bear.imcreated /mx (+170) "Created page with "{{stub}} <dfn>MX - tells other servers where to send email to for the domain (e.g. mail for example.com is handled by mail.example.com)</dfn> === See Also === - [[dns]]"" (view diff)
bear.imcreated /ns (+189) "Created page with "{{stub}} '''<dfn>a NS (aka NameServer) record in DNS tells the other DNS zone servers which servers are to be considered authoritative for zone updates.</dfn>''' === See Also ==..."" (view diff)
tantekthanks guys - this stuff is way too cryptic for most even dev-types, nevermind others. part of the hell that is configuring independent DNS presence on the web.
creven in "early" days like 96 cablemodem/DSL most people told me the messages never arrived. already all sorts of realtime blackhole/spamtrapped-subnets back ten
fiatjafwhat do you guys think of couchdb? of using couchdb and couchapps to implement some services that everyone interested can run in their own couchdb database?
fiatjafthere are a lot of indie tools for publishing content, I miss one (a really simple one) for receiving content. quick public notes people want to share to you, at your website.
LoqiA comment is a kind of post that is in reply to some other post, that makes little or no sense without reading or at least knowing the context of the source post http://indiewebcamp.com/reply
crmabe you hve a geeky friend with a super-stable server. she sets ACLs so you can control your stuff yet not have to pay off the ICANNSAICVERISIGNDONUTS DNS cartel yourself, or manage a server yourself
crif you carry it with you, network connectivity wll never be an issue to get your stuff. hopefully you rsync/git-push it to something with more reachability/uptime if it's intended to be public
crdid you additionally try their built-in compression features. or try gzipping your files and setting the approprahte HTTP header to inform the other end
cr"I did not want to think about caching recently accessed files in memory and am happy for now to leave that sort of logic up to a smart database. " - what? the kernel does this for files
totthe webserver, while encountering files that will go into a response, compares their mtimes vs the local index timestamp. anything out of date is 'reindexed'.
totso navigating to a new email ends up spewing a few symlinks, one onto a global 'timeline' dir walkable by date, another into the address dir (either to or from )
crive acted in a way that i don't accrue an online "persona". changing nicknames on IRC every day for 20 years. each opensource project contribution under a different pseudonym/email. i dont blog or tweet or have a personal homepage
cri guess this is more in line with the "rave" scene i remmeber from teh early 1990s. it wasnt so much about superstar DJs as more of a communal anonymous experience
tanteksnarfed, it's an interesting question. fundamentally indieweb is about owning your online identity and data (indie = independence / independent) through the use of your own domain (web).
tanteksnarfed, in some ways I like to hope we can keep the barrier to entry to at least the IRC channel fairly low to encourage folks curious about the indieweb to be able to easily explore it, engage in some inquisitive dialog.
tanteksnarfed, in some recent conversations with folks pursuing indie / social web related things, I've made it pretty clear that owning your data / domain is of top importance to us, so however they want to serve it, that's up to them, but certainly we can collaborate on many things, e.g. extracting one's data from silos and making it usable.
snarfedtantek: definitely. another alternative for people like cr might be, they own their content on their own domain, but it's entirely private, and they only posse it out to silo accts that change regularly
thierrymarianne, EOGreer and gRegor` joined the channel
crtrying to read Plusone.google on lynx. press 'A' 3 times to accept cookies, then it says "Bad HTML" and barfs an error. then shows some blank login page
crhaving a somewhat stable reputable identity can definitelyt help w/ antispam. but it neednt be tied to real names. that's jsut marketing-surveillytics crazytalk
benwerd_, snarfed, j12t, snarfed1 and loic_m joined the channel
thierry.marianne.ioedited /projects (+1064) "/* Not sure this summary about the grid belongs to the project page ... please let me know if this is good enough */" (view diff)
j12t, tilgovi, xxcoeurxx, EOGreer, LauraJ, erlehmann, michielbdejong, grantmacken, kensanata, thierrymarianne and eschnou joined the channel
michielbdejong.comcreated /events/2014-10-15-hwc-lisbon/ (+2645) "Created page with "Report: Present: Andrea, Eva, Guiseppo, Michiel, Pierre '''Broadcast hour''' Things we discussed during the first ("broadcast") hour: * what is indieweb * what is webmention,..."" (view diff)
fiatjafShaneHudson, petermolnar: I am developing a pure-couchdb (plus http command-line client) web-analytics platform (simple thing, but highly useful, I think). do you think it will interest anyone?
