manually curating it, or something else? # 03:15 gRegor` Evening ben, kbs # 03:20 kbs ah well - I think the theory of rel=me is not matching the practice in a helpful way :) # 03:21 kbs so in theory, a rel=me link can only appear on a profile page # 03:21 kbs ie - both the source and the target have to be a profile page # 03:21 gRegor` Well, for what purposes? Webmention? # 03:22 kbs oh, just for finding out all the 'validated' pages for a given profile # 03:22 gRegor` If it's authorship, the rel=me should have an h-card, yes. Is that what you mean? # 03:23 kbs sort of more like what indieauth needs to do - given a starting profile page, find all the linked profile pages with a chain of links back to the starting profile pages # 03:23 kbs so for example, if I give it snarfed.org/about # 03:23 kbs I expect it to give me back ryan's twitter page, g+ page and so on # 03:24 kbs but [to take that example] although [in theory] snarfed.org/about is ryan's "profile" page - the rel=me links on the twitter and g+ page actually point back to snarfed.org rather than snarfed.org/about # 03:25 kbs so what's happening [I think] is that while in theory only "profile" pages should have a rel=me link -- both as the source and the target -- in practice, this is not the case # 03:26 gRegor` Well, personally I have the rel=me links on every page of my site. # 03:26 gRegor` I have an "about" page which I guess is analagous to a "profile", but I've not really used that term. # 03:26 gRegor` I'm not aware of it being a problem to have rel=me links from multiple pages # 03:26 kbs as I learn't earlier today - I think the theory is that you'd use a rel=author page on every site to the about page # 03:27 kbs rel=me is actually used to identify equivalent profile pages, rather than authorship - if that makes sense # 03:27 gRegor` I have them in the HTML of the footer, but am going to move them to <link> elements in the header # 03:28 kbs therefore, the only page that ever has a rel=me link is a profile page # 03:28 kbs (both for the source, and target) # 03:29 kbs I thought so too earlier - but this is actually the difference between rel=me and rel=author [otherwise, having two such tags would not be useful] # 03:29 gRegor` I mean, I see what you're saying, but I don't see a problem with having my rel=mes on every page # 03:30 gRegor` Isn't rel=author only absolutely necessary if you don't include the author h-card on the page? # 03:30 gRegor` I do the latter. # 03:32 gRegor` I guess it violates DRY, technically, though it just appears in one template file that's re-used. kbs` joined the channel
# 03:33 kbs Well - it's no problem for anyone visiting the page # 03:33 kbs but it indicates [to machines anyway] something that's probably not what you intended # 03:33 kbs rel=me indicates that the current page is a profile page # 03:33 kbs and that the linked target is equivalent to the current profile page # 03:33 kbs` rel=author indicates the current page was authored by the target page # 03:34 kbs` anyway - it's something everyone is doing in practice # 03:34 kbs` which is why I was thinking that the distinction between rel=me and rel=author seems not to be true in practice # 03:34 gRegor` The use of "profile" is confusing to me, because it's not a very specific term. # 03:34 gRegor` "rel="me" is used on hyperlinks from one page about a person to other pages about that same person." from the microformats wiki # 03:34 kbs` profile basically is just whatever you'd think of as an "about page" # 03:35 gRegor` Every page on my site is "about" me # 03:35 gRegor` Not a strict profile, in the sense of a FB / Twitter profile. # 03:35 gRegor` The wiki gives the example of tantek's homepage with rel=me to his Twitter. # 03:35 kbs` yes - but tantek is a weird case because in fact his homepage happens to be his about page # 03:36 kbs` you'll probably notice that none of his other pages have rel=me links # 03:38 kbs he uses rel=author for all the other pages, pointing back to his "about" page, which happens to be his home page too claytor joined the channel
# 03:39 gRegor` Just noticed werd.io points his rel=author to a humans.txt # 03:40 gRegor` But he has an h-card on each post, so not a problem for pulling author info. Interesting choice, though. # 03:41 kbs h-card is not generally an issue, I found # 03:41 kbs but the theory of rel=me is not actually true based on how people are using it. Essentially I have to pretend rel=me and rel=author are equivalent, for all practical purposes # 03:42 gRegor` Sorry, guess I'm still not following you. # 03:42 gRegor` If I'm parsing a page for mf2 to pull in webmentions, for example, I look for an h-card associated with the post as the author. If I don't find it, I look for rel=author and pull in the h-card from there. # 03:43 kbs so it's not about looking for a h-card # 03:43 kbs it's about finding all the sites that you "own" # 03:43 gRegor` rel=me hasn't entered into it for me. But I've only done webmentions # 03:43 gRegor` maybe you mean more for indieauth # 03:43 kbs yes - more related to what indieauth does, but more about this # 03:43 gRegor` So you should go to the rel=author page and pull in the rel=me links from there, right? # 03:44 kbs that's what I'm doing to make it "work" but that's not the theory :) # 03:45 kbs to put it differently - if I see a page with rel="me" on it, it immediately means that it's a profile page # 03:45 kbs [and that the target page is also a profile page] # 03:45 kbs here I use profile in the sense of "about" # 03:46 gRegor` I guess. Because some sites do that. I don't think it makes it a hard and fast case that it should be that way. pfenwick joined the channel
# 03:47 kbs the reason why it matters is something like this. # 03:47 gRegor` Is there rel=profile? :) # 03:47 kbs probably may as well - everyone is piling on random stuff into rel= :) # 03:47 kbs I start with gregorylove.com/about # 03:48 kbs github takes me back (say) to gregorylove.com # 03:48 gRegor` No y ;) But yeah # 03:48 kbs I can no longer prove that gregorylove.com/about is associated with github squeakytoy2 joined the channel
# 03:48 kbs because there's no rel=me from gregorylove.com to gregorylove.com/about # 03:49 gRegor` I guess I see how it can be a problem with domains that have multiple users / authors. # 03:49 gRegor` It doesn't seem to be a problem for single-author sites like mine (and most of ours) # 03:50 gRegor` But I guess if we're building general-use scenarios, yeah, that's a problem. # 03:50 kbs it's because of an accident of implementation I think - when you use indieauth, you probably provide gregorylove.com rather than