#indiewebcamp 2013-10-17

2013-10-17 UTC
acegiak, andreypopp, bnvk, pfenwick and tpinto joined the channel
caseorganic joined the channel
#
acegiak
well this in an hillarious fail somewhere along the commenting flow
caseorga_ joined the channel
#
aaronpk
not sure how that happened... the php parser seems to handle it pretty well http://pin13.net/85C
tpinto joined the channel
#
acegiak
aaronpk: yeah, super weird
caseorganic, julien51 and tantek joined the channel
#
tantek
aaronpk - that is pretty funny re: rss/postgres
tpinto joined the channel
#
neuro`
Hello
josephboyle joined the channel
#
aaronpk
just added streaming API to webmention.io like I have on my blog :) http://webmention.io/#streaming-api
julien51, tpinto, glennjones, LauraJ, caseorganic and pfenwick joined the channel
#
@jihaisse
@fdevillamil tu va implémenter les #webmention ? ou pas du tout ? #indieweb
(twitter.com/_/status/390737667891032065)
Jihaisse joined the channel
#
Jihaisse
Hello there
josephboyle, pfenwick, eschnou, tpinto, glennjones, ShishKabab, lmjabreu and andreypopp joined the channel
eschnou, barnabywalters and melvster joined the channel
bnvk and andreypopp joined the channel
#
@ndebock
@clemnt sur le web if faut être marketmaker et non marketplace car les marketplaces seront remplacées par l'indieweb
(twitter.com/_/status/390810789285621760)
#
tommorris.org
edited /projects (+117) "/* WordPress */ adding P2"
(view diff)
LauraJ, bnvk, _6a68 and texburgher joined the channel
#
barnabywalters
acegiak: RE comment parsing: woah yeah that is way off!
#
barnabywalters
looks like it’s a result of my parser converting the hentry classname on the article element into a mf2 h-entry classname even though it’s nested inside a h-entry
#
barnabywalters
so it looks like an h-entry nested inside another, and the outer one’s name property is the textContent of the whole element!
#
barnabywalters
I need to make the classic microformat conversion support much stricter — if *any* mf2 classnames are found in a document, don’t convert *any* classic mf classnames
_6a68, LauraJ, adactio, bnvk and tantek joined the channel
#
tantek
barnabywalters: I think your parser is doing the right thing
#
tantek
The error may be in the markup
#
tantek
If the hentry and h-entry represent the same thing, they should be on the same element.
#
barnabywalters
tantek: such markup errors are common enough to cause this problem frequently enough that I want to change the behaviour
#
barnabywalters
it’s happened at least three times now
#
tantek
Otherwise it is very odd to have both
#
tantek
barnabywalters: That's a validation problem then
#
barnabywalters
tantek: I think the approach you’ve advocated in talks before may be causing some confusion
#
barnabywalters
IIRC you said “use a classic microformat classname on the <body>, then use mf2 for everything else”
#
tantek
A validate could indicate when you have uf2 markup and also old uf1 markup without uf2 markup on those elements
#
barnabywalters
yep, that’d be a useful thing to have
#
tantek
barnabywalters: Yes, just *a* classic microformat. Then use uf2 for that *and* everything else.
#
tantek
That "and" is crucial
#
tantek
Your suggestion of ignoring all classic uf markup on a page rather than element level is interesting.
#
barnabywalters
tantek: yeah, I’m aware of that, but I’ve seen several cases now of people putting just a classic mf classname on the body, and using mf2 elsewhere
#
barnabywalters
so we need to clarify how to use mf and mf2 at the same time
#
barnabywalters
seeing as it causes regular issues
#
tantek
Have you documented the permalinks of the "three times" the error has occurred?
#
barnabywalters
tantek: nope, I should be able to find them though
#
tantek
I've never seen it before
#
barnabywalters
unless they’ve been fixed
#
tantek
Well document the permalinks and when you saw them (that's what dt-accessed is for in h-cite)
#
tantek
I think better examples for putting microformats on body would help too that showed both.
