#tantekGWG - problem right now with building a real time search engine is that no one would bother using it unless it had Twitter firehose indexed, which is prohibitive (both from a pay-Twitter perspective, and from a build-infrastructure-to-handle perspective)
#snarfedkylewm GWG: if you guys come up with a better bridgy publish algorithm for figuring out what to publish based on p-name/p-summary/e-content etc, please do file an issue, i'll happily implement it
#tantekis still waiting for a more user-centric term for what is "Planet"
#GWGsnarfed: I think the image one is the first one.
#GWGBut, for Bridgy purposes, it could be used for styling
#kylewmGWG: even without an algorithm, I would like a GH issue to track example posts, what you wanted Bridgy publish to do and what it did (assuming that's ok with snarfed)
#tantek.comedited /Twitter (+922) "break down search & notification features into specifics in terms of indieweb features, move client support to a new section" (view diff)
#mkoIt has almost every piece of relevant information that you could want plus all of the potential places that I post, could be contacted, or otherwise reached.
#tanteksnarfed - for your review, in particular the expanded treatment of search and notification in terms of a bit more detail / description of indieweb equivalents
#tantekmany of those details are a first attempt - feel free to disagree/dispute/improve
#mkoI'd love to have that on its own page with better formatting and features around it, but there's no way to expose it to a parser without changing all of my rel=me links on social services to point to the about page of my website instead of my homepag.e
#tantek!tell snarfed for your review, in particular the expanded treatment of search and notification in terms of a bit more detail / description of indieweb equivalents. many of those details are a first attempt - feel free to disagree/dispute/improve http://indiewebcamp.com/Twitter#Features
emmak, wolftune, nagaway, androtest, androte46, androte9 and snarfed joined the channel
#Loqisnarfed: tantek left you a message 10 minutes ago: for your review, in particular the expanded treatment of search and notification in terms of a bit more detail / description of indieweb equivalents. many of those details are a first attempt - feel free to disagree/dispute/improve http://indiewebcamp.com/Twitter#Features
androtest joined the channel
#snarfedtantek: notification part lgtm. search part is a good description of what we'd need to replace twitter on our own site(s), but it's not really a description of twitter's own search, which it probably should be given the section heading
#snarfedmaybe it should go down into the Replacing section, and revert the feature one to the earlier text, with a big disclaimer?
#snarfedyou've worked hard on this page and the intent, so i'd defer to you
chrissaad joined the channel
#tantek!tell snarfed yes we probably need to at least document the question of what would it take to stop using twitter.com for search
#tantek.comedited /Twitter (+174) "/* Replacing */ add point about expanded search for use-cases that you use Twitter search for" (view diff)
#tantekGWG, kylewm, mko - yes u-url for all your profile URLs is sensible. In terms of which matters more, you can always order them (order is preserved when parsing), and use rel=me on them and also use u-uid to what you consider your canonical / unique URL for you.
#beartaxi to airport arrives at 0330 - bed time for me very soon
snarfed joined the channel
#Loqisnarfed: tantek left you a message 14 minutes ago: yes we probably need to at least document the question of what would it take to stop using twitter.com for search
#tantek.comedited /Twitter (+238) "/* Features */ fix contextual description of Twitter Features, expand search in "replacing" section" (view diff)
#tanteksnarfed - thanks for catching that - I was trying to keep too many things in my head and the right thing ended up in the wrong place - take a look now: http://indiewebcamp.com/Twitter#Features
#tanteksnarfed++ for so politely saying I think you got the search description wrong in the Twitter features description. Good example to learn from. :)
#gRegor`My webmention plugin was pretty tightly tied to the blog posts, so I've been working on abstracting it so I can display comments on any URL path, not just an article.
wolftune, sparverius, fmarier and GWG joined the channel
brianloveswords, wolftune, snarfed, cuibonobo and memecake joined the channel
#cuibonobojust found out from the site-deaths page that Svpply is going down… guess i need to grab my data
tantek joined the channel
#cuibonobojust for some perspective: my primary motivation for moving to indieweb is actually because of these more fragile sites: bookmarks, wishlists, pinboards, etc.
