#indiewebcamp 2013-02-05

2013-02-05 UTC
tantek, bnvk and scor joined the channel
#
@znmeb A6 I more or less subscribe to the "Indie Web Camp" philosophy http://indiewebcamp.com/Main_Page #journchat
mxuribe, mxuribe1, bnvk, tantek, tantek_, gardnr, sbp, eschnou, andreypopp, barnabywalters, adactio, Stevef, friedcell, scor and raphdg joined the channel
#
@edsu Open, Social, Academic Bookmarking: Save to AppDotNet http://hublog.hubmed.org/archives/001965.html via @invisiblecomma #indieweb
texburgher and tantek joined the channel
#
tantek fascinating, about.me bought itself out of AOL. apparently there's a future in identity hosting?
#
tantek http://blog.about.me/2013/02/05/about-me-goes-independent/
#
tantek which would seem to imply that there may even be a future in commercially enabling IndieWeb identities as well.
#
tantek as an upgrade/upsell version of about.me
#
tantek like when you're ready to graduate from about.me to your own URL
#
tommorris interesting
eschnou joined the channel
#
@t Congrats about.me! http://blog.about.me/2013/02/05/about-me-goes-independent/ If "personal identity" can be a business, perhaps #indieweb identity can too. (ttk.me t4NM2)
barnabywalters joined the channel
#
@dshaw RT @t: Congrats about.me! http://blog.about.me/2013/02/05/about-me-goes-independent/ If "personal identity" can be a business, perhaps #indieweb identity can too. (ttk.me t4NM2)
zztw joined the channel
#
dpk http://kottke.org/13/02/the-home-office-of-the-21st-century
#
dpk err. wrong window. but an awesome link nonetheless
eschnou and andreypopp joined the channel
#
zztw I want to talk about how content addressability can fit into the discussion presented here: https://unhosted.org/adventures/8/Collecting-and-organizing-your-data.html (via g+ http://kiq.me/I3f )
#
zztw anyone want to do a g+ hangout friday evening or saturday? comment here: https://plus.google.com/101533265029269684907/posts/YYCf6kAcjuq
#
barnabywalters zztw: evening in what timezone?
#
zztw or /msg me or mention me or whatever else
#
zztw pacific time
#
barnabywalters ah, probably way too late at night for me then
#
barnabywalters I like the unohsted link, I wasn’t aware of that guide
#
zztw sat morning? what's good for you?
#
barnabywalters I am at an event this weekend but for people in the UK, from 9:00 to 11:00 PDT is generally quite comfortable
#
zztw that's AM?
#
barnabywalters zztw: I’m posting that link as a bookmark, what’s your personal URL I can /via?
#
barnabywalters yep
#
zztw http://traviswellman.com/
#
barnabywalters cheers
#
zztw :)
#
barnabywalters so, if identengine.com is working properly, I should get a nice h-card if I use your @namw
#
@BarnabyWalters Unhosted Adventures — a great guide to setting up your own #indieweb server /via @wavis (http://waterpigs.co.uk/notes/916/)
#
barnabywalters or not. oh well, manual it is
#
zztw I broke it... the hosting situation is disfigured temporarily
#
barnabywalters if we’re going to do a regular hangout type thing, we should probably use something like jitsi (I think they do video conferencing)
#
tommorris so, this afternoon, I fell off the wagon on posting to my own site. just because of speed.
#
tommorris tweeting is a lot, lot quicker than posting to my own site at the moment.
#
tommorris got to make the process of posting to my own site ridiculously fast.
#
barnabywalters tommorris: yeah, as is getting from a tweet to an indieweb reply dialog, even with indieweb reply
#
barnabywalters my UI is actually quite speedy, but I certainly need to parallelize all the various requests to services I’m making
#
Loqi PARALLELIZE ALL THE VARIOUS http://loqi.me/6QA
#
tommorris well, now I have it on my own server, I can use a job queue
#
tommorris rather than write to the DB, just write to a job queue which can feed the DB in its own sweet time
#
barnabywalters is scared of job queues for some reason
#
tommorris so, I've sort of evolved through a variety of them
#
tommorris I started using Ruby DRb queues, which are a giant hack, to do inter-process communication between a Ruby and JRuby process
#
tommorris then more recently, I tried out Beanstalk'd, which is pretty lightweight but also pretty rliable
#
zztw tommorris: why is writing to a db slow? should never be slow...
#
tommorris well, it may not be writing to the DB that's slow but the HTTP requests
#
tommorris I need to do some profiling
elioat joined the channel
#
barnabywalters tommorris: as in, the POST request to create the note, or are you doing auto-POSSE now?
#
tommorris Auto-POSSE
#
barnabywalters cool — yeah, twitter’s API is variable, and FBs is hideously slow
#
zztw barnabywalters: i expected a not-google-plus sentiment. I totally agree we should have something like that but for now I haven't experienced the convenience of a g+ hangout anywhere else.
#
barnabywalters I was auto-posseing to FB for a while, but now use their feed dialog as a web action delegate. I should probably document this
#
zztw if you have a server to do it on, i'm down
#
barnabywalters zztw: yeah, G+ hangouts are nice. I haven’t tried out jitsi conferences but it’s gotta be worth a try
#
barnabywalters thinks jitsi is peer-to-peer, so no server required
#
zztw barnabywalters: let's do a g+ hangout to collaborate on setting up jitsi!