#
barnabywalters
I think that one got fixed
#
barnabywalters
this is another which taproot parsed weirdly, I think it was for the same reason: http://notizblog.org/2013/06/18/the-rise-of-the-indieweb/
#
tantek
Are those both Wordpress?
#
barnabywalters
yep, and I think the acegiak’s problem comment might have been caused by an outdated version of the same plugin pfefferle is using
#
barnabywalters
either that or mixture of classic mf in-theme and mf2 being added by the plugin
#
tantek
So that sounds like user/install error - I mean of course it's possible to screw up a Wordpress install with plugins or inconsistent plugins. We're not going to solve that.
#
tantek
And those have nothing to do with body either
#
barnabywalters
tantek: I’ve seen several cases where ignoring classic mf classnames if any mf2 are found would benefit, none where parsing would suffer
#
barnabywalters
tantek: they did when the problems were caused, they’ve just been fixed since
#
tantek
I think it's a bad precedent to set to ignore previous markup.
#
tantek
This is a Wordpress(even plugin) specific problem it sounds like
#
tantek
Not a hand authoring error which you made it sound like when you mentioned the advice re body
#
barnabywalters
yeah, looking at it more closely it seems like a clash between existing markup in themes and injected markup with a plugin
#
tantek
So it's actually *good* that the errors were revealed so quickly in the Wordpress setups.
#
tantek
So that's a theme/plugin conflict problem.
#
tantek
That's a separate/larger problem that does need addressing.
ozten and caseorganic joined the channel
#
tantek
However we shouldn't change the parsing model just because of oddities on one CMS platform.
squeakytoy, tantek, LauraJ, josephboyle, bnvk_, bnvk, benprew, andreypopp, jschweinsberg and _6a68 joined the channel
#
@kevinmarks
“And then I made a resolution: I would give myself time, every day, to build my own things again.” - @benwerd http://werd.io/entry/5243114abed7de2970c0b2b7/how-xoxofest-and-indiewebcamp-saved-me-in-a-way
(twitter.com/_/status/390884589310201856)
bnvk, spinnerin and tantek joined the channel
#
tantek
barnabywalters - could you add an issue to the top of http://microformats.org/wiki/microformats2-parsing-issues re: sites sometimes getting uf+uf2 combined markup wrong?
#
tantek
and add citations for the examples you mentioned
#
tantek
let's at least collect this issue and our reasoning about it
#
barnabywalters
indeed — doing
#
tantek
in case we get more/less data in the past
#
tantek
great - thanks!
spinnerin and realzies joined the channel
#
aaronpk
!tell shaners just got my flight for Farmhouse conf weekend!
#
Loqi
Ok, I'll tell them that when I see them next
BjornW, jernst_, brianloveswords, bnvk and wyomingplease joined the channel
#
wyomingplease
hi folks, earlier this morning I wrote the first blog post on my new blog, documenting my personal cloud stack. You can check it out here: http://petar.petrovic.io/2013/10/17/how-i-built-my-own-personal-cloud/
#
wyomingplease
I wondered that it may be a good candidate for addition to the Posts about Indie Web page at indiewebcamp.com
#
aaronpk
wyomingplease: cool!
#
wyomingplease
aaronpk: thanks, this is just the first phase
#
aaronpk
you should add some h-entry markup to the post! http://microformats.org/wiki/h-entry
#
Loqi
definitely
#
wyomingplease
aaronpk: yeah, I've thought about that, but since I'm just getting started, it will get better over time
#
tantek
wyomingplease - nice post!
#
tantek
I agree with not liking the term cloud in general - it seems to have been abused mostly by marketing/sales people
#
wyomingplease
tantek: thanks, Indie Web made me start blogging again :)
#
tantek
typically "in the cloud" can be replaced by "on the server"
#
wyomingplease
it's pure marketing in my opinion
#
tantek
wyomingplease - that's great to hear! and good on you for starting blogging again :)
#
tantek
there's a lot to be said about thoughtful blog posts
#
tantek
so, the term "cloud" originally had *some* meaning beyond just "server" - it came from "elastic cloud"
#
wyomingplease
tantek: I just want to build stronger digital identity online and get involved more profoundly in this indie web thing :)
#
tantek
definitely
#
tantek
something is not really "cloud" unless it is elastic in *some way* (either storage, bandwidth, CPU etc.)
