#indiewebcamp 2014-10-23

2014-10-23 UTC
#
KevinMarks_
yes, we'lll find a way
snarfed joined the channel
#
ShaneHudson
I think Loqi needs his own channel
#
aaronpk
no I just need to make webmention.io send better IRC messages like "gregor liked _____"
#
snarfed
aaronpk: hey, speaking of which…nudge on https://github.com/aaronpk/indiewebcamp-irc-logs/pull/7 ?
#
aaronpk
oh yeah
#
snarfed
yay thx
#
Loqi
yay!
jet__ joined the channel
#
aaronpk
should be live now!
fmarier and dariusdunlap joined the channel
#
snarfed
aaronpk: thanks! btw, i still occasionally see this webmention.io error: "Incorrect string value: '\\xF0\\x9F\\x98\\x9E @...' for column 'name' at row 1"
#
snarfed
want me to file an issue?
#
aaronpk
argh yeah
#
aaronpk
I thought I got mysql to accept emoji but apparently not
#
aaronpk
i'm a little behind on github issues on my stuff
#
snarfed
maybe try https://github.com/issues ? (i don't have enough incoming volume to know)
#
aaronpk
that's awful
#
aaronpk
I prefer looking at the issues in a per-project context :)
#
kylewm
snarfed: what's your reasoning for leaving the android article off your home page?
#
snarfed
kylewm: totally subjective i guess. i usually omit "how to" posts like that
#
kylewm
oh ha at first I thought you were pointing to an article explaining posts vs pages
#
kylewm
wouldn't have been surprised
jacus, dariusdunlap, tantek, mlncn-agaric, mdik, chrissaad, cmhobbs, ScruffyDan and benwerd joined the channel
#
tantek.com
edited /venue (-45) "improve dfn, -h1"
(view diff)
#
tantek
aaronpk - re: "maintenance of indie venues is interesting" - could you expand up on that perhaps in an "Issues" section on /venue ?
chrissaad, paulfitz, Pea1 and snarfed joined the channel
#
tantek
we're getting started in SF HWC
#
tantek
note: IWC Cambridge was awesome - demo notes still in IRC logs
#
tantek
and next week I'm co-chairing at first f2f mtg of W3C Social Web WG: https://www.w3.org/wiki/Socialwg
jet__, jfranusic and shing joined the channel
#
tantek
welcome shing!
#
tantek
and jfranusic !
#
tantek
benwerd: also at IWC: obra built importers for LJ export, and Twitter export that import via micropub!
#
tantek
benwerd: Known gained IndieAuth sign-in support
#
tantek
benwerd: also fell down a rabbithole trying to make Known work on PHP 5.3
#
tantek
benwerd: as a result now Known is confirmed to work on GoDaddy shared hosting, and Dreamhost shared hosting
#
tantek
… the goal was to make Known work on cheap hosting options
#
tantek
… if you're a new customer on Namecheap hosting, you get PHP 5.4, but if you're an existing customer they'll put you on PHP 5.3 and keep you there.
npdoty joined the channel
#
tantek
jfranusic: this is me joel.franusic.com
#
tantek
… I like NearlyFreeSpeech as a host
#
tantek
… it's cheap and does everything right
#
tantek
… my understanding it is run by one dude
#
tantek
… and it is architected so it can be run by one guy
#
tantek
… NearlyFreeSpeech.net
#
tantek
… it runs PHP 5.3 up to 5.6
#
kylewm.com
created /2014/Cambridge/Demos (+123) "add link to beginning of demos in IRC"
(view diff)
#
Guerillero
Digital Ocean all the way
#
tantek
what is Digital Ocean?
#
Loqi
DigitalOcean is a web hosting provider targeted towards developers and offers low cost cloud servers in data centers across the world http://indiewebcamp.com/digital_ocean
benwerd joined the channel
#
tantek
nice. Guerillero, could you add reasons / advantages to Digital Ocean to ^^^
#
tantek
thanks much!
#
tantek
jfranusic: if I were setting up websites for other people I would set them up on NearlyFreeSpeech.net
#
tantek
… easy to set up delegated access, while you can still help with config etc.
#
tantek
jfranusic: I am here to continue my path to removing non-self-hosted services
#
Guerillero
the short version is that it fits in my college student budget ($5/month) and was easy to set up
#
tantek
… I am ready to flip myself off of google hosted email services
#
tantek
… I want to use S/MIME for everything
#
obra
tantek: I actually only built twitter, not lj
#
tantek
(Guerillero that's a great set of reasons! definitely add it!)
#
tantek
… next will probably be getting Jabber setup
#
tantek
(obra, good to know. I think sandro setup the LJ export / import)
#
tantek
jfranusic: possibly raspberry pi at home too
#
tantek
benwerd: have you talked to j12t re: the scripts he has setup?
#
jfranusic
whois j12t
#
dariusdunlap
who is dariusdunlap
#
tantek
who is j12t
#
tantek
jfranusic: also I heard about tilda club
#
tantek
jfranusic: what are the indieauth / indieweb ways of doing friend graphs?
#
tantek
benwerd: people are using XFN
#
tantek
jfranusic: does Known support XFN?
#
tantek
benwerd: it will shortly
#
tantek
jfranusic: the idea is XFN + IndieAuth is a way of saying anybody who is a friend of ben's has permission to do x
#
tantek
jfranusic: how do you post your public key publicly? standard?
#
tantek
tantek: yes, h-card key property
#
tantek
dariusdunlap: dunlaps.net/darius
#
tantek
… was playing with the raspberry pi
#
tantek
… installed wordpress, and it's pretty painfully slow
#
tantek
… the beagleboneblack is faster
#
tantek
… and it's only a bit more expensive
#
tantek
… you can get some pretty powerful computers on the order of $250
#
tantek
… nice thing is they are super low power
#
tantek
… when they're not doing anything they draw 4-6 watts
#
tantek
… I'm using one of these boxes for a firewall at home and need to update it to a more powerful machine
#
tantek
kylewm: kylewm.com
#
tantek
… the thing I've been working on is server-side version of indie-config
#
tantek
… where if someone logs into your site
#
tantek
… they can click reply and have it work like a webaction to their site
#
tantek
… we got that working Saturday
#
tantek
… got it working with Ben Roberts working over this past weekend.
#
tantek
check out jessielikes.coffee
ScruffyDan joined the channel
#
tantek
shing: how does Known work on mobile
#
tantek
benwerd: it's responsive so the site just works on mobile
#
tantek
… one challenge is with image resolution and resampling
#
guerillero.net
edited /DigitalOcean (+217) "add some upsides"
(view diff)
#
tantek
what is gif?
#
Loqi
It looks like we don't have a page for "gif" yet. Would you like to create it? http://indiewebcamp.com/wiki/index.php?action=edit&title=gif
#
tantek
shing: shingwongphotography.com
#
tantek
shing: not a technical person, but I put it together with Twiter Bootstrap
#
joel.franusic.com
created /gif (+190) "Create page for GIF"
(view diff)
#
jfranusic
what is gif?
ScruffyD_ joined the channel
nagaway and Scruff___ joined the channel
#
kylewm
notes that many people don't want their social media profile to point to their personal homepage
#
kylewm.com
edited /gif (+75) "add definition"
(view diff)
#
tantek
shing, to get an icon for your site, check out indiewebcamp.com/icon
brianloveswords, nagaway, gRegor` and brianlov_ joined the channel
#
ben_thatmust
gah, the JS for this drag and drop / merge into folders is pretty heavy
j12t joined the channel
#
tantek
heavy JS is heavy
willowbl00 joined the channel
#
tantek
wonders if there is a heavy.js
#
ben_thatmust
anyone who is better at JS than I interested in working on the js with me
#
ben_thatmust
sorry, tantek, was searching, cannot find a heavy.js
#
tantek
ben_thatmust: maybe it would make a good April Fools joke
#
@bear
@ednapiranha I'm curious in this idea and I'm also seeing some parallels with some of the IndieWeb tenants
(twitter.com/_/status/525128027018395649)
#
tantek
s/tenants/tenets
#
tantek
bear ^^^
#
bear
bah - I always get that one wrong
#
@bear
@ednapiranha I'm curious in this idea and I'm also seeing some parallels with some of the IndieWeb tenets
(twitter.com/_/status/525128736094842880)
#
tantek
what are tenets?
#
Loqi
It looks like we don't have a page for "tenets" yet. Would you like to create it? http://indiewebcamp.com/wiki/index.php?action=edit&title=tenets
#
tantek.com
created /tenets (+24) "r"
(view diff)
chrissaad joined the channel
#
tantek.com
edited /principles (+380) "dfn"
(view diff)
#
tantek.com
edited /principles (+25) "scratch your own itche"
(view diff)
#
tantek.com
edited /principles (+24) "AKA tenets"
(view diff)
j12t, dariusdunlap and ScruffyDan joined the channel
#
gRegor`
ben_thatmust: I might be able to help with JS
reedstrm joined the channel
#
ben_thatmust
huh, actually i wonder if this already exists from the home screen for firefox os
#
ben_thatmust
okay off to bed
#
ben_thatmust
goodnight all
#
Loqi
sleep tight!
snarfed, j12t and eschnou joined the channel
chrissaad, musigny and lukebrooker joined the channel
#
KevinMarks
I wonder if webos cards have some code available
#
KevinMarks
Also android widgets and igoogle/opensocial gadgets have some overlap
annevk_, ShaneHudson_, snarfed, cweiske, hmans, thierrym and lukebrooker joined the channel
danfowler, tantek, Jeena_, leadballhummingb, hmans_, jacus_, davemenninger1, loic_m, glennjones, elima, pfefferle, eschnou and Erkan_Yilmaz joined the channel
friedcell joined the channel
#
GWG
!tell snarfed Now Google Plus backfeed isn't finding the links. I manually did that.
#
Loqi
Ok, I'll tell them that when I see them next
loic_m, j12t, glennjones, EOGreer_ and csarven joined the channel
Haxxa, Sebastien-L, Pierre-O, krendil, EOGreer and KevinMarks joined the channel
jsun, EOGreer, erlehmann, adactio, alanpearce, cweiske_, cweiske and kensanata joined the channel
j12t, parzzix, scor, modem and pfefferle joined the channel
eschnou, bitraten, EOGreer and cweiske_ joined the channel
adactio_, csarven, musigny, kerosene, phil80 and verdi joined the channel
cweiske_, j12t, scor, brianloveswords, reedstrm, xxcoeurxx, mlncn-agaric, KartikPrabhu_1 and KartikPrabhu joined the channel
paulfitz joined the channel
cweiske_ and indie-visitor joined the channel
#
Loqi
Welcome, indie-visitor! Set your nickname by typing /nick yourname
EOGreer, Sebastien-L and brianloveswords joined the channel
parzzix, Pierre-O, KartikPrabhu and reedstrm joined the channel
#
@dawn_evil_eu
For any requests, please, feel free to contact us via our site #game #news #rpg #startup #indieweb #indiegamedev #gamedev
(twitter.com/_/status/525277816582324224)
#
@IndieGameGuys
RT dawn_evil_eu: For any requests, please, feel free to contact us via our site #game #news #rpg #startup #indieweb #indiegamedev #gamedev
(twitter.com/_/status/525278090176778240)
#
@IndieGameDevBot
RT @dawn_evil_eu: For any requests, please, feel free to contact us via our site #game #news #rpg #startup #indieweb #indiegamedev #gamedev
(twitter.com/_/status/525278413620527105)
shiflett joined the channel
chrissaad, eschnou, snarfed and j12t joined the channel
#
dariusdunlap
Good Morning!