gregorylove.com/about # 03:50 gRegor` E.g. systems shouldn't assume single-author sites # 03:50 gRegor` I put rel-me links in the footer, hidden, so I couuld use indieauth, yes. # 03:50 gRegor` That's all they're intended for, so far. # 03:51 kbs you should ask kevinmarks or tantek when they show up next for the distinction - I had a long interesting conversation with kevinmarks today afternoon when I finally realized what the theory was, and also why it's almost irrelevant because nobody actually makes this distinction # 03:52 gRegor` I'll look it up in the logs # 03:52 gRegor` (I meant, on my site, that's all they're intended for, so far. Just to be clear.) # 03:54 gRegor` What time was the chat? # 03:55 kbs ah, let me look it up and paste some links to the time # 03:57 kbs ending at 12:23 - basically the key point [for me] rel=me should occur only on an about page, and point to an about page # 03:58 kbs and then some discussion about why this doesn't really happen in practise, etc # 03:58 Loqi kbs meant to say: and then some discussion about why this doesn't really happen in practice, etc # 04:02 kbs so also in theory, apparently we should all be logging into indieauth using our /about pages, should our "about" page actually not be the baredomain itself. I think part of the confusion is probably because tantek uses his baredomain to represent what we'd all consider an 'about me' page # 04:02 gRegor` Hmm. I think I get the gist of it, but still a bit confused. I'll think about it some more. :) # 04:02 gRegor` I guess take my site for example: How/why would you start with gregorlove.com/about instead of just gregorlove.com? # 04:02 gRegor` Like, what's the use case? # 04:03 kbs so my specific use case is - I have a little tool that given a url, starts looking for associated profiles. # 04:03 kbs ah, let me link to an example, one sec # 04:04 kbs that's the actual experiment I'm tinkering with # 04:05 kbs and I say, hey, this person looks interesting, find out more about him or her # 04:05 kbs and let's say, I end up in your /about page, and decide I'd like to know more about you # 04:06 kbs The tool starts crawling all the rel=me links, and checks that they eventually point back to the starting point [that's how I can tell it's "legit"] # 04:06 kbs right now, if I crawl your baredomain, it "works" - because you've linked it in such a way that I can see your associated domain(s) # 04:08 kbs Ideally, I'd do something quite simple - find the 'author' on any page # 04:08 kbs then chase rel=me links from there - which also doesn't quite work if I happen to start from your /about page # 04:08 kbs so I have to do some unholy hacks of merging rel=me and rel=author links and look out for specific domains before anything really works as intended # 04:11 kbs I do think that any reasonable person would probably start with the baredomain for any indieweb site # 04:11 kbs and that also does work simply because people just want to get indieauth working :) # 04:14 gRegor` Well, it makes an interesting case for me adding rel=author, which I don't have anywhere currently. # 04:15 gRegor` Is the "ideal" situation for you that my post permalinks would have rel=me to /about, and on /about I would have rel=me to external profiles? # 04:15 gRegor` With those external profiles linking back to /about? # 04:17 kbs gRegor`: yea - that's the intended purpose of rel=me links anyway, if you want to be precise about those tages. But as you also pointed out right at the top - in practice people aren't making this distinction # 04:17 kbs so I think any tools that actually use the data are having to make some adjustments # 04:17 Loqi kbs meant to say: gRegor`: yea - that's the intended purpose of rel=me links anyway, if you want to be precise about those tags. But as you also pointed out right at the top - in practice people aren't making this distinction # 04:17 kbs it'd also mean that you'd have to login with /about rather than the baredomain into indieauth - I don't think it's chasing rel=author links # 04:18 kbs would a rel=me linked bare personal domain, also have a separate about page? # 04:18 kbs gets tantek's style of using baredomain as the about page # 04:22 GWG I've come in in the middle of this. But it seems like an interesting discussion. KevinMarks2 joined the channel
# 04:24 kbs what would the appropriate rel=me target be, for those who use a separate /about page on their personal domain? # 04:25 kbs kevinmarks is the best test-case I've had for rel=me link chasing :) # 04:27 KevinMarks I'm just a Zaphod Beebblebrox - didn't you hear I come in six packs? # 04:29 kbs I think the confusion is also because both kevinmarks.com and tantek.com use the baredomain as their 'this-is-me' page. But (for example) ryan uses snarfed.org/about - so as I understand it, ryan would be linking his various rel=me targets from and to snarfed.org/about # 04:29 kbs and rel=author from everywhere else to snarfed.org/about # 04:29 GWG But kbs, to go a bit philosophical...aren't you more than just your bio? # 04:30 kbs (or, he could mutually rel=me link snarfed.org and snarfed.org/about I guess, dunno) # 04:30 GWG For example, my problem is that I am an author all over the place. # 04:30 gRegor` kbs: I guess depending how far down the rabbit trail you want to chase rel-me links, if you started with my /about, followed the link to my github, followed the link back to my root /, then followed the rel-me links from there, it would finally resolve. :) # 04:30 GWG But my personal site is just...me. # 04:31 gRegor` accidental italics, heh # 04:31 kbs gRegor`: I don't think you have a rel=me from baredomain -> /about # 04:31 GWG I'm trying to make it more than just an about page, which is why I'm here. # 04:31 gRegor` You don't need one. # 04:31 gRegor` You basically do two loops # 04:31 gRegor` . /about => github => / => github again => / and you have a match. # 04:32 kbs gRegor`: that matches / to github # 04:32 gRegor` I'm saying follow rel=mes until you get the match # 04:32 kbs gRegor`: yes - I know ;) but the match has to be on the actual page, not a parent # 04:32 gRegor` I think it's a pretty strong indicator, since it's in the chain # 04:33 kbs it makes perfect sense for a personal domain, but not in general # 04:33 gRegor` For your purposes, it would list "gregorlove.com and github.com/gregorlove" as validated. Not /about # 04:33 gRegor` Which should be fine. # 04:33 gRegor` Either it ultimately comes up with a match or it doesn't # 04:34 gRegor` You could then go through the rel-author method if that didn't match, I guess # 04:34 gRegor` You do multiple loops like that? # 04:34 kbs