fiatjafcouchdb databases are also easily moved, copied or continually backed-up using their replication protocol to and for various other backends (http://pouchdb.com/2014/07/25/pouchdb-levels-up.html), including the browser's idb, the filesystem and any leveldb-levelup backend
LoqiWebFinger is a discovery protocol for the web that aims to bring functionality found in the unix finger tool to the web, using email address-like identifiers or fully qualified URLs to represent people http://indiewebcamp.com/WebFinger
fiatjafanyway, my webnalytics couchdb tool isn't going to surprise you, as you are a core couchdb committer, it is just the user accessing your site posting events (pageView, clickSomewhere etc.) info to your couchdb.
tantekalso, we ask people here about their personal website as way of also seeing how much they use their own tools (like your webnalytics couchdb and site posting events) on their own personal site.
fiatjaftantek: I have a blog hosted at github, fiatjaf.github.io (don't judge me). I have some domains also, but I'm working on doing something with them.
LoqiThe Smallest Federated Wiki is an open source project led by Ward Cunningham to produce a wiki that federates using fork history, and was created & launched at the first IndieWebCamp in 2011 in Portland http://indiewebcamp.com/fedwiki
tantekbigbluehat - re: Loqi / the wiki not knowing about those terms - CouchDB, PouchDB etc. - it's an indicator that so far they are irrelevant to the indiewebcamp community
ShaneHudsonbigbluehat: Not that they have no value, but we just don't have time to document everything there is on the web. The going slowly approach seems to work well
Loqireply context examples are real-world examples of reply-contexts on both indieweb sites and silos, how they are presented, their visual design, UI, etc http://indiewebcamp.com/reply-context-examples
bigbluehatit's partly a re-implementation of springpad.com (which I dearly miss) and partly an implemenation of an older idea--which lead me to love springpad.com...while it lasted
reedstrmbigbluehat: some of us are very aware of that(those?) tech, just not yet seen their utility in an indieweb context. Don't need a Ferrari to get to the corner store ...
reedstrmI am surprised tiddler/fedwiki's not here, though. Yet. I looked at those a looong time ago, played around w/a tiddler on a stick for note taking.
LoqiThe Smallest Federated Wiki is an open source project led by Ward Cunningham to produce a wiki that federates using fork history, and was created & launched at the first IndieWebCamp in 2011 in Portland http://indiewebcamp.com/fedwiki
fiatjafben_thatmustbeme, tantek: so why do you use these h-cards? when commenting on each others' blogs you paste your domain so they know who are you?
RichardLitt, squeakytoy and vanderwal joined the channel
ben_thatmustbemefiatjaf, h-cards makes a simple way to identify "Me" if someone replies to my post and they want to pull in my icon and homepage link, they can, likewise if someone comments on my post, I know how to pull in their name / icon / url etc
tantekbigbluehat: re: "it wants to be your web site" - do you know anyone that is using couchdb as their primary content store on their personal site? Perhaps invite them here?
ben_thatmustbemehey tantek, based on the talks of mobile UI and the contact info at IWC, i have started moving type creation for my MP client to its own page each, and once I'm finished I'll have a folder / screen with just MP client buttons.
bigbluehattantek: yeah, the "personal" site question is interesting. I know http://writing.jan.io/ does. jchris.net did (seems his hosting account's expired). and other's I've heard of in the past do
tantekbigbluehat: that would be good to document (which devs have which couchdb sites and not) on the /CouchDB page (or whatever is the canonical capitalization)
finchdtantek: sinxe hte last time you asked me that, i've: bought a domain, paid for hosting... I've yet to relearn HTML form scratch or make any content, so no IndieAuth for me...
LoqiApache CouchDB (aka CouchDB) is a Document Storage Database with a straightforward HTTP API and the ability to do stateless master-master replication over (interruptible) HTTP connections http://indiewebcamp.com/CouchDB
LoqiApache CouchDB (aka CouchDB) is a Document Storage Database with a straightforward HTTP API and the ability to do stateless master-master replication over (interruptible) HTTP connections http://indiewebcamp.com/CouchDB
LoqiPouchDB is an open-source JavaScript database inspired by Apache CouchDB that is designed to run well within the browser http://indiewebcamp.com/PouchDB
tantekbigbluehat: perhaps add a == Selfdogfood == section to /CouchDB noting what you said above about which of its devs are using on their personal sites
ben_thatmustbemei'm going to take it out of my UI i think, just assume it can find it based on the page its given, the base url / homepage (if different) and any rel=me links that are internal to the site
ben_thatmustbemebut i now have my whitelist public so that people can easily use that method to find a link on my site (if its published in that list, not all are)
ben_thatmustbemei've been thinking about going through and backfilling all my old text messaging / AIM history once i get private messaging really worked out
mkoTechnically, I haven't figured out a good way to POSSE my Last.fm activities, so they're still PESOS (though I did post a manual one just to ensure that I could do it without having to PESOS).