#
wyomingplease
elasticity often means different things to different people
#
tantek
that elasticity is what distinguishes a "cloud service" from just a "web/internet (hosting) service"
#
wyomingplease
it depends on whether it is automatically elastic by design or you have to look after it and expand it on your behalf
#
tantek
I disagree - there needs to be *some* semit-automatic elasticity
#
tantek
"look after it and expand it on your behalf" is just traditional hosting services where you can switch plans
#
wyomingplease
well you're right, but still people call things "cloud" even if they are not automatically expandable
#
tantek
well two things: 1) don't call things cloud that are not at least semi-automatically expandable (and contractable)
tpinto joined the channel
#
tantek
and 2) call people out when they use "cloud" to mean "server"
#
tantek
as, in you mean "personal server", right? ;)
#
wyomingplease
that's exactly what I'm doing, but some people just don't want to bother
#
wyomingplease
a server can be for any purpose :)
#
tantek
of course
#
tantek
just suggesting that "personal server" is a better term for what you're doing (unless I missed something in your post)
#
wyomingplease
yeah, it's completely for personal purposes
#
wyomingplease
at least for now
#
wyomingplease
but I like it this way
#
wyomingplease
VPSes are damn cheap these days...
#
wyomingplease
and they're expandable so you can scale any time :)
#
tantek
are you paying a flat monthly fee? or paying per usage (bandwidth, CPU, storage) ?
#
wyomingplease
I pay flat fee, I like to predict my costs
#
tantek
right, so you have a personal server, not personal cloud :)
#
wyomingplease
yeah, but I used a little of marketing mumbo jumbo to better explain what I'm doing :)
#
tantek
don't give into the marketing mumbo jumbo - it only clouds the issue ;)
#
tantek
just adds to the confusion - doesn't better explain (maybe better promotes, but then that's marketing)
#
wyomingplease
right, but the general idea is that once you set everything up, you don't really need to think of it as a single server instance, since you've got a bunch of stuff on it and you just use it as any real cloud service
#
tantek
just use it as any other real server service
#
tantek
even "real cloud service" like Apple's iCloud is not actually *cloud*
#
tantek
that's the point
#
wyomingplease
well nothing is actually "cloud", but people still use the term a lot
#
tantek
no that's not true - if you pay per usage rather than flat fee per time, then that is certainly "cloud"
#
wyomingplease
cloud is just a glorified business model with heavy virtualization under the hood
#
tantek
it's that elasticity of paying only for the resources you use that makes it "cloud"
#
tantek
it's not just a business model. that flexibility provides end user benefit.
#
wyomingplease
sure, but I use the term cloud exclusively for the sake of clarity, I'm not really an expert on the marketing terminology
#
tantek
it doesn't help clarity, that's the point
#
tantek
you've got a virtual server so just say "server"
#
tantek
for the sake of clarity
#
wyomingplease
we agree that we disagree :)
#
tantek
you prefer to use "cloud" as a synonym of "virtual server"?
#
tantek
or is there some other meaning you are ascribing to "cloud" when you say it?
#
wyomingplease
I just think of it as a set of services where you don't really have to think about where the data actually goes, you just know that it's going somewhere where you can control it, or not if you're using a mainstream service, that's all
#
wyomingplease
I don't go deeper than that so I don't think of elasticity and stuff
#
tantek
"you don't really have to think about where the data actually goes" == virtual
#
tantek
ah, found a decent NIST defn of cloud computing essentials
#
jernst_
for me, "cloud" = "don't need to worry"
#
wyomingplease
yeah, once you set everything up the way you like it, you don't have to constantly look after it, a little maintenance from time to time is sufficient
#
wyomingplease
unless you're hosting Gmail or something
#
tantek
so if you have storage or bandwidth caps on your virtual server provider, then yes, you do have to worry - about termination of service when you exceed those caps
#
jernst_
well, worry-free doesn't exist. if not caps, then metered $, so i might have to worry more
#
wyomingplease
they don't terminate my account, they charge me over my base fee for exceeding usage
#
tantek
wyomingplease - ok so then you do have *some* elasticity
#
wyomingplease
tantek: technically, yes
#
tantek
I didn't say terminate account - I said terminate *service* - big difference
#
wyomingplease
I thought service too, my mistake
#
tantek
jernst - the NIST definition here (scroll a bit) is quite good: http://en.wikipedia.org/wiki/Cloud_computing#Characteristics
#
tantek
where it says "The National Institute of Standards and Technology's definition of cloud computing identifies "five essential characteristics":"
#
jernst_
I know it and I don't disagree, just words have a way of evolving
#
tantek
so for example, my virtual server has a hard monthly bandwidth transfer cap, if/when exceeded, my server returns an error page
#
tantek
not the greatest solution (cheap though), and not "cloud"
#
tantek
jernst - there's no utility in calling the same thing by a new name except for marketing/sales people that want to charge more $ for the same thing.
