#indiewebcamp 2013-01-14

2013-01-14 UTC
hober, tantek, mxuribe, tilgovi, Nadreck, andreypopp and zztr joined the channel
#
@metacosm RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
friedcell and jaquecoustaeau joined the channel
#
@davehawes RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
tantek, eschnou, danbri and barnabywalters joined the channel
#
@zeldman RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@darylws RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@gazshaw RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@Slath RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@jonvalentine RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@okke29 RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@nicomen RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@fuddlemark RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@iainspad RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@ranwar RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@BillSeitz RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@angiemckaig RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@AnHigherLove RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@squirrelfactory RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@BrianAkpa RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@fractality RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@whovingh RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@ApocalypseCowUK RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
danbri__ joined the channel
#
@DaveCZombie RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@lordmachin RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@wadpaw RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
mxuribe joined the channel
#
@webmonkey RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@tukeke RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@getswizzle RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@toastcowboy RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@idoclosecuts RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
eschnou joined the channel
#
@digitalWestie RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@4d47 RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@ReneeDeWolf RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@44sunsets RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@concept47 RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
eschnou, tantek and ianloic joined the channel
#
@MichaelZed RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@okinasevych RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@isemann +1 Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb (via @aral)
#
@codeknitter RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
tommorris hey tantek, just put up a new Twitter account for weird text quirks. http://tommorris.org/posts/8040 ;-)
Ajis and barnabywalters joined the channel
#
@vfritzon RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
danbri___ joined the channel
#
@Stevenpearson RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
@svs_nl RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
tantek, zztr and sdboyer joined the channel
#
@aral Moved my site to S3 :)
barnabywalters joined the channel
#
tantek tommorris - heh - the tests look interesting
#
tantek the space separate hashtag was unexpected
#
tommorris tantek: ah, it's not a space-separate hashtag, it's just a different hash symbol from Unicode
#
tantek weird
#
barnabywalters good evening tantek and tommorris
#
tommorris hey barnabywalters
#
tantek good evening barnabywalters
#
barnabywalters I should dig out the twitter parsing messups I’ve found — although most of them are just inconsistencies between different twitter apps :/
#
Loqi definitely
#
barnabywalters Loqi, are you listening for "I should"s now?
#
tommorris I'm gonnna hopefully do the <ins datetime> and <del datetime> stuff shortly. been a bit busy with wikidrama.
#
tantek tommorris - twitter also fails to link irc: links
#
barnabywalters tommorris: I’ll have a crack at implementing it too. As it’s only working on bare tags, a simple regex replace should suffice
#
tommorris tantek: will post a few more tweets later with lots more obscure protocols.
#
tommorris (I wonder if the IETF have a process for deprecating old protocols? I'm pretty sure nobody is using wais:// anymore.)
#
barnabywalters or gopher://, although I think there are some people trying to bring it back
#
tommorris well, they removed gopher from Firefox a while back, which is pretty much the death-knell.
#
tantek tommorris - check out the WHATWG URL spec for the latest
#
tommorris anyway, my server does a full parser of a post before storing anyway, so it's trivial to add another callback for checking for ins and del.
#
tommorris barnabywalters: be sure to read http://stackoverflow.com/questions/1732348/regex-match-open-tags-except-xhtml-self-contained-tags/1732454 before parsing HTML/XML with regex.
#
barnabywalters tommorris: are you writing your posts in markdown? what is the parser doing?
#
tantek http://url.spec.whatwg.org/#urls
#
tommorris well, it does markdown->html, then parses the HTML for a variety of things. can't remember what.
#
tommorris has to go and eat, but will have a look later.
#
barnabywalters tommorris: ah, love that ;) I am aware of the dangers. In this case I think it’ll be safe
#
tantek regexing HTML is usually a good way to leave yourself open to various injection attacks
#
barnabywalters tantek: I cannot find any examples (nor think of any, but I’m no cracker) which would affect finding <ins> and replacing it with <ins datetime="" cite="">
eschnou joined the channel
#
@kattekrab RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
barnabywalters tommorris: I wrote the auto ins/del datetime/cite adder! currently deploying…
#
tommorris excellent. I'll sort mine out shortly.
#
tantek barnabywalters - where did you see the cite attribute on ins/del?
#
barnabywalters tantek: http://developers.whatwg.org/edits.html#attr-mod-cite
#
barnabywalters my implementation links to edit history hash fragments, like this: http://waterpigs.co.uk/articles/lutherie/#update-2013-01-11T17:25:31+00:00
#
tommorris heh, so HTML5 is keeping ins/del @cite but not @longdesc. There seems to be some inconsistency here. ;-)
#
barnabywalters tommorris: @cite is a URI, @longdesc is non-visible text IIRC
#
tommorris nope, @longdesc is a URI pointing to a long description
#
tommorris and I agree with the removal of it. just seems rather inconsistent.