mkoThat's an example of what they look like right now. Again, now that I've got tons of data, I'm seeing what I can do to improve the visual design of the listens.
mkoIf you look at the HTML structure, each of the data elements has a p-* with the Musicbrainz ID of the various Tracks and Albums where it's available.
mkoThanks, gRegor`. I wanted a good way of denoting that I didn't want a lot of my posts directly tied to locations, but I liked the idea of being able to segment them as a separate type of "location."
mkoYeah. It annoyed me that Last.fm served images without HTTPS, so I cached the imagery for every track. I need to run a checksum comparison of images to remove duplicate files at some point.
mkoAnd when I imported Foursquare, I imported all of my posted photos, likes, and comments, though the likes and comments haven't been converted into IndieWeb mentions yet.
mkoAnd now I've got a push endpoint on my site, so anytime I check-in on Foursquare through the app, it'll check me in via my site as well. If I check in via my site, it'll check me in via Foursquare and ignore the push response for the checkin (so I don't get duplicate checkins).
mkoben_thatmustbeme: I used to run a startup that was called Notion which was all about democratization of your personal health data. I've got so much more to import that it's not even funny. I have something like 2 TB of photo data that I want to figure out how to expose on my site, around half a million non-Foursquare location data points, millions of biomarkers, and my entire genetic sequence.
mkoYeah. I've already got a few visualizations working locally, but figuring out how to expose them publicly is proving to be a challenge. It's over 30MB of text data. Probably going to have to splice it (no pun intended) up into each chromosome.
Mark87I want to make my pages a bit more machine readable so I can implement some ideas I have. i'm just looking for experiences or opinions to think about as I decide what syntax I want to use
Mark87I can start the conversation though. I have been a quiet fan of microdata for a long time now, what with schema.org, although i have an idea that goes in a slightly different direction than that. But at the same time, microformats have more traction, and really they are more useful than microdata because they are also css hooks
Mark87I recently wondered whether you couldn't replicate the machine-readableness i'm after with link tags with some sort of rel=context attribute like json-ld has
fiatjafI have bad prejudices against rdf or anything using xml, but I like triplestores, and every search for "triplestore" on google will return information about rdf.
danlykeI think the lengths that modern search engines are going through to parse English and understand context shows that if you provide the data, search engines will find ways to use it.
danlykeyeah, the counter to that is that you have to have *an* application in order to justify markup, but I think inclusion of inbound links (even if I'm still skeptical about the notification mechanism) is a good application, and that gives microformats2 a leg up.
tantekin fact, that's a pretty big deal, and a big difference between the way we design things for microformats / indieweb, and a lot of the theoretical academic junk out there
Mark87so you're not just marking up "Here is the author's name", "here is the content of the post", "here is x data". I want to also say "here is a url where you can POST a reply to the post", "here is a url where you can rsvp to the event", etc
Mark87KartikPrabhu__ that's exactly the sort of response i was looking for when i asked my original question. Now someone has to say something like "But you can't do X with microformats but you can with rdfa"
KartikPrabhu__Mark87: one X i have heard to represent arbitrary graph like data, but I haven't needed to do that yet, except to write arbitrary graph like data
Mark87hmm one negative to microformats is the possibility of false positives because developers may use keyword class names in non-microformat webpages. Avoiding this introduces the annoying requirement to prepend keywords with u-/h-/etc prefixes
mkoI've got streaming auto-scrobbling as well as manual syndicated scrobbling. I couldn't figure out how to permalink to a Last.fm syndicated scrobble, though, so I'm not officially syndicating. I'm technically "duplicating"
tantekMark87 - for generic microformats questions / issues (e.g. "possibility") unrelated to any specific practical indieweb usage, feel free to ask in #microformats
danlykeIf the connection to the database was UTF8 and the table was Latin1, the database would have to be coercing 8212(decimal) into an 8 bit number, so it'd choke on that. If the connection was Latin1 and the table was Latin1, you'd get garbage.
danlykeSo does kylewm. And apparently I need to do something h-card is on my blog pages, although I have the "which identity page is this" problem there.