#cuibonobothe personal website space is very robust and has a bunch of options to choose from
#cuibonobobut a solid place for a wishlist? not so much
crossdiver and jschweinsberg joined the channel
#pdurbincuibonobo: yeah, I'm into data liberation and avoiding vendor lock-in
#cuibonobopdurbin: do you know of any indie sites that that are being used as a replacement to, say, Evernote?
#cuibonoboi'm having trouble finding examples of indie sites being used for private data
#pdurbinhmm. I just use git for that. a repo only I can reach
#cuibonoboi gotcha. i'm looking to build something that allows my stuff to be searched / sorted / tagged — all the common website stuff, but applied to private data as well
#tantek.comedited /Twitter (+60) "/* Replacing */ order a bit by importance / likely frequency of use" (view diff)
#pdurbincuibonobo: would you build mobile apps too? my wife uses evernote but often from a phone or tablet
#cuibonoboi don't think i have the time or mental bandwidth to grok native device apps, but i *do* subscribe to the app mentality: my website would be an app that consumes my public data API, posting would be through an app that has been registered with the API, etc.
#cuibonobopdurbin: aha. so it needs to be indexed. fair enough. i'm actually between using flat files and then indexing with elasticsearch or using a database for both
#pdurbin"Everyone should own and control their data. So when you make a web site on Branchable, we don't lock you in. You can use git to download the entire source code of your web site at any time, and move it elsewhere if you desire."
#cuibonoboi understand that flat files allow for greater portability, but being able to search them requires building an index (or outsourcing to google) and the link between the files and the index can be very fragile
#aaron_pkflat file storage sounds cooler if you call it NoSQL
#cuibonobopdurbin: well, i feel like there's a trade-off: search solutions that work very robustly with flat files generally can't do faceted search, i.e. only notes, stuff tagged with x
#aaron_pkelasticsearch can index whatever you want from flat files
#aaron_pkI opted for a regular MySQL index for my flat files tho
#pdurbinaaron_pk: interesting. so you could switch your database-cache from mysql to elasticsearch or whatever
#cuibonobopdurbin: sure, but the parser between the flat files and the index is the 'delicate' part i had mentioned before
#pdurbinsure. you'll want a way to re-index everything if need be. blow away the index and re-index it all
#aaron_pkpdurbin: yep. or in my case more likely postgres
#barnabywalterscuibonobo: why is it delicate? in my experience it’s the opposite, because if something goes wrong you can throw away the broken index and just rebuild everything
#aaron_pkyeah I have a script that I used to build my index in the first place, and several times I've already tossed the DB out the window and rebuilt
#aaron_pk(mostly when I was developing my Post class which is responsible for writing the files and updating its index)
#cuibonobowell, i was actually looking to use mongodb as storage. so my process would be reversed: dump the database to flat files
#barnabywalterscuibonobo: ah, you might be confusing flat file storage with having a static site
#cuibonobobarnabywalters: nope. in mongo, each entry is a 'document' that doesn't need a schema, so a flat file storage dumped from mongo would look very similar to the flat files you use for your own site
#aaron_pkbarnabywalters: how did you pick the URL slug for venues? was hoping that would be documented on the /venues page
#barnabywaltersaaron_pk: er, I just ASCIIfy the name of the venue
#cuibonobobarnabywalters: but yes, i was looking to store the posts in mongo and have my site be an app that consumes that data
#pdurbincuibonobo: can you shove any flat file into mongo? does it have to be json or something?
#aaron_pkbarnabywalters: what do you do about conflicts?