#
tantek
as technical people we shouldn't buy into that kind of bulshytt ( http://indiewebcamp.com/bulshytt )
#
jernst_
well, I have long time ago given up attempting to freeze any definition of anything in normal language use, whether the sales guys stretch it or anybody else
#
jernst_
everybody is guilty of it in one way or another, (not just about cloud, but any term)
#
tantek
jernst - it's not about freezing (as yes, definitions do evolve), it's about avoiding needless duplication
#
jernst_
I own a great book about the roots of English words, and some of them have a rather mind boggling evolution behind them
#
tantek
the evolution is not a problem and has no bearing on the issue of duplication
#
jernst_
I rather focus on what the principles are instead of whether this word or that is "correctly" used
#
tantek
jernst - when people misuse words, it either indicates lack of knowledge or intentional deception, both of which are good to be aware of
#
jernst_
or literary talent, or probably a number of other things
#
tantek
or if you prefer your message in t-shirt form: http://stml.spreadshirt.co.uk/the-cloud-is-a-lie-men-A16241807
#
jernst_
I'm all in favor of precision when precision is warranted … but it's not my #1 goal in itself
#
tantek
jernst - it's more about clarity than precision.
caseorganic, HereticLocke, caseorga_, ozten and LauraJ joined the channel
#
tantek.com
created /gmail (+19) "r"
(view diff)
#
tantek.com
created /Gmail (+152) "stub with a simple definition"
(view diff)
#
tantek.com
edited /Google (+80) "add Gmail"
(view diff)
#
tantek.com
created /blogging (+284) "stub"
(view diff)
#
sparverius
does iwc have any SF beer-ups coming up
#
tantek.com
created /cloud (+3048) "stub with essential NIST definition, misuses, articles, see also"
(view diff)
#
tantek
sparverius - I don't see any on http://indiewebcamp.com/events but you could propose one - all it takes is two people agreeing to host it and posting the details on a wiki page.
caseorganic joined the channel
#
tantek
those of you sending webmentions from your site semi-automatically (i.e. without a manual curl) - do you do it synchronously in your posting UI? or do you do it asynchronously? either way, do you report webmention results back to your posting UI?
#
tantek
aaronpk - were you POSSEing replies when you posted this? http://aaronparecki.com/replies/2013/04/19/2/indieweb
#
tantek
!tell barnabywalters were you POSSEing replies when you posted this? http://waterpigs.co.uk/notes/1334/
#
Loqi
Ok, I'll tell him that when I see him next
#
tantek
!tell eschnou did you POSSE https://eschnou.com/entry/testing-indieweb-federation-with-waterpigscouk-aaronpareckicom-and--62-24908.html to Twitter? if so what was the tweet permalink? and are you displaying/linking-to rel=syndication POSSE copies now on your posts?