#
barnabywalters ah, either I misunderstoof or it changed
#
tantek well look at that - the it's even in HTML4
#
tantek somehow I'd either missed or forgotten that
#
tantek tommorris - cite attr on blockquote is similar, but perhaps slightly more useful
#
tommorris it's actually quite usable: something like Wikipedia could use it to point to a diff.
#
barnabywalters tantek: apparently browser makers have missed or ignored both of the attributes, as neither do anything useful
#
tantek barnabywalters/tommorris - how are you actually using the cite attr on ins/del? I've found no use for it (though as I was typing it, just thought of one).
#
tommorris well, if you had version controlled posts, being able to say "this chunk of text was added/removed in this revision"
#
barnabywalters tantek: well, initial plan is to write a bit of js which allows the user to select an update. js toggles various ins/del elements depending on their datetime, to recreate the document as it was at the time of the selected update
#
barnabywalters much like SFW’s history, but native HTML with a bit of JS
#
tommorris have any of you guys seen Wikitrust? it's a Firefox extension that lets you see how much you should "trust" Wikipedia edits.
#
tommorris see http://wikitrust.collaborativetrust.com/_/rsrc/1252650901873/screenshots/Politics_of_Denmark_2.png
#
barnabywalters tommorris: oh that’s cool! trust heatmap!
#
tommorris they use web of trust type algorithms to do a sort of 'blame' on the page
#
tommorris it's based on whether the person who added it is a new or experienced user, and how often their edits have been reverted.
#
tommorris I'm hoping that having +sysop and 50,000+ edits means mine don't go orange, but I haven't tried it recently. ;-)
#
barnabywalters installed it
#
tommorris but each of those 'blame' fragments could be marked up with <ins cite="{diff URL}" datetime="{when it was changed}" />
#
barnabywalters I’m actually thinking I might implement the reverter as a browser extension which adds a timeline scrubber to any webpages marked up with ins/dels with @datetime
#
tantek barnabywalters - for the scenario you describe, viewing changes by date/time - all you need are datetime attrs on ins/del.
#
tantek you don't need cite attr for that
#
barnabywalters tantek: true
#
barnabywalters what was the use you thought of?
#
tantek tommorris's answer seems more interesting, and is roughly akin to the idea that popped in my head as soon as I typed "found no use"
#
tantek the use I thought is clustering ins/del elements into transactions
#
barnabywalters so, associating ins/del elements which are part of the same edit?
#
tantek but for that all you need is a unique URL per "transaction"
#
tantek you don't actually need a *description* of the transaction
#
tantek right
#
tantek in particular, I've often wanted to cluster one del with one ins
#
tommorris I still need to build a modern version of acts_as_git for history-browsing.
#
tantek even just that would be useful
#
tantek usually I only have one edit per day at most
#
tantek so per-day viewing of edits works fine
#
barnabywalters arguably you could do that with datetime too, as it’s highly unlikely for multiple edits to happen at exactly the same second
#
barnabywalters unless you’re wikipedia
#
tantek but the "same second" also doesn't work for paired ins / del edits
#
tommorris barnabywalters: generally, they'll edit conflict, but sometimes you'll have people editing different sections of the same page at the same time.
#
tantek I'll often delete the text many seconds before I finish inserting the new text.
#
tommorris remembers trying to edit 'Amy Winehouse' the day she died. That was not fun.
#
tantek so if you're depending on precise seconds for ins/del pairing that will fail
#
tantek also, seconds is usually too imprecise as well
#
tantek er
#
tantek too overly precise
#
barnabywalters tantek: ah, I was still thinking under the assumption of the attributes being automatically filled in
#
tommorris one's editor might be able to actually give that kind of precision
#
tommorris Vim's undo-tree for instance: someone might be able to use that.
#
tommorris but if you are doing clustering, you probably want to cluster ins/del's
#
tantek undo stacks usually keep track of edits / inserts differently
#
tantek whereas what cite would be useful for is a higher level of "as the author I consider these part of the same 'edit' "
#
tommorris well, in Vim, you might do "cw", change word.
#
tommorris but I believe we may have reached Pareto's point of stop-what-the-fuck-are-you-thinking about now.