#
aaronpk
morning!
#
snarfed
GWG: looks like your G+ profile didn't link to your web site when you signed up for bridgy, so it doesn't know where to look for syndication links. https://www.brid.gy/googleplus/117352868113616785641
#
snarfed
your profile has the link now, so try signing up again
#
Loqi
snarfed: GWG left you a message 6 hours, 14 minutes ago: Now Google Plus backfeed isn't finding the links. I manually did that.
loic_m and musigny joined the channel
#
dariusdunlap
Interesting… IndieWebCamp.com auth failed on my domain when I input Dunlaps.net/Darius, but then succeeded when I specified https://Dunlaps.net/Darius.
#
dariusdunlap
I thought it was already working to handle https correctly. Maybe I broke something on my side.
#
GWG
snarfed, I just can't win.
#
@openwebdaily
If you’re not familiar with the IndieWeb effort, let @adactio introduce you to rel=me, IndieAuth, webmentions, POSSE https://adactio.com/journal/7698
(twitter.com/_/status/525288634036260865)
#
@simevidas
RT @openwebdaily: If you’re not familiar with the IndieWeb effort, let @adactio introduce you to rel=me, IndieAuth, webmentions, POSSE http…
(twitter.com/_/status/525288692295143424)
#
snarfed
GWG: lol. bad luck. it will go back and pick up links for older posts
#
dunlaps.net
edited /Getting_Started (+266) "/* Get a place for your content */"
(view diff)
KartikPrabhu joined the channel
chalettu joined the channel
#
GWG
snarfed, want to coauthor a document on what not to do?
#
snarfed
GWG: heh. usually I try to fix these cases do they Just Work in the future
#
snarfed
s/do/so/
#
Loqi
snarfed meant to say: GWG: heh. usually I try to fix these cases so they Just Work in the future
brianloveswords, gr0k and tfontaine1 joined the channel
Garbee, EOGreer, hober, tfontaine, ben_thatmustbeme, KartikPrabhu, gRegor`, mko, gr0k and indie-visitor joined the channel
#
gRegor`
Morning, indieweb
j12t joined the channel
#
aaronpk
morning!
#
mko
Good morning gRegor` and aaronpk
snarfed joined the channel
#
KartikPrabhu
aaronpk: do you have a post on how you collect your location info and store/upload it?
#
aaronpk
almost
#
KartikPrabhu
i have a friend who wants to do that :P
#
aaronpk
too many things to finish!
#
KartikPrabhu
aaronpk: ha! what hardware device do you use?
#
aaronpk
it's not fancy, it just runs 24/7 and uploads to my server in batches
#
aaronpk
yeah I really need to do a post on this setup
Mark87 joined the channel
#
aaronpk
I should make that app work with indieauth sign-in and publish it in the app store
#
Loqi
yea!
#
KartikPrabhu
why not micropub it for good measure :P
#
KartikPrabhu
gives aaronpk more things to do
#
aaronpk
I don't have a good answer for how to upload the location data via micropub yet
#
KartikPrabhu
aah i see
#
KartikPrabhu
h-geo type things not sufficient?
#
aaronpk
i could change the property names to h-geo stuff relatively easily, but the problem is more with sending multiple location objects in a single request
#
aaronpk
form encoding does not lend itself well to that
#
dunlaps.net
created /Small_Computers (+1163) "Created page with "{{stub}} These '''<dfn>Small Computers</dfn>''' are low-power computers that are useful as servers and special-use controllers for a variety of applications. They can all run we...""
(view diff)
#
aaronpk
also the "motion" field there will sometimes have multiple values
#
KartikPrabhu
is there a formal spec for form-encoded?
#
aaronpk
but the way everyone implemented "arrays" is not in the spec
barnabywalters joined the channel
#
aaronpk
val[]=1&val[]=2
tantek joined the channel
#
dunlaps.net
edited /Small_Computers (+49) "/* Hardware */"
(view diff)
#
aaronpk
I mean there's no reason I couldn't just send a JSON document to the micropub endpoint
#
KartikPrabhu
so the spec does not define a canonical way to send multiple values so people made up their own notation
#
KartikPrabhu
true. just use h=geo file=<file data>
#
aaronpk
interesting point, I could upload it as a file
#
KartikPrabhu
and your micropub endpoint should handle that
#
aaronpk
so the micropub request would still be form encoded
#
KartikPrabhu
second coffee time in the great Hobbit tradition
#
dunlaps.net
edited /Small_Computers (+35) "/* Hardware */"
(view diff)
#
reedstrm
KartikPrabhu: second? I think I'm on 4th.
#
tantek
coffee++
#
Loqi
coffee has 23 karma
#
dunlaps.net
edited /Small_Computers (+42) "/* Hardware */"
(view diff)
#
dariusdunlap
I’ve got to get on to some other stuff today, but at least that ’s started.
j12t joined the channel
#
aaronpk
so encoding this data as h-geo uses 44 bytes more per data point
#
aaronpk
due to the additional array brackets needed around each property
#
KartikPrabhu
arraybrackets--
#
Loqi
arraybrackets has -1 karma
#
KartikPrabhu
serisouly
#
aaronpk
notice how each property like "latitude" is an array here http://microformats.org/wiki/h-geo
#
KartikPrabhu
hmm I see you want to send a mf2 parsed JSON type JSON
#
tantek
because each property is both optional and potentially multivalued
#
aaronpk
that would make sense
#
aaronpk
tantek: in the case of my location data, it is not possible to have more than one latitude
squeakytoy joined the channel
#
tantek
if you're worried about 4 bytes per lat/long pair, save even more bytes by using OctalGeo.
#
aaronpk
44 lol
#
tantek
I mean either you're worried about a few bytes, or you're not.
#
dunlaps.net
edited /Small_Computers (+139) "/* Hardware */"
(view diff)
#
tantek
[ ] [ ] = 4 bytes right?
#
tantek
for a single lat/long pair?
#
aaronpk
i have 8 properties i'm sending
#
tantek
due to the arrays?
#
aaronpk
timestamp, latitude, longitude, altitude, speed, horizontal accuracy, vertical accuracy, motion
#
tantek
no altitude accuracy? ;)
#
aaronpk
oh and then add 2 more characters for each property because of the quotes, whereas I am currently sending numbers
#
aaronpk
(I don't get altitude accuracy from the phone)
#
tantek
one of the other points of multivalued properties (which we haven not explored much but went into the design) is that per object semantics could indicate cross-property ties
Pierre-O joined the channel
#
tantek
e.g. the 2nd latitude value is correllated with the 2nd longitude value etc.
#
aaronpk
oh yeah
#
tantek
assumes that your publishing / consuming is strict enough that you always have the same number of each of the properties
#
aaronpk
that seems dangerous to me tho if you're talking about hundreds of values
#
tantek
but it is a valid use
#
aaronpk
too easy to accidentally mismatch
#
tantek
you can check counts for validation right?
#
aaronpk
i suppose
#
tantek
and once you've done that, then there's no more mismatch risk either way
#
tantek
it's just two ways of serializing a 2d array
#
tantek
row-major vs. column-major
musigny and eschnou joined the channel
#
aaronpk
so this is one way one of my data points could be serialized as h-geo https://gist.github.com/aaronpk/cc82f5957dd66ba02f87
#
tantek
or we could come up with a new object since you have so much more data than h-geo
#
tantek
motion has two values?
#
tantek
when the other have one?
#
aaronpk
yes, motion has potentially multiple values, the others will have only one
#
tantek
huh - might want to pre-encode that multivalue then
#
tantek
since you want them all associated
#
tantek
e.g. "stationary, driving" rather than "stationary", "driving"
#
tantek
it's not that motion has multiple values, it's that a motion value has multiple flags that could be set.
#
tantek
and those flags have names
#
aaronpk
I'd rather not have to parse out a comma separated list from what could have already been parsed as an array
#
aaronpk
so for that reason it makes sense to have multiple objects in the "items" array
#
tantek
then I'd pursue compressing motion even more
#
tantek
what's the set of possible values for it?
#
tantek
also, "parse out a comma separated list " = one split() call amirite?
#
aaronpk
stationary, walking, running, automotive, and I think they are adding biking
#
aaronpk
depends on whether there are spaces after the commas
#
tantek
right
#
aaronpk
putting a comma-separated list inside a JSON-encoded string sounds ridiculous
#
KartikPrabhu
iirc python split() handles the spaces correctly
#
aaronpk
sounds like sending a form-encoded JSON document which is a pet peeve of mine
#
KartikPrabhu
aaronpk: also vertical_accuracy should be vertical-accuracy ?
EOGreer joined the channel
#
aaronpk
oh yeah guess hyphens are better word separators in mf2 land
#
tantek
KartikPrabhu: there's no such thing as "correctly" there - if something is doing magic whitespace work it could be lossy
#
tantek
e.g. split then join should maintain the data
#
reedstrm
KartikPrabhu if by correctly you mean it will split on what you tell it to split on, that is ',' There's no magic for whitespace
#
tantek
aaonpk, given that data and the practice you speak of ("think they are adding") I'd use a simple single character based string array since they are flags from a discrete set
#
KartikPrabhu
tantek: it depends on the split parameter.
#
tantek
KartikPrabhu: it shouldn't. split is fairly literal
#
aaronpk
oh yep here's cycling
#
KartikPrabhu
eh? I can say split("\t") or something
#
reedstrm
You can choose to split on ', ' (that's comma space)
#
aaronpk
oh there's a motion confidence too... I should probably be including that
#
tantek
reedstrm but then you have to assume spaces in the data, also not robust
#
tantek
aaronpk single character based string array = something like "swracu"
#
tantek
where if the value is absent, you put a space, not emptiness
#
aaronpk
in iOS, the "CMMotionActivity" object has several boolean properties, as well as "low, medium, high" for confidence value
#
tantek
that's a classic method of encoding named flags like that
#
tantek
even if/when some combinations of flags are unexpected/invalid
#
aaronpk
that seems relatively close to the iOS implementation too
#
tantek
e.g. for "stationary", "driving", you would have "s a "
#
reedstrm
KartikPrabhu - right. You can split on anything you specify. The only magic is that the default is 'split on whitespace, remove empty results', so it's not as reversible.
#
KartikPrabhu
h,mmm yes but it is useful
#
tantek
the point here is to be able to easily index into the string, pull out the value, test it whether it > " "
#
aaronpk
kind of like a bitmask but with meaningful characters
#
tantek
correct
#
aaronpk
bytemask
#
tantek
charmask
#
tantek
and because the start letters are unique (unlikely to be an accident), you can remember what they mean by glancing at the charmask
#
tantek
so it is still semi-human-readable
#
reedstrm
Sorry, not groking all the scrollback - what deserves this level of optimization? Ar there billions of these records?
#
aaronpk
reedstrm: i'm sending usually at least 200 points at a time per POST request
#
tantek
reedstrm - somewhere between millions and billions ;)
#
KartikPrabhu
reedstrm: aaronpk is known for taking his location data every 5 seconds soooo yeah
#
aaronpk
and I currently have something like 20 million or so, haven't counted them all up in a while
#
tantek
I was right! :)
#
aaronpk
it was 10 million a couple years ago
#
aaronpk
I cranked it up to 1 second recently too, but only when i'm moving
#
aaronpk
(recently being like a year ago)
#
reedstrm
Yeh, but that's storage. You're not storing all that flat, are you?