#barnabywalterscuibonobo: ah okay. Really the point of the whole storage debate is about the longevity of the *canonical* copy of the data
#barnabywaltersaaron_pk: currently, nothing. In the future, probably adding the locality/country name to the end of the slug
#barnabywaltersalso planning on showing webmentions for the venues, so a checkin will be a note with a rel-venue (or something) to the venue URL
#aaron_pkah ok. was hoping to steal some ideas from you on that :)
#barnabywaltersand then other people can also check in to my venues
#cuibonobopdurbin: it should be json, but other than that there's no schema per se. so your document could very easily be `{"document":"your data here"}`
#barnabywaltersaaron_pk: my current thinking is: ASCIIfy the name, if that already exists add the locality and city, if that already exists add an incrementing number
#barnabywaltersalthough maybe something more sophisticated like a geohash could be used
#cuibonobobarnabywalters: hmm. the 'canonical' copy is something that i've also found confusing regarding the indieweb. this is digital data. there is literally no difference between the original and the copy besides the story in your mind
#bretmeh locks... i wonder if it wold be worth it to have a single writer RCU type thingy over a file system.
#aaron_pkbret: a lot of my stuff is done in a queue so it's already serialized
#barnabywalterscuibonobo: there are huge, practical differences such as naming/addressing, access control, resolution, ease of backing up, ease of restoring from backups, etc
#barnabywaltersbut also the term canonical is used for a couple of different things
#cuibonoboyes, but the data itself is the same, which is why i take issue with 'canonical'
#cuibonoboin my personal case, if i use mongodb for storage and each item has a unique id, i can dump the entire contents of the database to flat files as id001.txt, id002.txt, etc.
#cuibonoboand mongo handles the index for me, so i don't have to think about it
#barnabywalterscuibonobo: and then what are the text files used for?
#cuibonobobarnabywalters: backup of a backup? porting to some other backend? i'm not sure! i was just trying to illustrate that there was no practical difference between what would be in my 'database' and what's in my 'file system'
#barnabywalterscuibonobo: there are huge practical differences like ease of backing up, ease of inspection, ease of searching
#barnabywalterse.g. the copy which gets changed when you make an edit
#barnabywaltersat that point that copy is the canonical, “true” representation of the content, and all the other, non-canonical copies are out-of-date backups or indexes
#cuibonobothere's a point where it's kind of blurry though, like if you post something that goes to a cache to be worked on before ending up in your store.
brianloveswords, frzn, fr0zen and vanderwal joined the channel
#tantekaaronpk: since you're into self-recording a lot of stuff - have you seen http://www.reporter-app.com/ ? (developed by Feltron - what he used for the latest Feltron.com annual report)
#tantek.comedited /Falcon (+436) "/* Working On */ combine the two repost of tweets tasks, move improve reply-context support to second spot" (view diff)
#tantekok that's enough of a braindump for now of things I've been meaning to get to
#tantekin particular since this past weekend's restyling efforts, I've been thinking a lot about in-stream web action buttons (only indieweb example I know of is adactio.com right-aligned links at the end of each post to favorite, retweet etc.)
#tantekhmm… I'm leaning towards use of "like" rather than the generic (but uncommon) "props", or the more strongly worded "favorite" (context: webactions verbs)
#KartikPrabhustill a bit confusing to me to use 'like' and 'favourite' to mean the same thing
#cuibonobotantek: there's other services that name it like the icon, i.e. 'star', 'heart'
#tantekcuibonobo: really? could you provide specific examples? we've been trying to document them on the /like and /favorite pages
#tantekKartikPrabhu: what is the difference to you between like and favorite? (Note that no services use both)
#KartikPrabhu'like' is me saying to the author 'hey I like this' v/s 'favourite' is marking it for myself... more like a bookmark
#cuibonoboif you look at their feeds, hovering over a post makes a giant heart appear, to mark it as a thing you like
#KartikPrabhunaming things after icons would make it more confusing... :P
#tantekgoing to categorize heart as an a variant of like
#cuibonoboKartikPrabhu: yeah i wouldn't advocate for that :)
#KartikPrabhuI have similar confusion about repost vs reshare. I realised that I don't want to 'repost' the entire post from somewhere else... but just reshare it my friends...
#KartikPrabhutantek: yes then... my intent is to not 'repost' but then what is the word for that... Just telling your friends, 'this is a good post. read it' ?