#
Loqi
Ok, I'll tell them that when I see them next
#
tantek
aaronpk - http://webmention.io/#streaming-api looks cool but I'm not sure how to use it yet. Minor nit - typo at end: ">/script>" should be "</script>"
#
tantek
It looks like the webmention spec at webmention.org is quite out of date in terms of resolved issues, rel=webmention etc. So much so that I think we need to point it out on the wiki page.
abrereton, tpinto and abrereton1 joined the channel
#
tantek.com
edited /webmention (+605) "supersedes Pingback, unlink .org from top since it is out of date, no idea when it will be updated, update curl to use rel=webmention, note grep caveats"
(view diff)
#
aaronpk
tantek: yes I believe so. as soon as I started POSSEing notes I did the same for replies
#
aaronpk
that one didn't go to twitter because it didn't find a tweet from eschnou
#
tantek
aaronpk - ah that was my question - as I can't seem to find a tweet from eschnou linked from anywhere
#
tantek
though I POSSE my replies even if there is no tweet
#
aaronpk
oh really? why's that?
#
aaronpk
to me it doesn't make sense to syndicate a reply to twitter unless the thing i'm replying to is also on twitter
#
tantek
e.g. my recent blog post was really a long reply to the Guardian article
#
tantek
which I obviously preferred to post on my site than in their comments form
#
tantek
is reviewing webmention discovery parsing libraries and is a bit horrified.
ozten and barnabywalters joined the channel
#
Loqi
barnabywalters: tantek left you a message 41 minutes ago: were you POSSEing replies when you posted this? http://waterpigs.co.uk/notes/1334/
#
aaronpk
horrified? heh
#
tantek
aaronpk - yeah, I'm considering just using phpmf2 to parse for the rel=webmention value
#
aaronpk
oh yeah, it does a good job of pulling out rel values
#
barnabywalters
checks logs
#
tantek
rather than the grep/regex scraping going on in all the libs
#
aaronpk
i think all the libraries are trying to avoid adding dependencies
#
Loqi
THINK ALL THE LIBRARIES http://loqi.me/85p
#
aaronpk
ALL OF THEM LOQI
#
barnabywalters
tantek: looks like I was, as the previous note (also a reply) was POSSEd: http://waterpigs.co.uk/notes/1331/
caseorganic joined the channel
#
barnabywalters
although that was a reply to a tweet
#
barnabywalters
I certainly hadn’t figured out original post discovery then
#
tantek
barnabywalters - ok, so 1331 was POSSEd but not 1334. Interesting. Why some and not others?
#
barnabywalters
RE sync/async webmention sending: I actually cheat and have an almost hybrid approach, documented here: http://waterpigs.co.uk/notes/4R_M2J/
#
barnabywalters
tantek: no idea — it’s difficult to justify stuff like this in hindsight ;)
#
barnabywalters
and there were some bugs in my twitter POSSE code for a while which prevented POSSE if the in-reply-to URL was not a twitter.com URL
#
tantek
barnabywalters - not looking for justification. more the feeling/thinking behind it.
#
barnabywalters
I remember thinking it would be weird for people to see my reply tweets to non-tweet content, as they’d always have to click through to get context
#
tantek
barnabywalters - see my blog post reply example above
#
barnabywalters
especially if the replied-to person had no twitter handle — then it just looks like a completely random tweet
#
tantek
well, maybe. some reply tweets are self-standing.
#
barnabywalters
tantek: indeed, and since ditching twitter as a place to post stuff the way I reply to stuff has changed completely
#
tantek
right
#
barnabywalters
fewer mindless “+1” style replies, more questions, longer responses
#
tantek
more thoughtful then
#
barnabywalters
yes. replies now have to be good enough for them to go alongside all my other notes, as I have no separate reply stream
#
tantek
is that good (more thoughtful, better content), or bad (higher barrier to expression) ?
#
barnabywalters
good for all the above reasons. the only thing this approach misses out on is small social interactions and positive reinforcement
#
barnabywalters
but there are other channels which are (at the moment) more effective for that — favoriting/liking, talking in IRC or IM
#
barnabywalters
or, heck, in person
#
tantek
right, I think that makes sense
#
tantek
interesting tradeoffs
#
barnabywalters
is writing a series of blog posts about this stuff at the moment
#
barnabywalters
as 2013 is my 10 year web publishing anniversary
#
tantek
interesting noting what post types accept webmentions and what don't
#
tantek
barnabywalters - is the *only* way you provide webmention discovery through HTTP link header?