#
aaronpk
you know what I should just be sending a geojson file via micropub
#
reedstrm
o.k this is where I bow out of the discussion, because to my eye, this is exactly the kind of data that a DB is good at, at the scales it's designed for.
#
aaronpk
reedstrm: read the top of that article ;)
#
reedstrm
Well if you insist, I'll give you my honest take on it.
#
reedstrm
(already read it btw)
#
reedstrm
The problems you describe is not using a database. It's which database.
#
reedstrm
The gold standard for GIS work these days (which thsi is) is PostGIS, PostgreSQL w/ Geo extensions.
#
aaronpk
I know Postgres doesn't have the locking problem that mysql has
#
aaronpk
but how about upgrading the database?
#
reedstrm
All that tablelocking/backups/non-version compatible stuff just goes away.
#
reedstrm
The postresql team is _vewry_ conservative about that. THe on-disk format seldom changes. If it does, there are tools for for doing an in-place upgrade.
#
aaronpk
that's good
#
danlyke
aaronpk for the most part I've just done package manager updates for a decade and a half now.
#
danlyke
The one time that didn't work, I fired up the old version, dumped it, fired up the new version and imported it. And I think it didn't work because I'd waited too many versions between distribution upgrades on that server.
#
reedstrm
I'm running production systems that have ~ 60 GB in them on disk. Tey'be been through every major postgresql upgrade since version 7.<something>
#
finchd
PSU has the ODOT database at portal.its.pdx.edu it is ~4TB on a single server, been running well for years
#
aaronpk
i should add a note to my follow up article about how writing this data to separate plain text files allows me to sync it between machines using btsync
#
Loqi
I agree
#
reedstrm
btsync?
#
aaronpk
what is btsync?
#
Loqi
Bittorrent Sync is a proprietary Dropbox replacement/alternative http://indiewebcamp.com/btsync
#
aaronpk
what would you recommend as an easy way to get incremental updates replicated to other postgres servers?
#
reedstrm
danlyke: and finchd: those are both postgresql testimonials, I'm assuming?
#
danlyke
reedstrm yes.
#
reedstrm
aaronpk: DB replication has been the big dev push the last 2 years or so :-)
#
aaronpk
i'm not talking about actual replication tho
#
aaronpk
because I want my laptop to be able to have a copy of the data and I don't think actual postgres replication would work super well for that
#
rascul
just realized my btsync of the wiki wasn't running
#
rascul
restarts
davidpeach joined the channel
#
reedstrm
aaronpk: you might be surprised. The main issue w/ an occasional disconnected replication like that is that you would need to set up the primary server
#
reedstrm
to keep it's log files essentially indefinitely.
#
aaronpk
ah right
#
reedstrm
What's the acutal size of a mysql dump, BTW?
#
aaronpk
haven't checked
indie-visitor joined the channel
#
reedstrm
ball park - tens of GB, or we getting into TB levels?
#
aaronpk
probably just tens of gb
#
reedstrm
Anyway, the main reason I'd suggest it, is the community around PostGIS - all the cool tools are being built against it. OSM uses it heavily
#
aaronpk
we use it a lot at esripdx too :)
#
tommorris.org
edited /advocacy (+345) "/* Good and bad practices */ expanding"
(view diff)
#
tommorris
tantek: ^^ based on the off-channel discussions. :)
#
reedstrm
I think a postgresql install using 'log shipping' replication would probably fit the bill pretty well. Then you can worry about the format for message passing around
gr0k joined the channel
#
reedstrm
tommorris: but the current discussion is on-topic and requested, correct?
#
reedstrm
aaronpk: w/o worrying about the storage format, per se.
#
tommorris
reedstrm: I haven't read it yet. :)
#
reedstrm
decoupling's a good thing.
#
tantek
tommorris++
#
Loqi
tommorris has 59 karma
#
aaronpk
I dunno, I'm still leery of maintaining a postgres database for this
#
aaronpk
but I agree the transport should not be affected/driven by the storage
#
tommorris
PostGIS is made of all sorts of awesome. That's my only contribution to the discussion, I'm afraid.
#
reedstrm
well, that's your call. My experience is that my postgresql dbs take essentially no maintenance, other than making sure I don't acutally run out of disk space. Even then, it's stop, add more space, start, and it auto recovers. Now, the mysql db backing my home DVR, on the other hand ... I won't start :-)
#
tantek
depending on any one source-base / implementation is a bad long term design decision. your dependencies should be on open formats and protocols, not any one open source project.
#
tantek
until someone else builds an implementation that can interoperate with PostGIS files, protocols, replication etc., PostGIS is just another monoculture vulnerability.
#
aaronpk
tantek: thanks, that is a great expression of what I was thinking
#
tantek
this is why file systems will outlast *any* databases.
#
aaronpk
I mean I do trust the Postgres community far more than the MySQL or CouchDB communities, but still
#
tantek
no database software interoperates at such standards. maybe CSV dumps/exports/imports?!?
#
danlyke
So there's SpatiaLite, which could be your second implementation.
#
tantek
aaronpk, but still, it is a *monoculture* community
#
tantek
and thus vulnerable to all the monoculture problems
#
aaronpk
oh no I don't wanna go anywhere near sqlite
#
tantek
this is what is very difficult to explain to advocates of any one open source solution
#
danlyke
(Still worth thinking about second formats and making sure your dumps are compatible)
#
tantek
danlyke - not just "worth thinking about" IMO, but *required*.
#
aaronpk
this flat file thing feels pretty solid to me
#
tantek
and this is also why it's safer to longterm keep your data in the thing that is interoperable (e.g. the dumps) rather than the thing that is *only supported by one codebase* (actual database files)
#
reedstrm
tantek - how close are you to the GIS comunity to paint it with your 'monoculture' brush?
#
tantek
reedstrm - does not matter. anyone dependent on a single open *source* project is a monoculture.
#
tantek
independent of what their specialty may be
#
danlyke
tantek, yeah, my experience with tab delimited SQL dumps (the last few I've looked at) is that they're easy enough to parse in Perl that that's as good an intermediate format as any.
#
reedstrm
tantek: so it's your particular dogma, I see.
#
tantek
reedstrm - not dogman, backed up by experience, long term evidence. see /monoculture
#
aaronpk
you do realize the exact thing tantek is worried about has happened with MySQL
#
reedstrm
Open standards and foramts, with an open source implementation.
#
tantek
all software dies. and it shouldn't matter if it does. it's your data that matters.
#
tantek
reedstrm *an* implementation. that's the problem.
#
gregorlove.com
created /500px (+121) "stub"
(view diff)
#
tantek
anyone involved in standards knows that that means you can't trust the supposed "open standards and formats" when there is only one implementation.
#
danlyke
yeah, I think one of the reasons the PostgreSQL users tend to be less worried about it is that PostgreSQL has a culture of "we're implementing the SQL spec" which means interop with Oracle/SQLServer/etc (with GIS extensions) has long been possible.
#
tantek
danlyke - that sounds like they're pursuing a path of being compatible with at least one other implementation, and that *greatly* mitigates the monoculture risks
#
tantek
however, one open source, and one closed source, is still very fragile / tenuous, historically speaking
#
reedstrm
But as long there are robust export tools to open formats, the existing open source implementation doesn't matter. If it stops growing, gets taken private, whatever, your code doesn't stop working. We're not talking a webservice here.
#
tantek
reedstrm is there an open test suite for these open standards and formats you speak of?
eschnou joined the channel
#
tantek
reedstrm, it's a timing problem. the one implementation might *die* before you have time to export all the things.
#
tantek
e.g. if you've backed things up in implementation-specific formats, all those backups are null and void when the single implementation dies.
#
gregorlove.com
edited /advocacy (+1) "/* Good and bad practices */ Oxfordize"
(view diff)
#
tantek
and if your'e backing up in "standard" export formats, but not actually re-using them, then you're taking *on faith* that the standard export formats actually maintain full fidelity of your data.
#
tantek
thus once the one implementation dies, you might not lose all your data, but you'll likely suffer some data rot / loss.
#
reedstrm
tantek: again, this isn't a webservice. Running my own code (since it's open source it's mine) Im' not sure how it "dies"
#
tantek
it dies when your device / OS / dev tools upgrade and it no longer runs
#
reedstrm
A backup isn't a backup until it's been tested by a restore. That's an article of sysadmin faith.
#
tantek
and the compiler for it no longer runs or doesn't compile it
#
tantek
that we agree with
#
tantek
s/with/on
#
Loqi
tantek meant to say: that we agree on
#
reedstrm
for that to happen, you would have had to have your head in the sand for a very long time, though.
#
aaronpk
in my experience that "very long time" only needs to be like 2 years
#
aaronpk
and 2 years is not that long to go without touching a code base once a system is set up
#
tantek
reedstrm - it happens all the time. go visit the computer history museum, or talk with @textfiles.
#
reedstrm
But it's plenty of time to see the writing on the wall and get your export story working.
#
tantek
and such deaths often happen slowly, quietly.
#
tantek
without any writing on the wall
#
tantek
the writing just slowly stops being written.
#
tantek
and life has a way of pre-occupying all your time for periods of time.
#
reedstrm
tantek - I'm getting lots of emotionally laden vague scare speak from you. Can you give me a concrete example of what you're talking about?
#
reedstrm
tantek well that's what priorities are for. Nothing says plain text files on a file system will last forever, either.
#
tantek
reedstrm - see above - go to the computer history museum. tons of examples there.
#
tantek
reedstrm - it's a matter of picking less fragile solutions over more fragile ones.
#
tantek
per the heat death of the univers, nothing is "forever". some things just have *more* longevity.
#
tantek
s/univers/universe
#
Loqi
tantek meant to say: per the heat death of the universe, nothing is "forever". some things just have *more* longevity.
#
reedstrm
looks meaningfully at the floppies full of thesis data that he has no equipment to read
#
reedstrm
Right, it's a balance - you've drawn a bright line in the sand, claiming that flat files are the one true way. I contend that there's essentially no difference between the software needed to read those files and the dreaded "monoculture" of a system like postgresql.
#
tantek
btw any reasoning in the form of, or any system that depends on "plenty of time to notice abc and do xyz" is form of future time-debt and not zero-cost. accrue enough of those and you reach failure. or death is failure, or passes on the future time-debt to someone else who may already have too much future time-debt of their own.
#
tantek
thus avoid depending on (or designing systems that depend on) any form of "plenty of time to notice abc and do xyz"
#
reedstrm
which your preferred solution suffers from as much as mine. Marks on acid-free paper are about the only thing that might pass that test.
#
tantek
reedstrm - historically there's been quite a big difference in longevity between storing things in open formats with multiple implementations, and storing things in a format (open or not) with a single implementation.
#
tantek
1980s desktop software all bears this out
#
tantek
like all of it
#
tantek
"as much as" - nope. a single implementation solution means you have to pay attention to that implementation (time debt). text files, or any multi-implementation format, you can ignore the implementations (no additional time debt).
#
reedstrm
Really? I used software that wrote very standards compliant and readable postscript files, back then. I can still read them.
#
tantek
and there were multiple implementations - back then. that's my point.
#
tantek
whereas good luck reading *any* document created by (and supported by) a single implementation in the 1980s.
#
reedstrm
(at leat, the ones I migrateed rom unreliable media)
#
tantek
physical media vulnerability is another problem (and source of future time debt)
KartikPrabhu and julian`` joined the channel
#
aaronpk
is there a layer diagram like the one for networks but for data storage?