I chase down all the links, finding loops # 04:34 KevinMarks there is s presumption that if you own / you own/about surely # 04:35 KevinMarks I think the separate about pages would make more sense for something liek techcrunch # 04:35 kbs KevinMarks - hm - does that work for something like a hosted site [like github] # 04:35 kbs someone could arguably claim / without necessarily claiming ownership over /user ? # 04:37 kbs to first separate out the 'how to make it work' from 'what should the rel tags be' discussion :) to take a specific example, let's say snarfed.org is a personal domain and snarfed.org/about is the 'about-me' page # 04:37 kbs in theory at least, should external rel=me targets go to snarfed.org, or snarfed.org/about ? # 04:38 gRegor` I think if it points to snarfed.org, the implication is "this profile is associated with the entire domain and sub-pages" # 04:38 gRegor` If it's a multiple author site, then the rel-me links should go to a specific user page. # 04:39 gRegor` So in absence of rel-author elements on my site, once you follow the loops and verify my root is associated with the external profiles, you can implicitly assume the original /about is as well. # 04:39 gRegor` Er, rel-me should be rel-author two chats up. # 04:40 gRegor` Corrected: If it's a multiple author site, then the rel-author links should go to a specific user page. # 04:40 kbs I do agree that if one assumes claiming ownership of a rel=me over a url is equivalent to claiming ownership over all urls with a longer suffix, that's indeed the case # 04:41 gRegor` That said, I don't like the idea of making systems have to go through loops like that - 4 HTTP requests - just to confirm my /about is indeed associated with my twitter. :) # 04:42 kbs it's isn't clear to me anyway, that the rel=me links are intended to claim anything like that # 04:42 kbs though in practice that's what I effectively end up doing... # 04:43 gRegor` True. I believe we're talking about additional information that can be extracted from rel-me relationships, outside their base purpose. pfenwick joined the channel
# 04:44 kbs it's a tricky, because it's used for validating claims about site ownership [eg: indieauth] and my cynical mind usually jumps to ways this could be abused squeakytoy joined the channel
# 04:46 gRegor` GWG: implicit information about paths below the "verified" link. E.g. my twitter profile links to gregorlove.com/, so once that's validated you can assume my twitter is also associated with my /about page. # 04:47 GWG Indieauth uses that, so does Google, but do many others use it to prove anything? # 04:47 kbs is planning to as well :) hence wondering all the places he's gonna screw up # 04:48 gRegor` Not yet that i know of, but kbs is working on something :) # 04:48 gRegor` Is this related to your PGP thing? # 04:48 kbs somewhat tangentially - it's the initial set of sites I'll consider for fingeprints # 04:49 kbs rel=me is supposed to point to a profile page # 04:49 kbs and not everyone has their profile page sitting at the root of their 'set of owned pages' on all sites # 04:50 gRegor` Here's my main question: Why do you need to verify gregorlove.com/about is associated with my Twitter, when you can follow the rel-me loop and verify that gregorlove.com is associated with my Twitter? # 04:50 kbs gRegor`: that happens only if I happen to start with the /about page to begin with basal joined the channel
# 04:50 kbs the hope is to start with any page # 04:51 kbs in your specific case, if I happen to start with /about, just because I look at it and decide to find out more about you # 04:51 GWG But, once you've decided I'm verified, what do you do with that information. # 04:51 gRegor` I don't have rel-author links anywhere # 04:52 gRegor` So I guess I'm out of luck for now. :) # 04:52 kbs gRegor`: well, if I happen tostart with the /about page on your domain, actually I'm out of luck I guess :) # 04:52 gRegor` I guess my site is mostly an exception, though, currently # 04:52 kbs no no - I think ryan also has a somewhat similar thing going on # 04:52 gRegor` kbs: You don't have to be, though. I'm saying follow the rel-me loops until you verify *something* # 04:53 gRegor` You'll end up verifying gregorlove.com, so gregorlove.com/about - where you started - is implicit # 04:53 kbs gRegor`: yes - using your assumption, you're quite correct # 04:53 kbs so here's where I'm kinda thinking all the ways things could go wrong # 04:53 kbs I put a rel=me link to plus.google.com # 04:54 kbs I guess I'd need to find the smallest subset of sites or something, and it still doesn't 'make sense' to me :) plus.google.com doesn't have any profile info, and I shouldn't be taking that sort of link into account addal joined the channel
# 04:55 kbs should I assume that my site is connected to all of plus.google.com? # 04:55 gRegor` No. You would look for rel-me links on plus.google.com, find none, and halt # 04:56 kbs why would I look for rel=me links on plus? # 04:56 kbs I start with my g+ subdomain page # 04:56 gRegor` Sorry, think I got confused. # 04:56 gRegor` You mean literally "plus.google.com" or your g+ full profile link? # 04:57 gRegor` So you start with the g+ and want to verify the domain it's associated with? # 04:57 kbs yea, sorry . I start with plus.google.com/+user -> somedomain -> plus.google.com # 04:57 kbs I want to find the pages that are 'owned' by the same user # 04:58 kbs ... have to disappear here, ... will think more too :) # 04:58 gRegor` 1) Find rel-me from g+ profile and load that page A 2) Look for rel-me in page A that has the base domain you just came from ('plus.google.com'). If it matches where you just came from, verified and halt. # 04:59 gRegor` 3) If no rel-me found on page A, look for rel-author on A and load that page B # 04:59 gRegor` 4) on page B, repeat #2 # 04:59 gRegor` I'm getting tired and probably not making sense. Haha. Especially with things hard to describe in text. :) # 05:02 gRegor` sleep well, Loqi tantek, paulcp and kbs joined the channel
# 05:50 kbs tantek: wish you were here earlier to clarify rel=me usage :) # 05:50 tantek hey kbs - did you go through the docs on microformats.org/wiki/rel-me ? # 05:51 kbs I thought I did, but I probably missed something I'm sure - let me reread # 05:51 tantek is this about rel=me in general or RelMeAuth / IndieAuth in particular? # 05:51 kbs rel=me in particular. If someone has a personal domain with a separate /about page - should all the external rel=me targets point to the /about page? # 05:51 tantek there's also microformats.org/wiki/relmeauth if that helps # 05:52 tantek and then the links from