#
barnabywalters
tantek: at the moment, yes. I need to fix that
#
tantek
I was curious if that's considered "to spec" or not - i.e. are all discoverers required to check the HTTP link header for webmention
#
barnabywalters
hm, interesting — the spec lists both in the example but doesn’t explicitly require either
#
tantek
plus the spec is still horribly out of date
#
aaronpk
tantek: all the new stuff is in 0.2, and sandeep wanted to explicitly version it rather than just change what's there
#
tantek
I thought we had this discussion - that's a horrible way to do a spec
#
tantek
especially when you put the domain in a rel value like that
#
aaronpk
I believe there was a discussion, with no resolution
#
tantek
top level should always have the latest
#
tantek
we've already learned this lesson many times over with W3C, WHATWG, microformats etc.
#
barnabywalters
tantek: yeah, sandeep didn’t want to do that for some reason
#
tantek
well if he wants to choose to have his URL become out of date that's his decision, but we need to note as much
squeakytoy joined the channel
#
tantek
(insert point about wikis being better for specs than github for this reason as well)
#
aaronpk
in the sense that wikis are collaborative, but github requires a small set of people to basically approve changes?
lukebrooker joined the channel
#
tantek
aaronpk - yeah, in practice the latter is more fragile
#
acegiak
morning all
#
acegiak
barnabywalters: yeah I see what you mean about the markup
#
barnabywalters
morning acegiak
#
barnabywalters
are you using pfefferle’s “put mf2 markup in wordpress” plugin?
#
acegiak
no, I'm using his sempress theme
#
barnabywalters
we think it might be that, clashing with microformats already in wordpress themes, which is causing the problems
#
barnabywalters
ah, interesting
#
acegiak
but I see the issue in the markup like you've said
#
tantek.com
edited /webmention (+1742) "/* Implementations */ split into IndieWeb vs. library sections, add IndieWeb implementations in rough chronological order from memory, leave ??? where specific dates unknown"
(view diff)
#
tantek
given that webmention.org is now horribly out of date, I'm going to start treating indiewebcamp.com/webmention as the spec
#
tantek
since we're able to actually keep that up to date, and resolve conflicts with subsequent edits when needed
#
tantek
at some point we may want to just move the reference to webmention.org to be a historical 0.1 spec, rather than anything "live"
#
barnabywalters
!tell pfefferle looks like there might be a bug with sempress/wordpress mf2 plugin causing mf2 classnames and classic mf fallback classnames to be on different elements, documented here: http://microformats.org/wiki/microformats2-parsing-issues#mixture_of_microformats2_and_classic_microformats_classnames_on_different_elements — any ideas?
#
Loqi
Ok, I'll tell them that when I see them next
#
tantek
alright, in the process of figuring out what I need to implement to automatically discover webmention endpoints in practice, I've started documenting some of your implementations http://indiewebcamp.com/webmention#IndieWeb_implementations (cc: aaronpk barnabywalters benwerd adactio)
#
tantek
feel free to update the entry for your site/project with more details regarding your webmention implementation
#
tantek
!tell benwerd when did you implement sending (and receiving) webmentions? also, what do you think of adding rel=webmention in addition or instead of the .org rel in idno?
#
Loqi
Ok, I'll tell them that when I see them next
#
waterpigs.co.uk
edited /webmention (+132) "/* IndieWeb implementations */"
(view diff)
caseorga_ joined the channel
#
tantek
barnabywalters do you yourself use phpmf2 to discover rel=webmention endpoints?
#
tantek
that's the implementation path I'm considering currently - I can't stomach all the grep/regex scraping
#
peat
sighs.
#
tantek
welcome peat - how can the indieweb help? ;)
#
barnabywalters
tantek: no, I look for a link header using a HTTP request library, then use PHP DOMDocument to search for <link rel="[http://]webmention[.org]" elements
#
barnabywalters
using phpmf2 is a good plan for looking for HTML links as it handles all the URL resolution stuff automatically
#
barnabywalters
thanks to aaronpk
#
aaronpk
that was fun in a weird way
#
barnabywalters
and sandeep wrote a tiny little state machine for parsing link headers
#
tantek
interesting. right, webmention is *only* on <link> elements right, not on <a>?