#
tantek heh
#
tantek ok, I'll note cite attr on ins/del as potential point for future extension if I ever get around to needing to pair ins/del more than on a date-specific granularity
#
barnabywalters yeah. My site is there for experimentation anyway — there’s no harm in getting this data now, as a use might emerge as I generate more and more of it
#
tantek for now it's just a theoretical edge case
#
tantek (for me)
#
barnabywalters tommorris: wikitrust is not working for me, I just get a weird TEXT_NOT_FOUND error
#
tommorris barnabywalters: weird. I haven't tried it for a while. might want to email them or poke 'em on some social media service.
#
tommorris I think it's a university project, so it could just be they've graduated and buggered off
#
barnabywalters hm, no news and no contact links :/ Probably safe to assume it’s a dead project
#
tommorris shame.
#
barnabywalters yeah, looks like a great idea
#
barnabywalters goodnight all
#
Loqi gute nacht!
#
@mudsurfer RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
tantek tommorris, that tweet reminds me, perhaps we should find out why (directly using) Twitter is an "addiction" for Aral per: https://twitter.com/aral/status/290589559018905601
#
@aral @tommorris Hey, it’s not nice to make fun of a guy’s addiction :-P
#
tommorris I expect it's probably UI.
#
tommorris I'm dorky enough to use my own posting UI, but across the various platforms (iOS, Android, web etc.) Twitter is pretty ubiquitous.
#
tommorris tantek: I think I may have found a way to use OSM data relatively painlessly
#
tommorris tantek: have a gander at http://geocoding.cloudmade.com/BC9A493B41014CAABB98F0471D759707/geocoding/v2/find.html?object_type=poi&around=51.51384,-0.10952&distance=5000&results=300
#
tommorris now change "find.html" to "find.js" and enjoy some JSON goodness.
#
tommorris that's someone else's API key though.
#
tantek huh - and now in the archives ;)
#
tommorris well, it's already linked from StackOverflow.
#
tommorris my plan is to have it so that I use OSM to lookup, but replicate the details on my own site.
#
tommorris tantek: your post on Dodgeball's picoformat-style checkin syntax is quite interesting to that end.
#
tantek oh you mean that little aside rant in my checkie praise post? ;)
#
tommorris yup.
#
tommorris well, I'm thinking, if I were to store venues as, say, /places/>{id}, then when I go to checkin, my software could find those 'first'
#
tommorris and you could go to that URI and get a map, previous checkins and any other metadata I care to add.
#
tantek agree with storing venue URLs locally on one's own domain
#
tantek don't know about using {id} - whose id?
#
tantek realizes that Aral wasn't listed on http://indiewebcamp.com/2012/UK/Guest_List
#
aaronpk I would do it as my own ID
#
tantek considers adding him retroactively as my apprentice since I so strongly encouraged him to show up
#
tantek well he has rel=me setup on http://aralbalkan.com/ so that should work for him signing into indiewebcamp.com now
#
@bubbiting RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
tantek.com edited /2012/UK/Guest_List (+39) "retroactively adding Aral as my apprentice since I so strongly encouraged him to show up and he did (and gave an inspiring presentation and demo)" (view diff)
#
@lindertobias RT @aral: Worrying trend using Facebook, Google+, and other walled gardens as blogs. Especially geeks should know better. #indieweb
#
tantek.com edited /Getting_Started (+655) "distinguish CMS vs. static hosted domain solutions, add Amazon S3 info, note community member Aral is using it, simplify lists" (view diff)
#
tantek aaronpk, tommorris, I find opaque IDs to be an anathema and a weakpoint architecturally and in terms of data integrity/longevity
#
tantek opaque IDs also smell of "database think" - which is typically a bad mindset for designing anything that's expected to last.
#
aaronpk true, which is why I didn't use opaque IDs for my new site, but I can't think of anything better for a venue
#
tommorris well, I might use a unique alphanumeric ID assigned by the user (i.e. me)
#
tommorris which ties in nicely with the dodgeball picoformat. ;)
#
aaronpk like a slug?
#
tommorris yup
#
tommorris /places/parliament
#
aaronpk what happens when there is more than one place with the same name, like starbucks
#
tommorris yeah, that becomes a problem
#
tantek so opaque ID dumb, so just use an algorithmic ID instead
#
aaronpk based on what input?
#
tantek obvious suggested algorithmic ID: date of venue URL creation converted to newbase60 epoch days. add a slug if you want.
#
tantek that's what I'd use
#
tommorris opaque ID may be dumb, but I don't see it as being that big of a deal.
#
tantek same as what I already have as an algorithmic id for my posts
#
tantek done
#
tantek opaque = fragile
#
tommorris ultimately, the URI is the identifier. how the URI gets constructed is fairly arbitrary.