Mark87i'm trying to parse the output of phpmf2 into something useable. Now I've decided the easiest way is to just convert it to RSS which I already have a robust parser for.
danlykeSo <article class="h-entry"><a class="p-author p-name" href="/archives/user.cgi?id=$author_id">$name</a> should be correct? Or do I need an h-card in that <a class="..."> too?
KartikPrabhu__for your case you'd want <article class="h-entry"><div class="p-author h-card"><a href="/archives/user.cgi?id=$author_id">$name</a></div>
KartikPrabhu__snarfed: while you are here, do you know if python functions can be run from other systems? I ask in anticipation of making a python mf2+HTML storage
snarfedwhenever people advocate or deep dive into specific backend tools here, we often say, "that's plumbing, we prefer to focus on end user use cases (and maybe interop) here"
tanteksnarfed - correct, I think I answered this before (so I should make it an FAQ), however when plumbing's unreliability/fragility becomes visible to the UX, then it's worth documenting as a source of problems.
danlykeI'll also note that I ran a short poll after the last discussion on another channel, and got a unanimous "the database-antipatterns page turns me off from that community" from people who actively maintain their own personal web sites.
tantek2) DBA tax = your website sucks more of your otherwise free time (also impacts your UX of using/maintaining your website since it is more demanding for non-content oriented tasks)
KartikPrabhu__the whole point is if you are a seasoned dev then you have no trouble with building a website in the first place! but for "normal" people managing and debugging a DB is a nightmare
gRegor`I have experienced (at work), putting UTF-8 into an ISO-8859-1 encoded HTML file and subsequent garbage text, until I ensured our designers text editors were saving as UTF-8. It may be rare, but it's not hypothetical (to me).
tantekit is MUCH worse for beginners. which is why it's even more ludicrous to see all the people emo-tweeting in reaction to /database-antipattern. it's as if they lack empathy for non-dev users.
KartikPrabhu__snarfed: If I make a python storage thingie with functions that insert and pull entries from a file system will people running PHP be able to use it on their website?
snarfedyeah. until very recently, i still didn't understand that we meant dbs were an antipattern just for personal sites until very recently. i thought the argument was that they were bad, period
snarfedKartikPrabhu__: that's kind of too high level and vague to answer. sure, yes, if your storage thingie is something lang-agnostic, e.g. txt or json or a db (duck)
kylewmKartikPrabhu__: I didn't mean to be snarky. I'm just thinking you'd want to have the python in one process and the php in another ... communicate via HTTP by default, drop down to something like protocol buffers if that turns out not to be fast enough
finchdKartikPrabhu__: probably want a REST API from your python datastore, so that all languages can use it without the interlang stuff. may not be fast enough, but own't know until you've tried. a la github.com/pdxacm/acmapi github.com/pdxacm/acmweb (which aren't in production, BTW)
tantekkylewm - for the "interpreting the output" aspects, we should document any common use-cases with specific algorithms like the /authorship algorithm
danlykeActually, I think a "microformats2 cookbook" would be a good start: If you're publishing a blog entry (Assumed to have an optional title, a user name and profile link, an archive link, and content), use this form. If you're parsing a reply, look for those elements as...
kylewmtemplates for microformatted html *kind of* defeats the purpose... it's really supposed to be possible to tack it on at the end after you have it all written and laid out how you want
danlykekylewm I'm just thinking hierarchy templates. You can put those elements wherever you want, but if you're searching for the content of a reply, the tree looks like /(h-entry)(?:h-\w+)*(e-content)/
danlykethat's what makes parsing such a bear, and publishing equally obtuse, when mostly what we want is a semantic that looks like RSS (maybe with a few extensions, eg: GeoRSS).
danlykeAtom has an author per entry. So, actually, a mapping of microformats to Atom would be a perfect cookbook for people wanting to mf2 enable their tools.
danlykeClearly there's more you can do with mf2, but a mapping to the Atom ontology would make that mostly just a matter of renaming a few CSS classes in any blogging tool.
danlykeYeah, easy way out if you implemented the RSS bits first. I had to go find the Atom spec (once I figured out that I wasn't putting author in my RSS entries) to check, because mostly I read single author RSS, or don't care who the author is (ie: pretty much Lumberjocks.com and HomeRefurbers.com)