#
barnabywalters zztw: awesome idea :)
#
barnabywalters ah, looks like jitsi only does conference *audio* at the mo
josephboyle joined the channel
#
barnabywalters hm, I think right now parallelizing requests is low-hanging fruit compared to setting up a job queue. I can always resort to that if need be in the future
#
tommorris well, if you've got a job queue, you can parallelize it by pulling stuff from the job queue in parallel
#
barnabywalters yeah, I can see the benefits, but right now I’d rather build, say, self-hosted venues and sheet music and make my note-posting a bit faster, than spend a load of time re-developing to make my note posting SUPER FAST :)
#
barnabywalters hence researching migrating from mysql to pgsql
#
tommorris barnabywalters: postgres stood up for the 150,000 hits or so yesterday. ;)
#
barnabywalters it looks awesome, esp. the GIS stuff and arrays (which is my immediate reason for migration)
#
aaronpk is considering psql now too. currently has no DB backend
#
barnabywalters still a scary process though :)
#
tommorris be careful, admin is harder. scaling is a bit harder.
#
tommorris but well worth it.
#
barnabywalters not sure I’ll have to worry about scaling my little site!
#
zztw has never considered anything other than postgresql
#
barnabywalters has successfully installed pgsql and created a user+database. yay!
#
Loqi woot
#
barnabywalters that’s enough scary unix stuff for tonight. goodnight all
#
Loqi don't let the bed bugs bite
Stevef and tantek joined the channel
#
tantek btw ICYMI: http://glennjones.net/2013/01/brand-new-microformats-2-parser/
#
tantek this has huge implications for improving support for use of microformats' IndieWeb use cases
#
tantek e.g. all the hAtom / h-entry stuff
#
tantek I think we can now finally start transitioning the last bits of dependencies on XML/Atom etc. to hAtom / h-entry
#
tommorris now that Ruby has 'execjs', which lets Ruby call JS (using, say, Node), it wouldn't be too hard to use a microformats-2 parser as part of a test suite
#
tantek that would be great
#
tantek how much work would it be to write simple Ruby wrapper library around Glenn's parser?
#
tantek so we can start parsing each others' h-cards etc.
#
tommorris not sure, I haven't really looked into execjs, beyond basic security auditing
#
aaronpk heh, I don't know if I'd want javascript in my ruby
#
tommorris well, the alternative is writing a microformats-2 parser in Ruby
#
tantek or port either the JS one or PHP one to Ruby :)
#
tommorris which shouldn't be too hard since we have a pretty good spec, test suite and some open source implementations
#
tommorris hell, I've written large chunks of an RDF/XML parser. everything should be simpler compared to that.
#
tantek *much* simpler
#
tantek since folks around here seem to be quite good at real-world / practical solutions to (semantic) markup questions, I wanted to bounce a few things off the channel
#
tantek in particular regarding uses for h-entry
#
tantek which seems to be the strongest semantic workhorse for numerous indieweb use-cases
#
tantek so we're all marking up permalink post pages, stream pages (e.g. home page), and archive pages with "h-entry"
#
tantek I'm thinking that *any* web page that has a name and a notion of published (created) and/or updated (modified) date is worthy of marking up with h-entry
#
tantek e.g. W3C specs
#
tantek what do folks here think?
#
tommorris well, not sure about that. I think that the point of it is to identify the atomic unit of content.
#
tommorris the problem then is that the semantics of an h-entry just collapses into that of being an HTML document
#
tommorris which is fine. I'm all for that kind of thing on a philosophical level, but there's pragmatic concerns
#
tommorris people are likely to use the data *for* some purpose. subscribing to a blog, subscribing to some kind of stream of blobs of content (even if that's not a blog
#
tommorris a list of things you've faved, or friends you've added)... if everything becomes an h-entry, I'm not sure how that would work out
#
tommorris it reminds me a bit of the 303 debate in Semantic Web land, where everyone had grand theological ideas about what a "resource" is. but it all sort of collapsed in the end when people put their theologies aside and just looked at how people would actually retrieve data
#
tommorris tantek: in short, yes, but I'd want to see the practical consequences. ;)
#
tantek well, all the usual goodness in being able to automatically know the datetimes of a document
#
tommorris how would it affect pulling stuff out of feeds.
#
tommorris let's say I've got a page, tommorris.org - which has 20 odd h-entries. there is then a further "root" h-entry, the document itself
#
tommorris you then have a weird relationship between the main h-entry and the subsidiary h-entries
zztw, friedcell and donpdonp joined the channel
#
tantek tommorris - in your example, tommorris.org has no notion of it's own published/updated date
#
tommorris some blogs do. I know Dave Winer's Scripting.com used to have a "last updated" field
#
tantek it's current content is literally just a container for h-entrys and a header/footer with some meta info :)
#
tantek sure, but we're talking about your example
#
tommorris ah okay, that's reasonable then
#
tantek ok, I'm going to try adding h-entry to one of my specs and then you can take a look at a live example
#
tommorris is a bit ditzy this evening after some Polish vodka and spending 6+ hours watching Parliamentarians discussing whether he can get married. ;)
#
tantek well, get married to someone of your preferred gender/orientation :)
#
tantek bison vodka?
#
dpk ahahaha:
#
dpk .tw 298917256677691392
#
dpk err
#
dpk wrong window again
#
Loqi hahahaha
#
tantek lol
#
tantek ;)
#
tommorris tantek: Universum. @hadleybeeman bought it for me a while back.
#
tantek hopefully you were toasting to good outcomes
#
tommorris no, drinking once Peter Bone started talking
#
tantek saw the bingo board
friedcell joined the channel
#
tantek ok, here's h-entry markup in an actual spec, take a look at the output from Glenn's parser: http://microformat2-node.jit.su/?url=http%3A%2F%2Fdev.w3.org%2Fcsswg%2Fcss3-ui%2F&callback=&filters=
tantek joined the channel
#
hadleybeeman Glad to hear that vodka went to good use, tommorris. :) And cheers... well worth celebrating.