#
reedstrm
There is not a zero _current_ cost to your position, however. Innovation by it's very nature involves doing things that no one else is doing. Hence, single-implementation.
#
tantek
aaronpk - yes. It was in a Longnow presentation a while ago.
#
tantek
reedstrm - of course for there to be implementations, there had to have been a point in time where there was only one.
#
tantek
you can reduce that time period by implementing based on openly created open formats and standards.
#
tantek
and encouraging multiple implementations from day one
#
reedstrm
I think we are actually both on the same side of the bell curve on this argument, actually. I just think you're throwing a lot of babies out with the bathwater curse of 'monoculture'
#
tantek
rather than having the arrogance to pursue a path where one (open source or not) implementation is sufficient / enough / rules them all.
#
aaronpk
guys this is fascinating but I wanna get back to my original question/problem
benwerd joined the channel
#
reedstrm
See, making open source parenthetical there goads me. The existence of the code means that I am never locked in.
#
reedstrm
aaronpk - sure, I think I can turn it back down to a low simmer :-)
#
tantek
that's the classic error of forgetting that free software is only free if your time is worthless
#
aaronpk
only if you are actually comfortable developing that code!
#
aaronpk
i use plenty of open source software I don't know how to change and have no desire to
#
tantek
I'd say that's true of *most* of the open source software I use.
#
aaronpk
even the ones written in langauges i am familiar with!
#
reedstrm
tantak you're making the classic error of confounding the two meanings of free. Freedom is never free.
#
tantek
two things. 1) I've enough such "innovations" die that I have no desire to invest any time in them. too risky to be worth my time. perhaps that *is* a personal age/experience thing. 2) every time I've seen a "monoculture" style open source project/community dominate an area it evitably gets displaced by something else and dies. Every time.
#
aaronpk
example: firefox. say mozilla blows up and stops updating firefox. are you *really* going to start digging into the code to make updates to it when you want it to support new browser standards?
#
reedstrm
tantek: I have a feeling we're actually of the same generation. You've just perhaps been bitten harder than I have.
#
tantek
reedstrm - not a mistake. a deliberate calling out that not "locked in" is insufficient. time cost is real.
#
sparverius
i could easily see a large project like firefox getting an outpouring of support
#
sparverius
one like libreoffice, or a popular window manager?
#
sparverius
maybe not
#
tantek
sparverius - these monocultures often follow a path of feature bloat (due to popularity / dominance / enterprise), and then slowdown, and then displacement, neglect, slow quiet whimpering decline, and then death.
#
aaronpk
ok anyway. here's what i'm thinking. my ios app will send up a geojson file in a POST request to the micropub endpoint
#
reedstrm
I never said it would be free of cost. Everything cost something. I think you're paying daily costs that are avoidable, and compound to more, in my estimation, than the future cost of migrating the data. Which will need to be migrated anyway, multiple times (see: lifetime of media)
#
aaronpk
geojson because that seems like a reasonably adopted standard at this point
#
tantek
aaronpk - is still worrying about 44 bytes per second? just checking here.
#
aaronpk
no, geojson is gonna be chattier than my current protocol
#
tantek
(and btw, the on-the-wire formats don't have to be as robust as what you store)
#
tantek
since on-the-wire is ephemeral
#
tantek
you can stick anything into an "e-aaronpk-geojson" property btw
#
aaronpk
i don't see a reason to use anything mf2-related for either on-the-wire or storage formats for this data
#
reedstrm
aaronpk: if you want to eventually migrate this down to internet-of-things sort of level, were devices may have much less bandwidth to report their location (or storage to accumulate it), worrying about the bytes might make sense.
#
tantek
reedstrm by that remark I sense we may have some similarities in background.
#
aaronpk
mostly since geojson is pretty well established
#
tantek
aaronpk: I thought geojson was basically just geomjson with geo assumptions
#
sparverius
which one of you is imaginator ;)
#
aaronpk
geomjson?
#
tantek
aaronpk did you forget that joke? I made it in-person at perhaps GeoLoqi's first office
#
reedstrm
tantek: does it horrify you to know that current EE students graduate without learning how to solder? At all?
#
tantek
reedstrm: wat. that. I don't even know how to respond. I mean, maybe some colleges but surely not at reputable ones?
#
aaronpk
here's what one update looks like as geojson https://gist.github.com/aaronpk/28bb536b7ca32a381726
#
reedstrm
Well, this one seems fairly well respected .... (I think there are enough clues scattered about to indicate which one I'm talking about)
#
tantek
aaronpk - I observed geojson had no *geographical* semantics at all, and only had *geometric* semantics. hence Geomjson.
#
tantek
reedstrm - hopefully breadboards at least?
#
tantek
on another topic, huge step forward for anyone who works on or uses or develops or publishes activity streams
#
tantek
finally a W3C draft has been published:
EOGreer joined the channel
Pea1 joined the channel
#
tantek
whether you use ActivityStreams or not, it's likely that if you're interested in indieweb-like things, it's worth your time to read the specs, for ideas, background, research if nothing else
#
reedstrm
Yeah, they do a little of that. Then it's FPGAs, VLSI, autorouters, fab shops, and pick-and-place outsourcing. just like real Engineers :-)
#
tantek
disclosure: I'm a co-chair of the W3C working group that produced those, and I myself *do not* *yet* support ActivityStreams 2.0 on my own site (no dogfood from me yet)
#
reedstrm
aaronpk: 248 characters hmm.
#
benwerd
I'm going to update Known to support this, ideally before the TPAC (although I won't be attending)
#
ben_thatmustbeme
pokes head in
gr0k joined the channel
#
reedstrm
yo ben_thatmustbeme
EOGreer joined the channel
#
ben_thatmustbeme
quite active this morning
#
aaronpk
reedstrm: coincidentally that is also what I am storing on disk in files right now. one file per day.
#
aaronpk
my files are 1-4mb per day
#
reedstrm
tantek very cool!
#
tantek
benwerd - you're publishing all the formats right?
#
benwerd
tantek - not quite all. h-feed, RSS, JSON that incorporates an older Activity Streams standard, and DogeON ;)
eschnou joined the channel
#
reedstrm
aaronpk: compressed?
#
aaronpk
not compressed
#
aaronpk
a 2.4mb file compresses down to 187k
mlncn-agaric joined the channel
#
aaronpk
(gzip)
#
tantek
also heads-up, I can't find Activity Streams 2.0 support on the editor's site either: http://www.chmod777self.com/ (apparent lack of selfdogfood)
#
reedstrm
'cause that sort of repeated data compresses embarrassingly well ... yeah like that :-)
#
aaronpk
yeah heh
#
tantek
I will ask in the #social channel in W3C IRC
#
reedstrm
roh roh, are there _any_ test implementations? 'cause standards w/o code often have ... interesting corner cases.
#
reedstrm
attempts scooby doo voice
#
aaronpk
the only thing i'm not sure about with my current format is the motion data
gr0k joined the channel
#
aaronpk
I am considering switching it to something like tantek suggested, a charmap. also because I should proabbly be storing the confidence of the motion as well.
#
tantek
reedstrm: no implementations yet AFAIK. exactly the questions I've been asking. implementations have been promised by *next week* for W3C's TPAC meeting in Santa Clara, and the Social Web WG f2f meeting: https://www.w3.org/wiki/Socialwg#Face_to_face_Meetings
#
reedstrm
well, aren't they mutually exclusive?
#
aaronpk
so that might look like {"motion": {"type":"s a ", "confidence":1}}
#
aaronpk
no they aren't
#
reedstrm
Sound like an enum to me
#
aaronpk
which is why apple implemented it that way
#
tantek
likes compressing data formats.
#
aaronpk
"stationary, automotive" is a common one
#
aaronpk
when the phone is resting on the dashboard in a car for example
#
aaronpk
walking, running, cycling are mutually exclusive
#
aaronpk
(as observed, not by any docs)
#
tantek
no swimming?
#
aaronpk
maybe in iOS 9
#
tantek
got to tri harder ;)
#
aaronpk
cycling was added in 8
#
tantek
likely waiting for waterproof iPhone 7
#
reedstrm
Ah, so stationary has nothing to do w/ speed, got it.
#
aaronpk
actually none of them do, since this is coming from the M7 chip, not the GPS
Pierre-O and chrissaad joined the channel
#
aaronpk
M8 in the new iphone i guess
#
@PageRankSEO
Indie web building blocks /by @adactio https://adactio.com/journal/7698 #indieweb #relme #hentry #webmention #micropub #indieauth …
(twitter.com/_/status/525342433538998273)
KevinMarks_ joined the channel
#
@cacheflowe
@JackSmithIV @waxpancake backup tools will likely be a paid feature. Though right now, just add .json to your url and you get your post data
(twitter.com/_/status/525328915653722113)
Erkan_Yilmaz joined the channel
#
aaronpk
i do that with my urls too
#
tantek
KevinMarks: what in particular is the potential antipattern you speak of?
#
KevinMarks_
"backup tools will likely be a paid feature."
#
tantek
KevinMarks: tools or service? time is a paid feature.
#
tantek
ah, export
#
ben_thatmustbeme
git clone gaia... this might take some time
#
@jonty
@cacheflowe @JackSmithIV @waxpancake That makes your platform inhospitable to anyone who cares about their data. Export is not optional.
(twitter.com/_/status/525331849066074112)
#
@waxpancake
@cacheflowe @JackSmithIV Yeahhh, I hope you’ll reconsider that. Charging to get your own posts out is pretty user-hostile.
(twitter.com/_/status/525333740550373376)
#
@r0unak
@cacheflowe you claim to be all about the user, but holding their data hostile until they pay up is the worst thing you could do.
(twitter.com/_/status/525334368248926209)
#
@tayhatmaker
@cacheflowe @JackSmithIV @waxpancake it'd be wise to reconsider imo. Other premium features would make more sense. Holding data hostage = 💀
(twitter.com/_/status/525334796646375424)
#
tantek
reedstrm, speaking of exports and importance thereof (in general, from whatever), I though you'd find this funny (apparently from someone working on Ello) https://twitter.com/cacheflowe/status/525337481097805824
#
@cacheflowe
@waxpancake @jacksmithiv Some people will never be happy with a free service. We honestly haven't gotten into the export details yet. #shrug
(twitter.com/_/status/525337481097805824)
#
tantek
"haven't gotten into the export details yet" = 🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨🚨
#
benwerd
Meanwhile I was just invited to Tsu.co, which is a social network that splits advertising revenue with you. Lots of people dancing around what needs to be done, nobody quite going all-in.
#
benwerd
(Yet ;)
#
tantek
awesome, and KevinMarks debunks their "isn't getting your post as JSON data enough?" https://twitter.com/kevinmarks/status/525342085189095425
#
KartikPrabhu
doh! ello is getting hammered in that thread
#
ben_thatmustbeme
what is tsu.co
#
Loqi
It looks like we don't have a page for "tsu.co" yet. Would you like to create it? http://indiewebcamp.com/wiki/index.php?action=edit&title=tsu.co
#
reedstrm
benwerd I just heard about a techblog that does that ...
jet___ joined the channel
#
reedstrm
(splits the add rev)
#
reedstrm
s/add/ad/
#
Loqi
reedstrm meant to say: (splits the ad rev)
#
benwerd
a *blog* that splits the ad rev?