the home page to the /about page should have rel=me # 05:52 Loqi tantek meant to say: and then the link from the home page to the /about page should have rel=me # 05:53 kbs ok, I've still not grasped rel=me clearly :) # 05:54 kbs so, it's not the case that a rel=me target points to a page that represents "me", necessarily # 05:55 kbs because the bare domain is not a page that represents "me" it's actually the /about page, no? # 05:55 tantek no - because everything at your home page and below represents "you" # 05:56 tantek thus your home page acts as the common pointer for you # 05:57 kbs in the /about example, would there also be a rel=me link from the /about page back to the bare domain? # 05:58 tantek kbs - you could, but it's also implied in the rel-me reciprocation algorithm squeakytoy2, basal and KartikPrabhu joined the channel
# 06:00 kbs hm, trying to parse that :) I see that / -> /about is an explicit rel=me link - how does /about go back to / ? # 06:05 kbs so to put it differently - would the relmeauth algorithm still work if I input /about as my user identity URL? # 06:06 kbs [and the external sites link back to the bare domain] # 06:13 kbs so gRegor` was pretty much spot on - I'm the confused soul :) # 06:16 GWG I'm still trying to figure out how to get people to care about anything I put on a site in the first place. # 06:17 GWG Which is where the whole POSSE thing comes in. # 06:18 GWG Except all they ever do is 'like' it on Facebook. squeakytoy2 joined the channel
# 06:20 tantek kbs - the bigger challenge is that no one's implemented the sub-page part of the relmeauth algorithm yet either # 06:20 tantek current implementations all depend on using just your home page to make it work # 06:21 tantek which might make more sense to standardize on since it's just simpler # 06:21 tantek and in the long run it's not really worth the extra code to support folks who are going out of their way to make extra work for themselves # 06:21 tantek but wanted to allow for it in the algorithm in case there was interest in developing it # 06:22 tantek it just makes more sense to people in general to have domain name represent them. not domain/about # 06:22 tantek e.g. companies are company.com not company.com/about # 06:23 kbs so the thought about that are companies that host users # 06:23 tantek so it's ignorable from an indieauth perspective # 06:26 tantek so we *could* try to make indieauth work with entering your domain/about - but no one's presented sufficient evidence that it's worth it - so meh # 06:26 kbs hm. I guess as a model of domain identity, I'm trying to follow github.com as owning github, and for me to own github.com/user [ie, as being the target of appropriate rel=me urls] # 06:26 tantek though the algorithm says theoretically how it could work # 06:27 tantek kbs - why should indieauth support silo-auth use case like that? quite the opposite, not worth spending any time on it. # 06:27 kbs not indieauth - just trying to understand what the target of a rel=me means # 06:27 tantek I mean, ok for github to be an ephemeral auth provider, just not a canonical identity # 06:27 tantek kbs - rel-me has to be symmetrical in order to "work" # 06:28 tantek and as such, can theoretically work between any two URLs that rel=me link to each other # 06:28 kbs yes, but now there's an implied understanding that linking to a top-level rel=me also gives ownership of all paths under that top-level rel=me target # 06:28 tantek if you can control the root, you can typically control everything under it # 06:30 GWG I think there needs to be a visual. # 06:30 kbs so the my rel=me link to github.com/user doesn't really mean I control github.com/user -- it's just used to make indieauth work # 06:30 GWG Maybe someone will create a visualization generator. # 06:30 kbs since as a practical matter, github.com controls github.com/user # 06:32 kbs I'm realizing that my confusion is probably stemming from the dual purposes rel=me is being used for :) 1. "about me" and 2. "auth" # 06:33 GWG I thought rel=me was being used to represent that both resources referred to the same individual. # 06:35 kbs but if I rel=me to youtube.com/user, youtube.com could rel=me mutually to google.com (say). Now, who really "owns" youtube.com/user is a bit perplexing to me # 06:36 kbs [because of the "all paths under" rule] # 06:37 kbs so if someone proves they own youtube.com, they also own youtube.com/user (which makes sense in a way) # 06:38 kbs eg: if I have myfamily.com and myfamily.com/son and myfamily.com/daughter -- the 'ownership' question becomes a bit foggy to me # 06:39 kbs the son and daughter can't have bare links back to the root from their g+ pages # 06:39 kbs but if the son links back to myfamily.com # 06:39 kbs it seems to me that he could login as the daughter # 06:39 GWG I think you may be overthinking it. # 06:40 kbs is just confused as heck :) and apologizes profusely # 06:40 kbs yes, that would work for the bare-domain # 06:40 GWG daughter.myfamily.com and son.myfamily.com # 06:40 kbs *nod* yes, that would be just fine # 06:42 kbs I guess the thinking is that an indieauth'able site must be sub-domainable # 06:42 kbs can't have different users under different paths # 06:46 GWG Things keep getting more complicated. # 06:48 kbs I think I'm realizing how little I know # 06:48 GWG Well, let me reframe the question KevinMarks joined the channel
# 06:49 kbs given a url, find all the 'valid' profile pages for the author of the content at that url # 06:49 GWG But at the end of the day, what do you want to accomplish? # 06:49 kbs I dunno - don't have anything more concrete than that at the moment :) # 06:50 kbs I'd like to type in tantek.com, and pull up all of his twitter, g+, dribbble, blog sites, github etc # 06:50 kbs to satisfy my curiosity about what tantek does, etc # 06:51 kbs I think I have something that I can make work at the moment # 06:51 GWG All that takes is a little analysis # 06:51 GWG All those links should be on the page # 06:51 GWG All you need is a form box and a script # 06:52 kbs but in the course of doing it, I noticed that I needed to adjust and refine my understanding of what rel=me actually means KartikPrabhu joined the channel
# 06:53 kbs and I wanted to go back and find out what it is _supposed_ to mean, just out of curiousity and also to see how badly I've botched my algorithm # 06:54 kbs and out of that came my poking around at what relmeauth does, and an ocean of puzzlement, etc :) # 06:56 kbs :) and my depleted store of brain cells aren't helping # 06:56 GWG Well, no specification can handle everything. # 06:57 kbs it's curiosity at this point