#
aaronpk
always <link> elements inside the <head> element
#
barnabywalters
tantek: pretty sure I just look for elements with rel~=webmention and nonempty href
#
tantek
that's a restriction right? as in, do *not* support <a rel=webmention> ?
#
aaronpk
makes sense to me, although I don't think it's ever been explicitly called out as a restriction
#
barnabywalters
the idea is to allow people to use regexes to discover the headers, a la pingback
#
barnabywalters
so in theory you’re not allowed to put the rel and href attributes the wrong way round
#
tantek
hmm - that seems bad
#
tantek
even sandeep's code looks for rel/href in either order
#
aaronpk
the pingback spec was explicitly written to allow searching for the pingback endpoint with a regex
acegiak joined the channel
#
barnabywalters
s/doesn’t/does
#
tantek
aaronpk - that's because Hixie wrote it before he wrote the HTML5 parsing algorithm ;)
#
tantek
also single quotes vs double quotes vs no quotes
#
tantek
e.g. rel='webmention' or rel=webmention
#
tantek
the DOMDocument approach sounds reasonable
#
aaronpk
hah wow
#
barnabywalters
parsing HTML is not a big deal. we shouldn’t be advocating using regexes to parse HTML
#
tantek
agreed
#
aaronpk
especially if the mf2 libraries return rel values as well as the php one does
#
tantek
aaronpk - they do, however they don't distinguish <link> rels vs. <a> rels
#
aaronpk
good point...
#
aaronpk
not sure that matters though
#
barnabywalters
is that a problem?
#
aaronpk
it shouldn't matter assuming everyone agrees that all rel values apply to the whole page
#
tantek
and I've specified in the rel registry that webmention is only allowed on <link> http://microformats.org/wiki/existing-rel-values
#
barnabywalters
be liberal in what you accept, etc…
#
tantek
aaropnk - it's not the first rel value that only works on <link>
#
tantek
there are many other <link>-only rel values
#
aaronpk
yea that's fine. is it a problem that a parser would find a rel value on an <a> tag then?
#
tantek
I believe the threat model is that someone may insert a new <a> hyperlink in a comment or something with a rel value - but I've yet to see this happen in practice
#
tantek
it may just be for validation for authors
#
barnabywalters
if you’re letting people inject arbitrary HTML they you probably have bigger problems than webmentions going to the wrong place
#
aaronpk
barnabywalters++
#
Loqi
barnabywalters has 15 karma
#
tantek
it's not arbitrary HTML. it's many sites that allow *some* HTML including <a href>
#
tantek
though I've yet to see one that lets the rel attribute thru
#
acegiak
barnabywalters: I'm not actually even seeing the h-entry class on the page that asploded
#
acegiak
only hentry
#
barnabywalters
if it starts to actually cause problems then people will fix their purifiers
#
tantek
interesting, idno returns HTTP Link: for webmention as well
#
barnabywalters
acegiak: there’s .h-entry on the <body>
#
acegiak
oh, ok
#
acegiak
i missed htat
#
barnabywalters
protip: command-F in view-source window :)
#
barnabywalters
(or control-F if that’s your thing)
#
tantek
super-pro-tip - use pin13.net to parse it and show the JSON
#
barnabywalters
tantek: pin13 offers no way to activate BC mode, which is what’s causing these problems
#
tantek
oh I thought BC mode was on by default - no?
#
barnabywalters
unless there’s a hidden query param somewhere which does it
#
tantek
that would be quite handy for demonstrating these problems!
#
tantek
in fact, why not have it on by default?