#
tantek it's only arbitrary to folks who haven't learned that some URLs survive better than others ;)
#
tantek so example
#
tantek just as I have "t4Mz1" for a *T*ext note post I posted today (the first)
#
tantek if I created a venue today I'd use "v4Mz1" (for the first venue I created today)
#
tantek (I long ago reserved "v" for "venue" in my URL shortcodes taxonomy: http://tantek.com/w/Whistle#design )
#
ianloic why do you believe that these URLs will survive better? (also, hi!)
#
tantek ianloic - hi!
#
tantek ianloic - not "believe" - based on experience. URLs based on opaque ids (and links to them) die all the time. URL shorteners. news media sites that go down with their article IDs etc.
#
tommorris will probably do it very lazily with incrementing integers, because he's lazy. He is fully aware that tantek will probably have good reason to say "see, I was right!" in a year or so.
#
tantek whereas when those URLs happen to have *some* kind of information, a date, a slug, a topic, anything - you can typically reconstruct it or even relink it to another location
#
tantek or even redirect after the fact
#
tantek but when you loose your database of opaque ids - forget it
#
tantek s/loose/lose
#
Loqi tantek meant to say: but when you lose your database of opaque ids - forget it
#
tantek e.g. you upgrade your CMS, wipe previous database/tables etc. - oops, IDs gone.
#
tantek happens all the time with media/content sites
#
tantek redesign, URLs break, opaque IDs dead
#
tantek if the IDs are generated and/or include other information, you have a chance of scripting it back together, perhaps even with automatic redirects
#
tommorris problem is, if I lose my database of opaque IDs, I also lose the content that goes with it. the solution to that is to, oh, make sure it's backed up. ;-)
#
tommorris and that will probably at some point hopefully include some kind of peer-mirroring
#
tantek tommorris - maybe. again, just looking at broader experience here of what's historically happened on the web
#
tantek designing by "probably at some point hopefully include" is not a good strategy
#
tommorris for permalink short URLs it makes sense
#
tantek better to just design more robustly up front
#
tantek assuming you never get to "probably at some point hopefully"
#
tommorris well, you can already do http://tommorris.org/posts.xml?page=400
#
tommorris the aaronsw situation makes it seem like a good idea to have some peer backups. a process for backing up our friends things.
#
tantek yes, peer backups have been discussed many times. too low on anyone's priority list to ever get implemented.
#
ianloic tantek, I don't disagree that opaque IDs are bad, but I'm not convinced that your "algorithmic IDs" are better. They look opaque to me...
#
tantek ianloic - they are until you publish the algorithm openly, and then they're forever non-opaque
#
tommorris might be as simple as writing a little bash script that you run monthly to backup your friend's sites.
#
tantek and then maintaining those scripts / configurations across server moves etc.
#
tantek nevermind the auth needed to make it work securely
#
tantek etc.
#
tantek aaronsw thought in advance about the continuity of his stuff - http://www.aaronsw.com/2002/continuity
#
tommorris should probably put up a similar page just in case
#
tantek ianloic - in the case of my algorithmic ids, I've published the algorithm several times, on my own site, in interviews, and now in this IRC channel. so discoverability is high :)
#
ianloic tantek, fair enough I guess.
#
tantek reimplementing an algorithm is way easier than recovering an opaque data table
#
tantek so yeah, basically, opaque IDs dumb (or maybe lazy).
#
tantek (if it's for ephemeral stuff, sure go right ahead, authentication tokens etc.)
#
tantek btw - ianloic - great to have you here - and hope you can make it to IndieWebCamp 2013: http://indiewebcamp.com/2013
#
tommorris tantek: on an unrelated note, here's aaronsw in 2006 at wikimania - https://www.facebook.com/photo.php?v=10151202729530773
#
ianloic tantek, at this stage I'd assume that I won't - I'm doing very non-web stuff professionally at the moment.
#
ianloic It's kind of great :)
#
tantek ianloic - it wasn't a professional invitation :)
#
tantek tommorris - Flash player required :(
#
tantek repost in open format, open content license?
#
ianloic tantek, I have a mortgage, I'd have a hard time justifying this as anything but a vacation to my wife
#
ianloic tantek, though these days she's more involved in web stuff than me. She was at CityCamp Oakland in a professional capacity.
#
tantek ianloic - perhaps you can save up for your "hobby" budget then ;)
#
tantek assumes there is life beyond the union of "professional" and "wife".
#
hober mumbles something
#
tommorris tantek: could nag someone about it.
#
ianloic tantek, potentially.
#
ianloic tantek, w/ your URL shortener - how do you find your type prefix in practice? Do you find it useful? Do you think of it more as a namespace or as hungarian-notation?