I guess - [thanks KevinMarks for the pointers to the code, I've been mining the site-specific code for ideas] # 07:01 kbs at any rate, serves me right for thinking I had rel=me figured out :) # 07:02 GWG I'm still working on figuring out the philosophy. The code will come, hopefully. # 07:03 GWG Especially since I'm a horrible coder. # 07:03 kbs do let me know the philosophy as you figure it out :) much more entertaining anyway # 07:05 kbs at least, I _think_ one thing is that family.com/son and family.com/daughter doesn't play nice with indieauth - but it works with subdomains, as GWG poitns out # 07:06 GWG I've always found understanding concepts to be more important than protocols for how I approach things. # 07:08 KevinMarks mu uncle set up marks-family.co.uk with really bad URLs for all of us plus frames layout # 07:08 GWG KevinMarks, how much do you consider that part of your identity? # 07:10 GWG When I arrived first in this room, I compared myself to a multiple personality. # 07:11 GWG Isn't part of the goal of establishing relationships between pieces of your web activities to tie elements of your identity together? # 07:11 GWG Admittedly, in some cases, that is exactly what you don't want to do...but... glennjones joined the channel
eschnou joined the channel
# 07:18 kbs if you selectively go marks-family.co.uk/kevin [rel=me] g+/kevinmarks [rel=me] marks-family.co.uk (bare domain) - do you also get to assert indieauth ownership over marks-family.co.uk/dad ? :) # 07:20 GWG Okay...that's just...not a good URL # 07:21 GWG But, your identity seems fairly well established. # 07:22 GWG Unless people confuse you a lot with the Stargate character. # 07:24 GWG But you appear in enough places to easily establish your identity when needed. # 07:24 GWG I have a friend who has the same name as a major sports personality. # 07:24 KevinMarks "Sales Representative for Harley-Davidson Footwear at Wolverine World Wide" # 07:25 GWG I only hate myself when I google myself. # 07:25 GWG But that is mostly because of what surfaces at the top. # 07:25 kbs oh, your uncle must be doing a geneology thing I guess - that's what shows up when I search for the path # 07:27 GWG I seem to recall a documentary where someone went to find a bunch of people with the same name as them. # 07:28 GWG Google Me, I believe it was called. # 07:29 KevinMarks google me being a name we were denied for profiles by marissa # 07:29 GWG Google Profiles annoy me greatly. # 07:31 GWG KevinMarks, it isn't the idea of them. It is the fact I have a gmail and an apps address, and now under the new Google Plus philosophy, every account wants me to create a Google Plus profile, and I can't tell the thing I already have one, and to use that one and stop asking me to join. # 07:32 GWG But the Google Plus revolution was long after you were involved in Profiles. # 07:33 KevinMarks and they promised they'd have fixed the apps vs accounts problem by now # 07:33 GWG But, you'd think someone over there would think you might have two email addresses and only need one Google profile. # 07:34 KevinMarks the problme was you could have 2 google accounts with the same email address but different passwords # 07:35 GWG Well, I know how hard some of these things are to manage, and how many different problems Google is trying to solve. # 07:35 GWG That doesn't mean it doesn't annoy me when I'm being asked to set up multiple Google wallet accounts. pfenwick joined the channel
# 07:36 GWG When I can't comment on an Android App because I now need a Plus account, I buy apps through my gmail account, and tied the Google Plus account to my apps account # 07:39 GWG kbs: I've spent the last few days trying to figure out how to get people I interact with out of their silos and over onto my site. # 07:40 GWG So, I've been trying to gradually ramp up posting on said site and syndicating it over to the various networks I set up accounts on, backfeeding the responses to my site. # 07:41 GWG So, I have webmentions and the related indieweb tech starting to be set up. # 07:41 GWG Except the problem with silos is that people like them # 07:42 GWG So, the goal is for me to own my content... # 07:43 kbs I'm not much of a social-posting-type guy unfortunately # 07:43 kbs but I have a feeling there are enough here who'll probably be able to share their experiences with that aspect of it [interaction outside silos.] # 07:43 GWG But engagement requires people, who want to hang out on social networks... cweiske joined the channel
# 07:44 GWG kbs: I just...if I'm going through the effort to put something out there, I'd like people to actually see it. Otherwise...I might as well talk to myself. # 07:44 GWG I don't want to have the readership of a major blog or newspaper. I'd like to at least engage those people I'm affiliated with. paulcp joined the channel
# 07:46 GWG So, the point of syndication to a silo, as Willie Sutton allegedly said, is that is where the people are. # 07:46 kbs not knowing too much about this [I'm a newbie here as well] is it that you aren't able to pull and push comments in and out of (say) facebook easily? # 07:47 GWG Apparently, clicking like is easier for people than expressing an opinion. # 07:48 kbs oh, that's interesting - I'd somehow have assumed likes would be technically easier to pull than sync comments both ways # 07:48 GWG At the moment though, I'm working more on the creating of the process than the creation of the content. # 07:48 GWG Forest for the trees type problem # 07:48 GWG What's the point of building a nice house if you aren't going to live in it # 07:52 kbs are you already on the irc people list? just curious to see what sort of things you write about. # 07:52 GWG kbs: Not yet. I have about 5 sites. # 07:52 GWG And each one completely disconnected from the others. # 07:53 GWG But, the one I'm working on now is a bio site. # 07:54 GWG Basically, I'm going to consolidate a lot of randomness there. # 07:54 GWG Anyway, on that note, I'm off to bed. # 07:55 GWG kbs: I've been struggling with this problem for a while. I have a lot to say, but I can't keep it in order. # 07:55 kbs be interesting to hear what you've figured out # 07:55 GWG But I'll have the first pieces up. grantmacken, friedcell, friedcell1 and glennjones joined the channel
rtaibah joined the channel
# 09:15 rtaibah I’ve just install the WP webmention plugin on my blog http://rtaibah.com/blog the FAQ’s are quite lacking. How do I webmention somebody? catsup, glennjones_ and melvster joined the channel
bupkes joined the channel