#KartikPrabhucuibonobo: I'm also interested in moving my notes from evernote to my site. Haven't quite figured out the private posts and log in bits yet
scor joined the channel
#cuibonoboKartikPrabhu: yeah. access control is tricky thing.
#cuibonoboi have a bunch of stuff in it. the primary draw for me is 1) just writing a note. no need for titles or anything else. 2) tagging. 3) search. by far the most important feature.
benwerd joined the channel
#pdurbinI've started using Google Keep a bit. It's nice.
erinjo joined the channel
#Loqierinjo: bnvk left you a message 1 week, 5 days ago: yo, just testing out something
#cuibonoboright now i'm in an after-evernote phase where i'm dropping markdown files in a 2014 folder and using a timestamp as the title. i handle search through spotlight, but it's a little wonky
#benwerdwaves from the Reclaim Your Domain hackathon at Pepperdine
#tantekhey erinjo benwerd - how's the weather in LA?
#KartikPrabhucuibonobo: once you have a notes posting on your site, using markdown content should be straight forward, so that is a good first step.
#benwerdtantek: it's overcast, which is frankly disappointing
#benwerdI have a convertible for the weekend, which illustrates some kind of law or other
#KartikPrabhuthe other day I tried to export my 'annotations' from Google Play Books and failed!
#tantekjust updated his software on his site motivated by providing a better markup example to copy / paste into the wiki.
#tantek.comedited /webactions (+3302) "why, how, add explicit markup example for how to add webactions to your posts (hope that helps make more happen!), move some background at the top to a history section, been long enough. and prefer like over props" (view diff)
#tantek!tell barnabywalters not sure if it was you or not complaining about "props" but I now agree and have switched it all to "like" as the generic form, e.g. on http://indiewebcamp.com/webactions
#tantekcuibonobo: apparently I actually documented the rant against "share" a bit more thoroughly on the wiki, with a bit of debate too! http://indiewebcamp.com/share
#tantekeventually I get to "building wiki page support" for my own site: http://indiewebcamp.com/Falcon#Indie_Wiki but until then I don't really have a concrete suggestion to use one wiki software over another. :/
#KartikPrabhutantek: are you sure you want to use those emoji things... not sure they show up in every browser and having a mystery square is not good UI.UX
#tantekKartikPrabhu: I suppose live testing will tell right? ;)
#tantekKartikPrabhu: your post ate a *ton* of heap memory, likely due to the SVG. So SVG still seems like an impractical memory pig. Not going to touch it.
#KartikPrabhustrange since that post has no actual JS running on it except marginalia.js
#tantekKartikPrabhu: use of SVG causes all sorts of additional memory structures to be created for rendering etc.
#tantekbenwerd was that an emoji note POSSEd to Twitter? Or a new chicken post variant?
#KartikPrabhutantek: oh I see... did not know of that problem. in that case my indieweb post must be using monstrous amounts as all the illustrations are SVG too
#benwerdtantek: an emoji note, but I might upgrade my chicken plugin finally ;)
#benwerdDigital lead at a university pointing at a student's own website and domain: "we want this to be the future of the learning management system".
#kylewmseeing the same thing as Fedora 20 on Ubuntu 14.04, fwiw
#tantekbret was it you that asked about why have a URL: [ … ] field / input / UI in notes?
#tantekwhen the URL is already in the address bar?
#tantekanother reason: browser UIs where the address bar is not (easily) accessible, e.g. mobile, or kiosk mode etc., and not having to scan back up to the top of the window for it either.
#KartikPrabhutantek: isn't the 'not having to scroll' not relevant since on desktops the address bar is always present at the top, while in mobile you might have to scroll a little bit but the address is hidden per the first argument... So the second argument does not add anything really?
#breti like it. although most url fields designed for sharing I avoid since its usually a disgustingly long tracking url... definately not the case with your layout though :)
#breti would say that the "share this url:..." pattern can be really great, but often misused/abused by darkpatternists
#tantekKartikPrabhu: the difference is that "scan" literally means what your eyes are doing, the muscle movement of your eyes scanning up - nothing do with with scrolling.