#
tantek
KartikPrabhu: yeah that thread makes it pretty clear how clueless Ello is and thus fairly ignorable (unless you see some actual UX innovations that are worth documenting)
#
benwerd
Why yes, I would love my two cents back
#
tantek
in which case, better save some screenshots, this silo will run out of runway in 2-3 years tops.
#
KartikPrabhu
tantek: ello jumped on the FB privacy/ad disaster
#
KartikPrabhu
maybe we need another disaster to remind people about why owning your data is important
#
reedstrm
benwerd: It's a long standing techblog from the early days of 'the new journalism' so, yeah, actual revenue. Seems to have a rep for getting scoops. Now if I could just remember the name ...
#
KartikPrabhu
tantek: not bothering with an ello account at all
#
KartikPrabhu
honestly can't stand the horrible font
#
tantek
also not bothering.
#
aaronpk
until there is some new UI thing worth checking out
#
@inorganik
@cacheflowe don't listen to the haters. If JSON data doesn't satisfy them nothing will. They'll always find a way to find something wrong
(twitter.com/_/status/525336396199112704)
#
@milkandtang
@waxpancake @cacheflowe We want a paid platform but are bothered when they charge for stuff? I'd call backup a power-user feature.
(twitter.com/_/status/525336589908467712)
#
tantek
data longevity? definitely a power user feature. ;)
#
reedstrm
Those who care about such things are clearly the minority.
#
@kevinmarks
#indieweb @t: if we want a vouch url we return http error 449 which msft made up which means "add a parameter"
(twitter.com/_/status/525347960079810561)
#
reedstrm
Heck, I'd estimate that 40% of those who are care in this chat room!
#
danlyke
reedstrm I've observed before that based on consumer behavior, privacy and control of personal information have negative value to most purchasers.
#
tantek
I see KevinMarks is testing an update of Noterlive ;)
#
reedstrm
yeah. Until they get hacked/lose their baby pics/ etc.
#
gRegor`
I wonder if the @ doesn't work as a notification on Ello unless you follow each other
#
@kartik_prabhu
if you take my content for free, maybe you should also give it back for free? just a thought #silos #indieweb
(twitter.com/_/status/525348509407383552)
#
@patr1ck
@cacheflowe Yes. Taking my data for free but not letting me get it back without paying for it is lame.
(twitter.com/_/status/525345944519577600)
eschnou joined the channel
#
tantek
nice of Ello to raise awareness about the need to #ownyourdata
#
gRegor`
No permalinks in the JSON, but why would I want to use their ugly URLs anyway? :) I'd rather just use the ID
#
@r0unak
@cacheflowe you believe that you own the user data, and they don't. And that is hostile.
(twitter.com/_/status/525347856027496448)
#
gRegor`
It's encouraging to see all that response rather than just a shrug
#
aaronpk
oh wow now they're debating with people on this?
#
tantek
aaronpk - yeah, they're lolworthy arguing it
#
@cacheflowe
@inorganik Thanks :) I want to respond when I can, but yeah, everybody's gotta have *something* to complain about.
(twitter.com/_/status/525336614638460928)
#
gRegor`
Life's so rough, getting $5.5M
#
tantek
I wonder if raising that kind of $ contributes to such an irrationally defensive attitude?
#
@cacheflowe
@waxpancake @JackSmithIV So charging a buck or two for an export feature is hostile? This is ridiculous, tech-bubble madness.
(twitter.com/_/status/525340134938402816)
KevinMarks_ joined the channel
#
@oliverdore
Indie web building blocks | @adactio on IndieAuth, Webmention, POSSE and more to encourage a more independent web. https://t.co/tQNec2dqhB.
(twitter.com/_/status/525350294381948929)
#
gregorlove.com
edited /Ello (+214) "/* No Export */ twitter quote"
(view diff)
chrissaad joined the channel
#
KevinMarks_
ah, no, that's just the "posts the last note on reboot if the tab was open" bug
#
aaronpk
oh dear
#
tantek
gRegor`: indeed. how about actual hostility in those tweets?
#
gRegor`
Hm?
#
tantek
"ridiculous … madness"
#
KartikPrabhu
more like incredulous idiotic arrogance
#
@cacheflowe
@patr1ck Regardless, speculation about a feature that hasn't even been developed is silly. Don't use the platform of you don't want to :)
(twitter.com/_/status/525351042792554499)
#
gregorlove.com
edited /Ello (+228) "/* Plumbing */ Funding section"
(view diff)
#
tantek.com
edited /Ello (+416) "related ycomb thread"
(view diff)
wolftune joined the channel
#
KevinMarks_
hm, the "success" of my google email tweet means my mentions are noisy today
#
gRegor`
Well with that warm Ello welcome . . . anyone want an invite?? ;)
grantmacken joined the channel
#
aaronpk
omg 231 RTs kevin?!
chrissaad joined the channel
#
tantek
wonders how well Known handles 231 RTs ;)
#
@IndieGameGuys
RT oliverdore: Indie web building blocks | adactio on IndieAuth, Webmention, POSSE and more to encourage a more ... http://t.co/mKGQwCvlll;.
(twitter.com/_/status/525352041888362496)
#
KevinMarks_
24,000 impressions
willowbl00 joined the channel
#
KevinMarks_
wow, twitter backgroun image in chrome is weirdly broken
#
@4c4d
@cacheflowe @patr1ck fwiw i have an ello account and i fully support the idea but please at least try to understand #indieweb concerns
(twitter.com/_/status/525353643156193281)
#
ben_thatmustbeme
yay, finished that git clone, only took 40 minutes
#
Loqi
yay!
#
aaronpk
uh, I just got a random baby's blood work faxed to my fax number
#
aaronpk
that seems like a thing that should not be possible
pauloppenheim joined the channel
#
aaronpk
it just showed up in my dropbox
#
aaronpk
there isn't even a phone number on the document that I can call back
#
@meyerweb
I have seen these UIs, and seen floor staff struggle with them. It really is this bad. “How Bad UX Killed Jenny”: https://medium.com/@designuxui/ef915419879e
(twitter.com/_/status/524179582652407809)
chrissaad joined the channel
#
@cacheflowe
@patr1ck We've been responding *constantly* to users on the system. Not as much to Twitter mobs.
(twitter.com/_/status/525356797713199104)
#
tantek.com
edited /User:Tantek.com (+227) "move process h-card issues to top of community working on, expand invitation per IRC brainstorm"
(view diff)
#
aaronpk
well at least they're dogfooding :)
#
kylewm
someone should advise him to stop talking on twitter until he cools off
#
aaronpk
i dunno, this is pretty fun to watch
#
reedstrm
yrs so ... inclusive of them.
#
reedstrm
s/yrs/yes/
#
Loqi
reedstrm meant to say: yes so ... inclusive of them.
#
tantek
he's also sharecropping on Twitter right? Ello doesn't POSSE to Twitter does it?
#
ShaneHudson_
"So charging a buck or two for an export feature is hostile? This is ridiculous, tech-bubble madness."
#
ShaneHudson_
Wow, who is this guy lol
#
tantek.com
edited /User:Tantek.com (+166) "/* working on */ invitation - incorporate / bi-directional merge with uf wiki coverage"
(view diff)
#
@benwerd
@timdavies Heading back to Oxford in December. Would love to connect with the OII and discuss indieweb / @withknown if there's interest?
(twitter.com/_/status/525358550189871105)
#
tantek
ShaneHudson_: apparently one of the builders of Ello, and someone who just got millions of $ in VC
#
ShaneHudson_
The tech world is messed up.
shiflett joined the channel
#
benwerd
(Hey, maybe I should charge for parseable HTML)
#
aaronpk
benwerd: :P
#
tantek
!tell benwerd is de-dupping of RSVPs on the Known to-do / issues list? e.g. see http://known.kevinmarks.com/2014/homebrew-website-club-october-22-2014 and note the duplicate RSVP from the FB POSSE copy.
#
Loqi
Ok, I'll tell them that when I see them next
#
benwerd
!tell tantek yes, it's on my very very very near term list (i.e. today) ;)
#
Loqi
benwerd: tantek left you a message 29 seconds ago: is de-dupping of RSVPs on the Known to-do / issues list? e.g. see http://known.kevinmarks.com/2014/homebrew-website-club-october-22-2014 and note the duplicate RSVP from the FB POSSE copy.
#
Loqi
Ok, I'll tell him that when I see him next
#
tantek
Hey benwerd that's not a bad seed of an idea
#
Loqi
tantek: benwerd left you a message 8 seconds ago: yes, it's on my very very very near term list (i.e. today) ;)
#
tantek
simple valid HTML+microformats - free
#
tantek
JSON version - charge $
#
tantek
XML version - charge $$
#
benwerd
SOAP - you'd better have a three-year enterprise license
#
reedstrm
Hey ,what about the SGML?
#
tantek
JSON-LD version charge $$$ (because triples)
#
kylewm
benwerd: note that my three facebook.com RSVPs on that note are not Known's fault -- my fault for creating three events accidentally
#
KartikPrabhu
remor.se a content silos that converts everything you write into morse code and charges for every other format you want
#
reedstrm
KartikPrabhu++
#
Loqi
KartikPrabhu has 72 karma
#
tantek
benwerd - immediate implementation is to provide rel=alternate links to those, and the just return 402 Payment Required
#
tantek
s/the just/then just
#
Loqi
tantek meant to say: benwerd - immediate implementation is to provide rel=alternate links to those, and then just return 402 Payment Required
#
KartikPrabhu
booo that domain is taken :( much remorse
#
benwerd
kylewm: huh - but nonetheless interesting to dedupe on profile ID.
#
aaronpk
hey bridgy people, do you know what would happen if many people created a facebook event with one of my event URLs at the end? would all the event RSVPs end up on my site?
#
tantek
I'm not kidding about this. The Atom version of my posts costs me more bandwidth, so why shouldn't I pass that long to the consumers?
#
benwerd
tantek - I think it's reasonable to provide a JSON-LD version of the 402 page for everyone. Call it freemium
#
tantek
s/long/along
#
aaronpk
tantek: I like the idea of my own site returning "402 payment required" along with an html credit card form
#
Loqi
tantek meant to say: I'm not kidding about this. The Atom version of my posts costs me more bandwidth, so why shouldn't I pass that along to the consumers?
#
aaronpk
why should silos be the only ones that can charge!
alanpearce joined the channel
#
tantek
benwerd - the JSON-LD version of the 402 page should require 3 forms of identity along with credit card
#
kylewm
aaronpk: yes if all those people creating the events were signed up for bridgy
#
benwerd
It is not at all unreasonable to charge for different formats
#
KartikPrabhu
was listening to some web payment stuff the other day
#
aaronpk
we already have web payments. html forms where you enter your credit card number :P
#
KartikPrabhu
but but protocols
#
tantek
aaronpk - that has a dependency on the credit card oligarchy
#
aaronpk
that's why there was a :P at the end
#
tantek
it's like saying, we already have web image messaging, HTML forms where you enter your fax number
#
reedstrm
tantek hopefully implement reciprocity deals w/ common discussion partners.
#
tantek
reedstrm - reciprocity?!? why the heck would I spend the extra time to consume more complex formats?!?
#
tantek
I would charge 100x more for requests to *consume* those formats
#
reedstrm
well if you charge and they charge, ... be happy in your own personal silo ...
#
tantek
reedstrm - the economics of simpler formats are already winning that attrition war
musigny joined the channel
#
benwerd
sidles away from this conversation and goes to find food (where he will be charged differently for consuming different things)
#
tantek
it's those with more complext formats / protocols that will end up in smaller and smallers ilos
#
tantek
s/ilos/silos
#
Loqi
tantek meant to say: it's those with more complext formats / protocols that will end up in smaller and smallers silos
#
reedstrm
meh, then don't bother serving it at all - too much hassle to charge.