# 10:13 Loqi bupkes: tantek left you a message on 4/2 at 1:15pm: Welcome! And nice domain name :) # 10:15 bupkes !tell tantek thanks :) # 10:15 Loqi Ok, I'll tell him that when I see him next # 10:34 bupkes hello all. I'm a non-coder but am ploughing through codecademy JS course to try and get some grip on things. # 10:35 bupkes I've got POSSE, webmentions working on my site bupk.es, thanks to some WP plug-ins and things. # 10:36 bupkes Jolly interesting stuff. Looking forward to being able to build some things to share :) scor, pasevin, CheckDavid, cweiske, rtaibah, krendil and barnabywalters joined the channel
# 13:40 GWG But corporate training is the big part of it # 13:43 GWG I was here late last night, contemplating form over function # 13:46 GWG That the need to perfect the functionality of the site creates an attractive site with webmentions and what have you....that people don't come to # 13:49 ben_thatmustbeme basically if friends want to read your content they have to click through to your site, not just read and comment on the syndicated site pasevin joined the channel
ben_thatmustbeme joined the channel
# 13:58 ben_thatmustbeme there we are, properly seating the power cable on one of my hard drives is always good # 14:00 ben_thatmustbeme apparently my computer does have an internal speaker... didn't think it does, but it certainly let me know # 14:01 GWG Without it...how will you hear beeping? KevinMarks3 joined the channel
# 14:06 GWG I'm flashing back to when I used to put in a card for that CheckDavid and pasevin joined the channel
# 14:18 ben_thatmustbeme grr, for some reason php isn't installing any modules, so no curl... which makes testing webmention parsing difficult rtaibah joined the channel
pasevin_ joined the channel
kbs joined the channel
# 14:41 kbs sorry about the rel=me blathering :) KevinMarks2 joined the channel
paulcp and ttepasse joined the channel
# 15:46 GWG So, anytime anyone talks Indieweb...Loqi pastes it here? # 15:56 aaronpk using twitter's realtime search api, so it shows up pretty quick, often even before a webmention of a posse'd tweet kbs joined the channel
# 15:59 GWG I once again will say I feel like I could create a full hobby out of setting up this tech, and as a result, not have any content to POSSE # 15:59 GWG The barrier to posting stuff online is that you wonder if anyone is interested. # 16:00 GWG "There wasn't any hot water in my shower again." # 16:00 GWG How many of those interested you? # 16:01 aaronpk if I knew why the omelet was significant it may interest me # 16:01 GWG Well, for a while there I was chronicling my adventures with induction cooking. # 16:01 GWG Or rather...my triumph at not burning things # 16:02 GWG But, vanity is an age-old argument in social sharing # 16:02 aaronpk don't forget you can always reply to other peoples stuff too, and you know, have conversations # 16:02 GWG I haven't figured out how to pull in conversations I didn't initiate # 16:03 arcatan setting up the indiewebcamp seems like a lot of work, but to be able to take part in conversations might make it worthwhile # 16:03 arcatan (/me isn't yet there) # 16:03 arcatan and by the indiewebcamp, i mean support for indieweb stuff like webmention. weird typo. # 16:04 GWG aaronpk: I meant from other services. # 16:06 aaronpk GWG: not sure what you mean. have you seen how others are doing replies? # 16:07 aaronpk sure, compare my replies to a tweet and also directly to barnaby: # 16:09 GWG How did the tweet end up on your site? # 16:09 GWG You can't be pulling in every tweet? # 16:09 aaronpk but it works for anything i'm replying to, not just tweets # 16:11 aaronpk it's basically just parsing the URL I reply to for an h-entry # 16:11 aaronpk with twitter there's a special case for converting a tweet to an h-entry, but everything else is straight mf2 # 16:11 GWG Yes. But since I'm using Wordpress, I always wonder if someone has done the work for me. # 16:12 GWG Especially since I'm a horrible coder. # 16:15 kbs aaronpk: out of curiosity, do you have an analogous example for facebook [meaning, eg: commenting on posts written by others, and not necessarily public, etc.] # 16:16 GWG kbs: If it isn't public, then you get into issues of courtesy # 16:16 aaronpk kbs: I don't do anything non-public other than strictly private messaging # 16:16 kbs GWG aaronpk *nod* thanks, was curious # 16:17 aaronpk i'd love to get there, but there's a few things blocking me # 16:18 kbs nice. Be interested to see your solution when it's ready # 16:21 kbs [the twitter integrations are very neat - look forward to see how it co-evolves with the other silos] # 16:23 GWG The next question I'm figuring out is...if I know people share on various networks, how to allow them to seamlessly do so. # 16:29 GWG I have to count on people not to accommodate me pasevin and KevinMarks2 joined the channel
# 16:53 kylewm GWG: can you clarify? trying to help other people to seamlessly share your content? # 17:06 GWG People don't seem to want to come to a site and comment there. # 17:07 GWG Or mention it on their own site, as many of them don't have it # 17:07 GWG They want to bring it back to the silo they prefer wildtype joined the channel
# 17:15 GWG I'm a skeptic, basically, and I want to cover all bases. Kopfstein joined the channel
# 17:22 GWG kylewm: So I've gathered from my time here. KartikPrabhu joined the channel
# 17:30 GWG I was commenting on this earlier. # 17:32 kylewm that "I ate this for breakfast" is not interesting to other people? # 17:33 kylewm I used to keep a cooking blog but I was basically just reprinting recipes from vegan cookbooks :) # 17:33 kylewm but sometimes I still make something I want to remember # 17:34 GWG I was commenting I was worried that I'd focus on the technique and forget people come to a site for content # 17:36 kylewm kbs, yep! and here I thought everyone in the bay area had gone meat hipster :) # 17:37 kylewm I'm still trying to figure out if I have anything to say GWG, 5 years on twitter I rarely rise above bad puns or pictures of my cat # 17:37 kylewm so I'm mostly interested in the technology aspect at this point # 17:37 kbs kylewm: indeed, and therefore - cool :) # 17:40 GWG kylewm: I'm trying to figure that out. # 17:40 GWG Bear in mind, I control 6 Twitter accounts # 17:42 kbs looks forward to a new test-case for his rel=me obsession :) # 17:45 kbs nice tip on stanford inn - and looks like a very scenic location. paulcp joined the channel
# 17:51 kylewm nice! it was beautiful, i ate waaay too much :) barnabywalters joined the channel
pasevin joined the channel
pasevin_ joined the channel
# 18:01 bret kbs, i haven't been keeping up. Whats this with you have been doing with the rel= pasevin, pasevin__ and bnvk joined the channel
# 18:07 KevinMarks wondering if I need a variant of NoterLive for making @pmarca style tweetposts # 18:16 kbs in that process, I discovered that rel=me linking [together with validation] involved some more nuanced understanding of how they work [in theory and practice] and there were many longwinded conversations while people patiently educated me :) # 18:18 GWG kbs: I think a discovery webapp tool thingy could have interesting possibilities. # 18:19 kbs thanks :) it's been pretty fun to play with as well tilgovi joined the channel
# 18:45 aaronpk they've been slowly rolling it out over the past couple weeks i think KevinMarks joined the channel
# 18:51 KartikPrabhu aaronpk: yup. Maybe they are moving to the stream that dustin curtis mentioned in his criticism bnvk joined the channel
# 18:58 kylewm intentionally confusing and difficult to navigate, good short term business decision, seems bad long term though paulcp_, yaf, bnvk, ttepasse, eschnou, friedcell and snarfed joined the channel
# 20:12 snarfed !tell rtaibah hey, saw your question on the WP webmention plugin. it tries to send webmentions to every link in a post when you publish or update it. # 20:12 Loqi Ok, I'll tell them that when I see them next # 20:13 GWG snarfed: I have a question for you. # 20:14 GWG Does it work for a page rather than a profile? # 20:14 GWG I have a second site I post content to that I'd like to backfeed. # 20:14 GWG G+ actually, but there might be a use case for both # 20:15 snarfed i'm not sure if it works for G+ pages. probably not, but you can definitely try! # 20:15 GWG How would I do it? The option is oauth for profiles # 20:15 snarfed when you sign up with G+, it should give you a choice of which account/page to use # 20:16 snarfed if the page doesn't show up in that list, then i guess they're not supported yet # 20:16 GWG snarfed: Any Wordpress oriented Indieweb advice? # 20:17 GWG I was commenting to aaronpk that I haven't figured out how to pull in conversations from other sites that didn't start on my site # 20:18 snarfed ah, yes. bridgy used to do that (find all links to your site, even when posted by other people), back before it supported webmentions # 20:18 snarfed if you write code at all, i happily accept pull requests! :P # 20:18 GWG I can fix bugs, but that is the best I can do. # 20:19 GWG I feel bad when I ask about things, knowing you wouldn't want my code anywhere near you # 20:19 snarfed reviewing/helping with code is always easier than doing it all myself # 20:19 GWG My training is in Library Science ttepasse joined the channel
# 20:23 GWG I forgot to clarify that to kylewm earlier # 20:24 GWG That one is my testbed for implementing these standards # 20:24 GWG snarfed: I've created a multiple online personality # 20:24 GWG Every interest, I have a website for # 20:24 GWG And the people who wanted to share a site to write with me vanished # 20:25 GWG So, suddenly I have 6 Twitter accounts. # 20:25 aaronpk well crap, a friend just commented on one of my facebook posts with her address # 20:25 aaronpk my facebook is totally public, so I assume anyone can see the comment on facebook, but now it's going copy that to my site # 20:26 snarfed yeah, bridgy is careful to only send public posts/comments, but still # 20:27 GWG This is why I have the Wordpress comment system set to Approve All # 20:27 snarfed aaronpk: alternatively, bridgy won't poll your fb for a bit more, feel free to just delete it from fb (i assume you can) # 20:28 GWG I don't want things coming through I didn't check # 20:28 GWG Unless the flood comes to be too much # 20:28 aaronpk hm not sure I want to delete from FB... since she seems to be fine with it # 20:29 snarfed ok. easiest thing to do then is make bridgy poll and send it now, then scrub. sound ok? # 20:29 aaronpk this is more motivation for me to get comment blacklist/delete/moderation working on my site # 20:29 aaronpk (the other motivator being the weird pinterest spam bots that keep replying to my tweets) # 20:31 aaronpk it's funny, all the people who replied to that post on facebook think the butter idea is weird. but everyone who replied on twitter is like yeah! and is giving suggestions for how to make it better. # 20:33 GWG But I hate coffee, so I'm not a good judge pasevin joined the channel
# 20:35 GWG I still remember when I drank cream as a kid # 20:35 GWG And how much I regretted it later KevinMarks_ joined the channel
# 20:38 snarfed you can always just write html directly into your wordpress posts krendil joined the channel
# 20:39 GWG snarfed: The comments are either coming from you as the brid.gy developer, or those people over at SNAP # 20:39 GWG I'm using their comment retrieval code # 20:40 GWG I'm assuming your pulls are marked up, and theirs are not # 20:41 snarfed oh i see. interesting. looks like both are pulling comments back, at least for twitter # 20:41 snarfed you installed pfefferle's indieweb plugin bundle? # 20:41 GWG No. I pulled each piece from the github source # 20:42 snarfed but the bundle isn't updated quite as often as the component plugins a___ joined the channel
# 20:43 GWG It is sparse on documentation outside the code # 20:43 snarfed regardless, you have it all working great, nice work # 20:43 snarfed you might want to disable either snap or bridgy for twitter so you don't get dupes # 20:46 GWG snarfed: I maintain gadgetwisdom.com bear_ joined the channel
# 20:46 GWG I run the server for valuewalk.com, though I dont' maintain the site natwelch joined the channel
# 20:47 GWG I want to add a lot of the stuff I've been playing with to those sites as well netweb joined the channel
# 20:48 GWG snarfed: Now I have to figure out how to markup a podcast # 20:49 GWG That is why I asked about the G+ page ttepasse joined the channel
# 20:49 GWG Someone runs an active one for the ite # 20:49 snarfed i always search the wiki when i hit a new task like that, usually someone's thought about it before # 20:50 GWG Now that I have the base site up, I can start adding more content and expanding outward and connecting up all my content once and for all melvster joined the channel
ben_thatmustbeme joined the channel