#
tantek
that's a challenging trade-off to evaluate yes.
#
gRegor`
@cacheflowe doesn't even link to his ello from his homepage http://cacheflowe.com/
#
tantek
however I think the MVP of a 402 page is not bad to send a message.
#
gRegor`
Or his Twitter bio
#
KartikPrabhu
wow! firefox account signup just told me my birth year was "1990 or earlier"! I feel old
#
tantek
how would you characterize that as a criticism gRegor` ?
#
reedstrm
tantek: hehe. However, it's almost a promise to actually send that if someone would pay up. Then your stuck supporting it.
#
kylewm
KartikPrabhu: do years even go back that far??
#
tantek
benwerd, but seriously, charging for complex formats seems like a good idea, especially when its enterprise (those with lots of $) who ask you to publish them.
#
gRegor`
Probably just an oversight / stale website. Last post is from April
#
KartikPrabhu
kylewm: I'll show you youngun'
#
reedstrm
y'allcvan get off my lawn, right now.
#
KartikPrabhu
<shakes fist>
#
reedstrm
s/cvan/ can/
#
Loqi
reedstrm meant to say: y'all can get off my lawn, right now.
#
gRegor`
Guessing he's worked on Ello for quite some time
#
tantek
reedstrm - good point. I'll start with a $1m check in the mail please. per format per request. that should be a good starter 402 page.
#
reedstrm
we have a winner!
chrissaad joined the channel
#
tantek
reedstrm - could even do it kickstarter style
#
tantek
and then offer it "for free" once the minimum has been met.
#
ShaneHudson_
My site needs $8m funding btw. Let me know if anyone finds someone 'clever' enough to see the ivestment opportunities, they seem to be everywhere in SV
#
tantek
then you just have to set a high enough minimum. and wait for some philanthropist/angel/vc to come along and "throw money at the problem"
#
tantek
ShaneHudson: start returning a 402 page that says so ;)
#
tantek
KartikPrabhu: ah the splashy-webpage-before-you-have-anything-working project antipattern
#
aaronpk
it's a nice splashy web page tho
#
tantek
spec != working ;)
#
KartikPrabhu
tantek: yup. didn't the responsive image people start that?
wagle joined the channel
#
KartikPrabhu
for the podcast I listened : http://5by5.tv/webahead/82
#
ShaneHudson_
KartikPrabhu: You are probably thinking of the spec for element queries
#
reedstrm
"The purpose of PaySwarm is to build payment into the core architecture of the Web." Not even April 1!
#
KartikPrabhu
Manu made it almost seem like web-payments will solve poverty, world hunger and cure cancer all at the same time
#
aaronpk
interesting, it seems to use URLs as identifiers for people to pay, that's nice
#
tantek
KartikPrabhu: common for many projects
#
tantek
I suggest if you're interested in web payments, that you start here...
#
tantek
what is payment?
#
Loqi
payment in the context of the indieweb refers to a feature on an indie web site that provides a way for the visitor to that website to pay (currency, gift card credit, etc.) the person represented by that indie web site http://indiewebcamp.com/payment
#
aaronpk
e.g. "source": "https://bluebank.com/people/john/accounts/party-funds", "destination": "https://redbank.com/people/jane/accounts/hosting"
#
KartikPrabhu
tantek: yup have seen that page :)
#
tantek
the rest (payswarm web-payments etc.) is so far away from any kind of indieweb implementation that it's ignorable IMO - unless someone wants to try building it on their own site here?
#
KartikPrabhu
aaronpk: I don't think I want my bank to be my identity
#
aaronpk
indeed
#
reedstrm
I think there's a Freakanomics story on how more people are willing to do community group activities for free than if you offer to pay them.
#
KartikPrabhu
tantek: will you angel fund it? :P
#
tantek
KartikPrabhu: no, that's for your itches to fund to your scratches. ;)
#
tantek
pure self-interest, as adactio put it
#
reedstrm
cf: stackexchange vs. (some we'll micropay you to answer tech questions site that cratered and I can't remember the name of)
#
aaronpk
the indieweb version of this payswarm thing would be using your own domain as the identity and linking to one or more banks from your website
#
tantek
reedstrm: indeed I think I've read articles like that in a bunch of places.
#
aaronpk
rather than using the bank URL as your payment identity
#
tantek
aaronpk - why do you need a "bank"?
#
aaronpk
"bank" in a loose term
#
aaronpk
some entity that is actually transferring the money
#
tantek
aaronpk, why isn't the transfer merely some signed crypto tokens peer to peer?
#
aaronpk
same reason that being able to delegate to an authorization server is useful, because not everyone wants to be/make their own auth server
#
aaronpk
because blockchains are not sustainable
#
tantek
what are blockchains?
#
Loqi
It looks like we don't have a page for "blockchains" yet. Would you like to create it? http://indiewebcamp.com/wiki/index.php?action=edit&title=blockchains
#
reedstrm
There was a recent SMBC something like that - the economic analysis of 'the ultimate game' : I give you $100, w/ the stipulation that you have to offer some fraction of it (any fraction at all) to aaronpk. If he accepts your offer, you both keep it. Otherwise, deals off.
#
aaronpk
one nice design decision of payswarm is this: "While this algorithm outlines how transfer records are transmitted and recorded in a decentralized fashion, it does not outline how each authority ensures that currencies are exchanged between financial institutions."
#
reedstrm
Rational economics says he should accept any offer at all.
#
tantek
yeah - like I said, IMO payswarm is ignorable because it's so far off in the weeds. if you find something useful, maybe document it on a /payment-research page
cweiske joined the channel
#
aaronpk
i think there is likely some useful things in there
#
tantek
reedstrm - yes I think's a classic game. I don't have a citation, but no need to repeat it here. Perhaps you have a citation.
#
tantek
(games theory game)
#
gRegor`
What's the b/ 20?
#
tantek
aaronpk - "some useful things" -> document the research
#
aaronpk
gRegor`: sorry, fatfingered my tab switcher
#
aaronpk
should have been /b 20
#
gRegor`
Ah
#
KevinMarks_
do I need to wake up indiecreddit again, or should i wait until next april?
#
KartikPrabhu
KevinMarks_ next April and make a really fancy website for it, with parallax scrolling and all that
#
reedstrm
KevinMarks: better start coding now
#
KevinMarks_
ultimatum game
#
tantek
KevinMarks - the joke gets more interest(ing) with each year.
#
aaronpk
bahahaha
#
aaronpk
tantek++
#
Loqi
tantek has 110 karma
#
KevinMarks_
the really interesting thing about the ultimatum game is how it varies by culture
#
tantek
now redirects KevinMarks to #indiechat ;)
#
KevinMarks_
in cultures of obligation, people reject too high a share
Pierre-O joined the channel
#
GWG
!tell snarfed Bridgy just sent a bunch of missing mentions
#
Loqi
Ok, I'll tell them that when I see them next
jet___ joined the channel
#
GWG
I need someone to bounce thoughts off of again
#
GWG
Still trying to plan something out
#
GWG
It keeps growing.
#
GWG
Is that what they call feature creep? Or is that something else?
#
reedstrm
GWG: smells like it: with out any details, hard to tell. Usually feature creep involves adding 'just one more feature' to a milestone/release.
#
reedstrm
closely related phenomena: "While we have the wall open" during a home remodel. The "since I'm in this code anyway" refactor. Any one have others?
#
aaronpk
I totally did the "while we have the wall open" thing IRL
#
GWG
reedstrm: I do the...while we have the wall open one all the time
snarfed joined the channel
#
@ShaneHudson
@cacheflowe Okay, well I look forward to seeing what you do. Please take some advice from http://indiewebcamp.com/ and use the funding well.
(twitter.com/_/status/525372913504235520)
#
GWG
reedstrm: Usually its...I'm calling in the electrician....do I need him to do anything else while he is here...because I get a better deal if I fill his time more effectively
#
aaronpk
managed to add network jacks into every room of the house that way :)
#
GWG
aaronpk: Resale value!!
#
snarfed
GWG: glad to hear it!
#
reedstrm
hehe - that's the problem with all of these: they actually work, some of the time.
#
Loqi
snarfed: GWG left you a message 13 minutes ago: Bridgy just sent a bunch of missing mentions
#
reedstrm
unfortunately it was 20 years ago and is all cat3, right? :-)
#
GWG
snarfed: But how did they come back after it reported no post links found? I thought you might be tinkering
#
aaronpk
it's cat5!
#
tantek
ShaneHudson: for ending a thread on a positive suggestion
#
tantek
ShaneHudson++
#
Loqi
ShaneHudson has 11 karma
#
GWG
But, my problem is that I want to add features, but I see this massive way it could go if I continue along the same path...except I want to find an intermediate point in case I don't.
#
snarfed
GWG: when it finds the first new syndication link (per silo), it goes back and reprocesses
#
KartikPrabhu
GWG: add features that you'll use
#
KartikPrabhu
nothing more nothing less
#
GWG
I'd use them all
#
KartikPrabhu
hmmm ok can't help you with prioritsation
#
GWG
KartikPrabhu, it isn't quite that. It is more...I know I want to go from A to B...but A to B is big...I need to figure out how to work incrementally toward B, and figure out intermediate point C
#
GWG
This started with adding the layer so posts can be responses to other content...Reply, Like, etc.
#
KartikPrabhu
i think that is called creative thinking/problem solving :P
#
GWG
I started with simple metadata...a URL, a title, and a citation/content box.
#
tantek
GWG - why not start with a sketch instead?
#
GWG
I want to add to that with more metadata, including date, author name, etc.
#
tantek
build to that sketch, ship it. then expand later.
#
GWG
tantek: I haven't gotten to the scoping problem yet
#
GWG
The design problem I'm clear on how to figure out.
#
tantek
GWG, still talking reply contexts right? Start with implementing each level one at a time: http://indiewebcamp.com/reply-context#Levels
#
gregorlove.com
edited /Ello (+285) "/* Indieweb Support */"
(view diff)
#
GWG
This is the issue that is stopping me from moving forward right now
#
GWG
If I get into profile pictures and author names...the eventual endgame for that is a complete profile of the author, because it leads into all sorts of other things in the future. That includes things like authentication, messaging, etc.
#
GWG
Or, the one at the top of my list, whitelisting on the assumption people I wrote about should be a whitelisted group backward.
#
reedstrm
GWG followed all that except the last "whitelisted group backward" ?
tantek_________ joined the channel
#
GWG
I'm not sure what I meant by backward either
#
GWG
But, anyone I talked about should be considered trusted by me, and added to a trusted group
#
reedstrm
right, it works without it.
#
GWG
The issue is this basically...
#
GWG
Right now, if I don't want to do full profiles, then the author data for what I'm responded to is stored in the data for the post itself
#
GWG
If I do want to do full profiles, I would have to store it elsewhere.
#
GWG
Switching between the two later seems like it would be a pain in the neck
tgbrun joined the channel
#
reedstrm
Consider the 'stored in post' part a convenience cache for the full-bore profile. Implement it first. Ah what are you doing full profiles for others, anyway? Don't they have pages for that?
#
GWG
reedstrm: Storing profiles is a method of caching the data. Keeping it in the post means I'd have the same data repeated for people.
tantek__________ joined the channel
#
tantek__________
GWG, there is no eventual endgame. (Re)think, (re)design, (re)build, (re)deploy, (re)use, (re)iterate.