# 20:55 GWG I may not achieve what I want to do # 20:55 GWG I still can't even get my music metadata fixd # 20:55 aaronpk i'm considering only temporarily returning the status of webmentions # 20:55 aaronpk just storing in a cache for like 30 minutes rather than writing to some sort of permanent storage # 20:56 aaronpk so if you send me a webmention you'd get a URL you can use to check the status, but the URL will only work for like 30 minutes # 20:57 aaronpk i'm trying to think if there is some sort of benefit to keeping the logs around permanently # 20:57 snarfed aaronpk: hm. interesting that you're starting from a default of don't store, as opposed to a default of store # 21:01 snarfed seems like the code to read/write from cache isn't much easier than from disk...? # 21:01 aaronpk i don't have a database right now, so the alternative is storing on disk with the rest of the stuff # 21:01 aaronpk but then i have to make a scheme for generating files and avoiding name collisions and stuff # 21:01 snarfed temporary is definitely a good start. and you could keep it in memory indefinitely, ie until you hit memory pressure or restart # 21:01 aaronpk seems like it'd just be easier to throw it in memcache with an 8-char token # 21:03 snarfed name generation seems doable though, right? source + target, maybe truncated plus an index or timestamp suffix, or maybe hash of source + target? # 21:04 aaronpk i was thinking just sequential numbers after a date like my notes # 21:04 aaronpk but if I do memcache then i'll just randomly generate a string # 21:04 snarfed oh right. if you're handing out the id, do whatever's easiest # 21:04 aaronpk turns out if you use 10-chars of newbase60 it's equivalent to a UUID so you can generate them all day long and never get duplicates # 21:10 aaronpk i guess i'm trying to think why i would actually need that data to persist ttepasse joined the channel
# 21:19 bret any one notice chrome slowing down on older hardware in the last few months? ttepasse joined the channel
# 21:25 snarfed aaronpk: might be a good excuse to come up with some general purpose key/value store for p3k that you're ok with # 21:27 aaronpk but still, is there really a good reason to store these permanently? # 21:27 aaronpk this is for everything including garbage webmention/pingback requests as well as blatant spam attempts (source doesn't link to target) # 21:29 snarfed i'm surprised that you of all people are debating whether to store data :P # 21:29 snarfed convenience and debugging for clients is the main one, i guess # 21:29 snarfed and if you said, ok, still transient, but keep it for more like 1w or 2w, that would be good # 21:30 snarfed if you already had a storage answer that you liked, would you still debate? # 21:30 aaronpk mostly because of the amount of spam I already get # 21:30 aaronpk (things that are not even remotely valid webmentions or pingbacks) # 21:32 snarfed ie with a reasonable data model, how much space per day would the storage take? # 21:33 snarfed good data if you ever want to hack together your own wm spam filter # 21:34 aaronpk I have an app that does it, I use it to "scan" receipts and stuff by taking a photo # 21:34 aaronpk ok now I have to mount webcams on my whiteboard and run that script on it continuously # 21:35 snarfed i love me some command line imagemagick, but that’s just…wow. ttepasse joined the channel
# 22:01 GWG is eating chicken. Does not think this is noteworthy ttepasse joined the channel
# 22:14 GWG I'm worried that no one has signed up for IndieWebCampNYC tilgovi, CheckDavid, lukebrooker, ttepasse, pfenwick and lukebrooker_ joined the channel
# 22:52 aaronpk I guess lately I've been leaning towards not storing stuff unless I have to # 22:53 aaronpk example: access tokens generated by my site are self-encoded, so I don't need an access token database # 23:06 GWG No one has registered for it, I can see, KevinMarks # 23:07 Loqi KevinMarks_ meant to say: hence me pinging sara # 23:07 GWG I haven't been in Manhattan in months. I live in Queens iangreenleaf joined the channel
# 23:15 GWG I'm curious as to what an Indiewebcamp is like # 23:18 bret its fun! there is usually a meetup the day before at a bar where you can meet people and say hi # 23:19 bret the next day there are some brief intros and then people do session pitches # 23:19 aaronpk man I am not a designer at all. i can't even figure out where to put these really simple elemnts on the page # 23:19 bret then there is like 2 or 3 session rounds broken up by lunch, then the next day you come in and work on stuff # 23:22 bret ah for checking if the receiving end likes what it sees? # 23:22 aaronpk so if you send me a webmention you can see if it worked or why it failed # 23:22 aaronpk cause I do all webmention processing asynchronously # 23:29 aaronpk oh you don't see it, good :) I have php errors visible when i'm logged in irzhelni and lukebrooker joined the channel
Loqi snarfed: tantek left you a message 40 minutes ago: in particular, I think a talk on the concept of Backfeed and using Bridgy to demo it woudl be great! http://opensourcebridge.org/events/2014/proposals
@krynsky Contemplate this & take action: The Eternal Life of Your Data http://insights.wired.com/profiles/blogs/the-eternal-life-of-your-data #lifelogging #digitalpreservation #indieweb (twitter.com/_/status/452605804000968704)
ben_thatmustbeme good evening all
gRegor` Evening ben, kbs
kbs in theory - "you" are represented by the page http://gregorlove.com/about/
KevinMarks I think bare personal domains are OK as rel-me targets
tantek kbs - this is also documented here: http://microformats.org/wiki/relmeauth#users_with_separate_contact_pages
@realHermitian Amazon TV: another hardware gate for a problem with a software solution. #indieweb (twitter.com/_/status/452730788820582400)
@_Puwt Tweeps dapatka DISKON 25% utk pembuatan Web hosting & toko online di http://www.indieweb.biz dengan belanja di @zloozshop (* :D (twitter.com/_/status/452817053846470657)
@zloozshop RT @_Puwt: Tweeps dapatka DISKON 25% utk pembuatan Web hosting & toko online di http://www.indieweb.biz dengan belanja di @zloozshop (* :D (twitter.com/_/status/452823348771176449)
@kyle_wm @dshanske PS, saw your webmention to http://indiewebcamp.com in the IRC channel, come join us if you'd like! (twitter.com/_/status/452831584685543425)
@codepo8 How advertising cookies let observers follow you across the web - http://www.theverge.com/2014/4/4/5581884/how-advertising-cookies-let-observers-follow-you-across-the-web - this is what Mozilla fights against #ownyourdata (twitter.com/_/status/452843664621203456)
@kevinmarks @benwerd sÍ•Í–o̘͇Ím̶Ì̬e ̞̰̀a͕̬r̘eÍ̫̹͈͖̼̦̮ ͎͇̻v̹̪̜er͚͙ÌÌœyÍ Ì±Ì³ ÍÍ“Í“Í…Í–sÍ¡pÍe̻̜̕c̱̰̩̗i͚̫̹̖̯a̵̞̯̤l̨̜̟̜ (twitter.com/_/status/452855608358604800)
KartikPrabhu hello all!
GWG http://waterpigs.co.uk/notes/4VSLTX/ - I may have to check the webmentions plugin code to answer this
aaronpk see polling discussion http://indiewebcamp.com/webmention#Brainstorming
bret heres a good site death: http://blog.canonical.com/2014/04/02/shutting-down-ubuntu-one-file-services/