#
GWG
tantek_________: I just hate writing migration tools
#
reedstrm
Ah got it. Well, I'm always in favor of the DRYer approach. Store a profile, but allow it to be partial. A miss in a cache just means go look it up.
#
GWG
tantek_________: Also, the endgame I'm referring to is for my current vision, not the end of time
#
tantek__________
As reedstrm said, don't migrate but treat the earlier version as optional cache for the later version
#
GWG
Then I got to the difference between people and organizations
#
tantek__________
Your vision will always outpace your code & design. Again, there is no "eventual" nor "endgame".
#
reedstrm
Had to code that for search results: author info on each returned item, but wanting facet filters for by author. Answer was to dry it out, return all the people as part of the result set, and ref. it in the actual item display.
#
GWG
I want to have, for example, a tantek profile because I may respond to him regularly. But if I'm commenting on an article in a journal, the author is often superseded by the publication.
#
tantek__________
Always treat the author as primary. Publisher as optional affiliation.
#
GWG
tantek_________: I've just added another field to my dataset though.
#
tantek__________
If there is no name for the author, show "unknown author" to make it explicit
#
tantek__________
Then ignore publisher / publication completely for this iteration.
snarfed joined the channel
#
rascul
too much underscore!
#
GWG
Okay...simpler question
#
GWG
How long should you cache a profile picture?
#
tantek__________
First implementation: until someone complains about something looking wrong.
#
reedstrm
simple - he called cache invalidation simple!
#
Loqi
reedstrm has 7 karma
#
GWG
reedstrm: Simpler. It's a question that calls for a definitive answer.
#
GWG
As opposed to a question that was open-ended
#
reedstrm
o.k. simpler question, not topic. :-) Lacking any other info, if you have the modification date of the object, I scale cache time to the age: if they haven't updated it since 1992, it's likely to be good for a long while. Lacking that, 1 day/week/etc. (ie. pick a number)
snarfed joined the channel
#
reedstrm
(algo stolen from defaults for squid, back in the day)
snarfed joined the channel
#
GWG
reedstrm: To start, I'm not caching at all.
#
GWG
I just want to build it so it is easy to add later
#
reedstrm
what lang/platform? Dropping in memcached is pretty simple, usually.
#
GWG
reedstrm: WordPress, and thus PHP.
#
reedstrm
Just make sure you've got a 'get' method you use consistently to hook.
#
GWG
reedstrm: That is why I'm planning. I want to abstract some things so I can hook in later.
#
GWG
Hmm...then the logical thing to do is write a function that retrieves the data from wherever it is stored. Then change where it is stored without changing the function that retrieves it
#
GWG
is reading nicknames-cache in the wiki
#
reedstrm
GWG exactly - a getter method.
#
reedstrm
or storage abstraction layer, if you've got lots of these.
#
GWG
reedstrm: Well, if you've ever looked at WordPress, not as many as you would think.
#
reedstrm
suppresses shudder
#
reedstrm
Let's just say I'm not a PHP guy and leave it there, o.k. ? :-)
#
GWG
reedstrm: I'm not much of a programmer in general. I've always understood code, but rarely wrote it. I've gotten back into it via Indiewebcamp, because I want to add some functionality, but it is mostly centered around metadata.
#
reedstrm
The way to be a programmer is to ... program. Surprisingly few of the people who go to school to learn to program read code. This confuses me. It's like authors who don't read.
#
GWG
reedstrm: I studied Pascal in high school. And a little Java in college. Not exactly relevant. All the current languages I did on my own
#
GWG
Unless you need a BASIC program written for the Commodore 64, or you want me to activate the secret 320x240 mode on a VGA adapter using C...I'm pretty out of date on a bunch of things
#
GWG
Ah, Mode X.
jsun_ joined the channel
#
GWG
I feel nostalgic just thinking of it
snarfed joined the channel
#
GWG
Anyway....back to Indieweb
#
reedstrm
Enlightment usually happens around the 3rd language - it's all the same.
#
cweiske
6 languages and I still like PHP most
shiflett joined the channel
#
cweiske
s/6/8/
#
Loqi
!calc s/6/8/
#
Loqi
cweiske meant to say: 8 languages and I still like PHP most
#
GWG
I just like key indexed arrays for metadata storage.
krendil joined the channel
#
ShaneHudson_
I used to say I've got varying experience with about 26 languages.. but now I just say I know JS
musigny joined the channel
#
reedstrm
yup hashes are great, whatever they're called.
#
ShaneHudson_
Seems knowing a language is more about knowing the quirks than programming it
indie-visitor joined the channel
#
GWG
Hello, tgbrun
KartikPrabhu joined the channel
#
tgbrun
GWG: Thanks for the shoutout!
#
tgbrun
I am a noob and have some questions re: indieweb, but I don't want to interrupt any ongoing conversations that are going on
#
GWG
tgbrun: Ask away
#
GWG
I was just rambling on about my current obsession. I need a break to digest
#
GWG
Besides, I'm on vacation till tomorrow
#
reedstrm
irc's always about the interleaved threads, anyway :-)
#
gRegor`
Welcome, tgbrun. Ask away.
#
GWG
I'm not in the right frame of mind to code on a Chromebook converted to Fedora
#
tgbrun
GWG: I have a website and am sloooooowly implementing indieweb principles. I link my posts to facebook and twitter through nextscripts and am looking to get the comments and likes, etc. back.
#
ShaneHudson_
That will send you webmentions, which you can show on your site :)
#
ShaneHudson_
what is webmention
#
Loqi
Webmention is a simple way to notify any URL when you link to it on your site http://indiewebcamp.com/Webmention
#
GWG
tgbrun: Nextscripts drives me crazy, but it does the job. It supports comment import, but not like import.
#
tgbrun
I have installed rel-syndication but that adds an "also on" tag on even pages, which I don't share.
#
GWG
tgbrun: Either way, extra welcome for being a WordPress user
#
GWG
tgbrun: There is an alternate to rel-syndication
#
ShaneHudson_
Right I'm going to grab an early night, goodnight all
#
Loqi
don't let the bed bugs bite
#
tgbrun
I also looked at syndication link and web comments plugins but I don't see how to set them up to find the comments, etc. I have subscribed to bridg.ly. Is there anything else I need to do?
#
snarfed
tgbrun: this css should hide the "also on" on pages only:
#
snarfed
div.u-syndication { display: none; }
#
snarfed
body.single div.u-syndication { display: block; }
#
GWG
tgbrun: Web comments? Not familiar with that. Do you have Semantic Linkbacks?
#
GWG
As for Syndication Links, which is my alternative to WordPress Syndication, aka rel-syndication, it is 100% manual entry right now.
#
GWG
snarfed: By the way, I couldn't get Bridgy to pick up a u-photo class
dariusdunlap_ joined the channel
#
snarfed
tgbrun: if you've signed up for bridgy, and you have the indieweb plugin installed, that should be it. details: http://indiewebcamp.com/WordPress#backfeed
#
snarfed
GWG: ugh, sorry to hear it. did you look at the bridgy publish log? any clues?
#
tgbrun
yes to semantic linkbacks and to semantic comments. Do these plugins play well together, I'm confused about which I should have and which don't work with others
#
GWG
snarfed: It is having trouble with the markup around the img markup. <figure>
#
GWG
tgbrun: Semantic Comments requires Semantic Linkbacks.
#
GWG
tgbrun: I wrote Semantic Comments to change the display of comments to show the mentions separately from the comments
#
GWG
tgbrun: I've been refining it. I have a change planned for it
#
tgbrun
GWG: I have both SC and SL installed and activated
#
snarfed
GWG: yeah, i see the blank photo property. looks like your <img> elem doesn't have a closing /. maybe that's it? silly if so, but meh
#
reedstrm
html/xml xhtml html5, it's all good. not.
#
GWG
tgbrun: You need someone to like something and send it in for Semantic Comments to do anything
jet___ joined the channel
#
GWG
snarfed: I think it was when I tried to implement responsive images
#
GWG
I'd better fiddle with that plugin
KartikPrabhu joined the channel
#
snarfed
tgbrun: are you tombruning.com? if so, looks like bridgy is failing because you currently requiring js and cookies for comments. that's probably due to a plugin; you'll need to loosen that
#
GWG
tgbrun: Do you have a comment plugin enabled?
#
GWG
Semantic Comments would conflict with that
#
GWG
Note to self...write that in the description
#
tgbrun
GWG: So I should deactivate SC?
#
GWG
tgbrun: Do you have a comment plugin, such as JetPack comments, Disqus, etc enabled?
#
GWG
Semantic Comments wasn't designed to coexist
KartikPrabhu joined the channel
#
tgbrun
GWG: Jetpack comments aren't activated
#
tgbrun
Webmention for comments is activated as well
#
GWG
tgbrun: Then Semantic Comments isn't the culprit
#
GWG
tgbrun: It doesn't use Javascript
#
GWG
Or cookies
#
GWG
Well, not specifically
#
tgbrun
GWG: I wonder if it is wp-spamshield
danlyke joined the channel
#
KartikPrabhu
Q: over at #microformats -- is there a microformat for a table of contents ?
#
snarfed
tgbrun: sounds likely
#
GWG
tgbrun: It could be.
#
GWG
tgbrun: Either way, if you ever need help, happy to try. If its with one of my plugins...feel free to file any feature request, bug, etc.
#
tgbrun
I deactivated spamshield and ran another poll, bridg.ly found the response, but it failed for another reason. Still investigating...
EOGreer_, EOGreer and indie-visitor joined the channel
j12t joined the channel
#
tgbrun
GWG: disabling wp-spamshield did the trick. I am now getting comments and likes back from Facebook
#
GWG
tgbrun: Great
#
tgbrun
GWG: now that I disabled and will uninstall spamshield, which did a great job of stopping spam (and useful comments as well) what would be a good way to stop spam?
#
GWG
tgbrun: Do you accept comments on your blog?
#
tgbrun
GWG: yes, I would rather folks comment on the blog rather than on FB, unless that is a non-optimal solution
#
GWG
tgbrun: There are many plugins out there for spam on WordPress
#
tgbrun
GWG: I'm sure, but I'm looking for one that plays 'nice' with indieweb
#
GWG
Akismet?
fmarier joined the channel
#
tfontaine
I always had good luck with Akismet as well
#
Loqi
GWG has 23 karma
#
tgbrun
snarfed: Thanks for your help,
#
tgbrun
GWG++
#
Loqi
GWG has 24 karma
#
tgbrun
snarfed+
#
GWG
I think snarfed left
#
tgbrun
I still appreciate his help
#
GWG
tgbrun: He'll see it in the logs when he next comes in.
#
bret
KartikPrabhu_: are you using pandoc?
#
bret
(for the article
jet___, chrissaad and chalettu joined the channel
#
bret
!tell KartikPrabhu Your use case is turning your paper/book into another format, such as LaTeX, from the HTML/MF2 copy, and don't want to degrade the organizational structure
#
Loqi
Ok, I'll tell them that when I see them next
#
bret
(maybe you already had that understanding, it just occurred to me)
dariusdunlap, Mark87, benwerd, KartikPrabhu, gr0k and chalettu joined the channel
#
reedstrm
bret: re: AS2.0 spec - check todays logs, we had the heads up earlier from the WG co-chair. Who hangs out here. (tantek) :-)
#
bret
my bad, have not quite gotten that far back yet
#
reedstrm
no problem, just shows you're right, this crowd is interested in that.
#
reedstrm
bret: you r comment re: pandoc and maintaining semantics - are you imaging the HTML/MF2 as canonical?
#
bret
KartikPrabhu was talking about writing an article in html/mf2
#
bret
so yeah in that case, it would be
#
bret
npdoty was doing something similar with a markdown -> html/latex system
#
bret
(and others via pandoc)
#
reedstrm
Now it's my turn to miss something in the scrollback :-)
#
KartikPrabhu
pandoc is annoying. I have to write YAML side files to get author details and abstract etc...
#
Loqi
KartikPrabhu: bret left you a message 19 minutes ago: Your use case is turning your paper/book into another format, such as LaTeX, from the HTML/MF2 copy, and don't want to degrade the organizational structure
#
KartikPrabhu
reedstrm: I asked on #microformats
#
reedstrm
Ah, I'm not hanging there.
#
KartikPrabhu
bret: yes. I want the HTML+mf2 to specify the semantics and structure and then other formats could follow as needed
#
Loqi
gives KartikPrabhu the HTML
#
KartikPrabhu
thanks Loqi
#
Loqi
you're welcome
#
bret
KartikPrabhu: an html/mf2 to latex document converter would be super cool
#
bret
(.tex that is)
#
KartikPrabhu
bret: yes! I don't know of one but that looks like a huge undertaking specially trying to move around TeX stupidity
#
bret
i'm not sure how well the math would survive, you might have to rely on mathjax or something for the html math stuff
#
KartikPrabhu
bret: yes doing that currently
#
reedstrm
we're doing something similar, though not MF2 specifically - html5 but using CSS3 as the transform language, if you will. Currently targetting outputs to web, epub, and Pdf (via princexml, I'm afraid. Propriarty bit)
#
bret
reedstrm: neato!
#
bret
im not terribly familiar with mathjax, does it support more advanced environments and extensions? (like align and aboxed?)
#
bret
if you just did the subset of TeX that aligns well with mf2 + a few new properties it wouldnt be that gigantic of an undertaking
#
bret
reedstrm: is that project public?
#
KartikPrabhu
bret: yes mathjax does align and all that too!
#
bret
tight, i should really play with that
#
bret
is there a google + to atom feed generator anywhere?
#
KartikPrabhu
should find a way to make mathjax processing faster
#
bret
KartikPrabhu: perrrdddyyy
#
joskar
I've spent the last half hour wondering why my code for accepting webmention comments wouldn't work. Turns out I was refreshing the wrong page :)
#
bret
looks great
#
KartikPrabhu
joskar: ha :)
#
KartikPrabhu
i've done that before :P
#
KartikPrabhu
should write more of that relativity series
snarfed, Pierre-O, chalettu and RichardLitt joined the channel
tantek joined the channel
#
j12t
It's quiet on this channel, so let me ask the QOTD:
#
j12t
In the Indie Web of Things, does each thing have its own domain?
#
Loqi
j12t: benwerd left you a message 1 week, 5 days ago: Interesting feedback from timmmmyboy - would like packages vs an entire distribution (helps run Reclaim Hosting)
#
benwerd
j12t: or subdomain?
hober joined the channel
#
tantek
benwerd++
#
Loqi
benwerd has 43 karma
#
tantek
j12t - I'd say it depends on when you want to give your things an identity with a separate security context (then (sub)domain), or part of an existing security context (your choice, like your indieweb of posts).
#
j12t
(Context: I just chatted with Scott J about that "physical web" proposal which, in may ways sounds like LID. http://upon2020.com/blog/2014/10/the-google-physical-web-proposal-and-lid/ )
snarfed joined the channel
#
j12t
Whose (master) domain would it be anyway? Vendor, or buyer?
hober joined the channel
#
tantek
vendors should never own identity
#
tantek
except their own
#
tantek
things that likely deserve their own security context: a person, an organization, software that runs autonomously (agent, AI)
#
j12t
tantek: so you are saying that if I go to Lowe's and pick up a thermometer, and to the Home Depot and Office Depot and so forth, when I get home I'll then proceed to enter them all in my GoDaddy control panel?
#
tantek
no, your indiebox control panel of course ;)
#
danlyke
j12t I've been thinking about local URL shorteners with common enough formats that a Greasemonkey script can replace them.
#
tantek
then it handles whatever obfuscation it needs to with wherever your (sub)domains are hosted
gr0k joined the channel
#
j12t
danlyke: what problem would that solve?
#
tantek
another particular concern of getting into the "indieweb of things" is privacy about the existence of your things, e.g. at home, which would imply that we may need to address allocation of (sub)domain based identities which don't leak outside your home.
#
j12t
tantek: well, subdomains leak via DNS. Sub-URLs wouldn't.
#
danlyke
ie: local links happen as http://www.flutterby.com/resolve.cgi?id=[device-and-document-public-key] , there's something common to any site which uses this scheme to do resolution (so links on your site are only dependent on your server), but eventually browsers can use something like the Bittorrent blockchain to resolve that key. Releasing us from DNS tyranny.
#
tantek
j12t - that depends, you could dynamically allocate them / randomize them from your indiebox as needed
#
tantek
danlyke - take a look at existing algorithmic URL shorteners that have solved this problem
#
tantek
what is algorithmic?
#
danlyke
tantek, my hope isn't URL shortening, it's an address to IP mapping that avoids the DNS tax.
#
Loqi
Algorithmic is a term often used to refer to data such as URIs or other identifiers. An identifier is algorithmic if there is a way to decode it into additional information, typically about the thing it identifies http://indiewebcamp.com/algorithmic
#
j12t
danlyke: are you talking about a version of content-addressable addressing?
#
tantek
danlyke - like I said, take a look at the existing solutions those of us that have solved this have deployed
#
tantek
j12t lol at content-addressable addressing
#
tantek
I had an epiphany this morning about the domain name rental / stealing problem.
#
tantek
I think there might actually be a really simple solution
#
j12t
For the stealing part, or for the prevention part?
#
tantek
for recovery
#
tantek
and longevity
#
tantek
even if you let your domain expire
#
tantek
here's how
#
tantek
1) always post permalinks that include the date of the post in the permalink. e.g. either /YYYY/MM/DD/whatever-you-want or /YYYY/DDD/whatever-you-want per ISO8601
#
tantek
2) when storing a link to a / someone else's post, *always* store the end result of redirect resolution, and then rel=canonical following.
#
danlyke
j12t maybe... I'm trying to figure out how to give, say, my sister some sort of identity/server/document mapping that doesn't require her to pay an extra $10/year (and the attention to it), because that's part of the barrier to her moving off of Facebook/etc.
snarfed1 joined the channel
#
tantek
3) submit the link you stored to all the internet archives for long term archival
#
tantek
4) if years down the road one of your stored links fails to return "the same thing" as before (e.g. change of rel=canonical, 404, etc.), go pull that URL out of one of the internet archives, *USING* the date embedded in the very URL itself as the time-window to look for it.
#
tantek
boom - link repaired back to an archived version.
#
tantek
does not matter if you let the domain expire.
#
tantek
BONUS: you can even use this technique to *verify* rel=canonical changes by retrieving the *archived* version as above, comparing it to a new rel=canonical / redirect destination, and if it's "similar enough", updating your storage of the canonical URL.
jet__, erlehmann_ and erlehmann joined the channel
#
danlyke
so tantek, what portion of distributed URL de-shortening out there addresses my concerns about eventually being able to do that local to the browser?
snarfed joined the channel
#
bret
snarfed++
#
Loqi
snarfed has 49 karma
#
tgbrun
snafed: Your suggestion for the css changes to have the "also on:" not appear on pages, just posts worked great
#
tgbrun
snarfed++
#
snarfed
tgbrun: glad to hear it!
#
Loqi
snarfed has 50 karma
#
tgbrun
Now if I could find a way to do 'notes' as well as posts in WP, my day would be complete :)
#
kylewm
hmm, the algorithmic URL shortening doesn't really have any bearing on DNS distributed or otherwise, afaik
#
kylewm
danlyke^^ probably just confirming what you already know
KartikPrabhu joined the channel
#
danlyke
kylewm that's what I thought, just wanted to make sure I'm not missing something.
KartikPrabhu and tantek_ joined the channel
#
tantek_
kylewm, danlyke, it does, when the algorithm unpacks to a URL with a path starting with a date
#
tantek_
Per the above algorithm I just documented.
#
tantek_
What it does it make your rental of a domain name over a time period actually work for permalinks without having to rent the domain name for all time.
benwerd_ joined the channel
#
tantek_
Additional step you can take. Make sure that the date in the URL is the same as the date in the dt-published in the h-entry of the page
#
KartikPrabhu
tantek_: is this also supposed to prevent someone taking my domain name and then posting something else at the same URL?
#
tantek_
Kartik it raises the barrier for that considerably.
caseorganic joined the channel
#
KartikPrabhu
hmmm... how does the date in the URL help?
#
danlyke
I think the point of the date in the URL was that it helped in retrieval from the archiving silos.
#
KartikPrabhu
don't these archiving sites store a date anyway?
#
KartikPrabhu
does not know really
#
kylewm
they do, but you correlate the date in the permalink with the date in the archive
#
KartikPrabhu
as in if they are archiving do they not allow for date-based querying?
#
kylewm
I am surprised there doens't seem to be any promising leads in the "namecoin" type dns system
#
kylewm
that was overstated
#
kylewm
but I'm surprised that nothing has "caught on" yet
#
aaronpk
I think the point is given a URL that is dead, you don't know anything about it otherwise
#
aaronpk
If there's a date, you know which version of the URL in an archive to retrieve
#
KartikPrabhu
should the date in the permalink change if I later edit the post then?
#
aaronpk
Guess it's more important if the domain is taken over by those squatter pages cause those are sometimes still archived
#
KartikPrabhu
seemingly I can only retrieve the originally published page and not any revisions I made later (in this scheme)
tantek__ joined the channel
#
danlyke
the harder part is that archive.org won't give me stuff because the domain changed hands and they interpret the new robots.txt (or whatever) as not allowing access to the old stuff.
#
tantek__
aaronpk exactly
#
danlyke
(Got this when trying to retrieve Elf Sternberg's "Balkanize Usenet" manifesto from archive.org. I can see that they've got it, but he had it on his old ISP's web server and that domain is now some completely different company)
#
aaronpk
danlyke: yeah that seems weird. I feel like they should stop archiving new content after they find a block by robots.txt
#
kylewm
danlyke: I think that policy may have changed recently? mine suddenly is showing up
#
aaronpk
Can we bug textfiles about that? ;-)
#
kylewm
even though the robots.txt is still there
#
tantek__
Thus when linking to a URL without the date, you must indie-linkwrap it within a local bookmark link that *does* have the date in the URL path
#
danlyke
given that the date portion is primarily useful from the standpoint of the linker (rather than the linkee), embedding that in my document somehow seems more important than trying to convince targets that somehow they should URL-date their documents.
#
KartikPrabhu
note the "tagged" person is not marked on the photo in anyway that I can see
#
kylewm
yeah that's a newish feature on Twitter
jet__ and tantek___ joined the channel
#
KartikPrabhu
I mean't wrt to the person actually being marked on the photo itself
#
tantek___
Perhaps there is a simple regex to look for a date in a URL like \/dddd\/(dd\/dd|ddd)
indie-visitor joined the channel
#
Loqi
Welcome, indie-visitor! Set your nickname by typing /nick yourname