#
barnabywalters
tantek: IIRC not by default because simple class replacement is hacky and quite slow
#
barnabywalters
PHP mf2 is not very fast anyway
#
acegiak
I'm thinking I might switch to a different theme and use the plugin to add the mf2 markup
#
barnabywalters
it gets up to multiple 10ths of a second on larger documents
#
barnabywalters
acegiak: I would really recommend adding the mf2 by hand
#
barnabywalters
you’ll have more control, learn more about how they work, and be able to add more details
#
barnabywalters
the plugin can’t hook into many elements
#
tantek.com
edited /webmention (+110) "/* IndieWeb implementations */ werd.io also sends the HTTP Link: header"
(view diff)
#
tantek
so it looks like it will work for current implementations if I look for HTTP Link: header with rel="http://webmention.org/", and if not found then a <link> with rel~=webmention
#
tantek
as in the selector link[rel~=webmention]
#
aaronpk
also sends a Link http header
#
acegiak
barnabywalters: what about the "focus on UX" principle? I'm trying to make the workflow as simple as possible
#
tantek
acegiak - what does this have to do with workflow?
#
tantek
the microformats should all be automatic and not take *any* time in your publishing workflow
#
tantek
(unless you're writing a structured blog post with lots of references inside which could use microformatting)
#
acegiak
that's pretty much the opposite of what barnabywalters just said?
#
tantek
but certainly the generic h-entry markup should all be automatic in posting workflow
#
tantek
he meant you should add the microformats in by hand into your templates etc.
#
tantek
not into your posting content!
#
barnabywalters
acegiak: at the moment the tools which have a good UX (the wordpress plugin) don’t do a very good job of adding mf2, due to wordpress’s limitations
#
tantek
and when did you start sending/receiving webmentions?
#
acegiak
that makes sense
#
barnabywalters
it’s a nice drop-in quick fix, but if you care about good markup (which I’m assuming is the case, as you’re here ;) editing by hand is currently the most effective approach
#
tantek
editing *your templates* by hand
#
tantek
rather than editing *your posts* by hand
#
barnabywalters
when pfefferle drops by we can figure out if it is an issue with sempress, and hopefully resolve it
#
tantek
maybe leave him a question !tell about it
#
barnabywalters
tantek: I left one eariler
#
acegiak
has the same issue
#
barnabywalters
the fact that this is happening at all points to a need for better documentation and validators
#
tantek
barnabywalters - we're using the pin13.net validator currently - what needs to be "better" about it?
#
acegiak
cause I might just fix the issue on mine and then pfefferle can merge if he wants to
#
barnabywalters
acegiak: h-cite markup there looks good to me!
#
barnabywalters
the nested <a> elements are a bit weird
#
tantek
yeah - there's almost no reason to ever do that
#
tantek
usually means you have a bug
#
barnabywalters
tantek: automatically detecting problems like this and bringing them to the user’s attention
#
barnabywalters
which is what bnvk and I are working towards on indiewebify.me
#
barnabywalters
offering documentation, support etc
#
tantek
barnabywalters - since this problem came from mixing microformats - that's where having pin13.net's validator do BC by default would be useful
#
barnabywalters
well I can change phpmf2 to do BN by default without too much trouble
#
barnabywalters
I doubt it’ll cause any problems
#
tantek
I think that it should do that be default because that's what the spec says to do by default
#
tantek
the option should be to omit
#
tantek
if you think that will make a perf difference etc.
#
tantek
we're going to be in a hybrid markup world for a while, so might as well have the tools assume/support that
#
barnabywalters
sure — any progress getting new e-* parsing rules in the spec?
#
barnabywalters
that will be a bigger change, breaking BC with previous versions
#
barnabywalters
phpmf2 is getting a bit messy, it needs cleaning — may as well do API changes at the same time as other BC breaking parts
#
tantek
barnabywalters - not yet no - hence why I at least linked to the etherpad and summarized issue resolutions on the issues page first
#
barnabywalters
is it considered finalised and ready for implementations though?
#
barnabywalters
I have no issues with it
#
tantek
I think so since we have unanimous agreement on the resolution
#
tantek
and no one has suggested any alternatives since
#
barnabywalters
so I’ll clean phpmf2 and the shim, update mf-cleaner and then release shiny new versions
#
barnabywalters
sounds like a good weekend project
#
tantek
also sounds like good motivation to update the parsing spec accordingly at the same time
#
tantek
I wonder how glennjones's parser updating is going.
#
aaronparecki.com
edited /webmention (+81) "filled in dates of p3k implementations"
(view diff)
#
aaronpk
git history ftw
#
tantek
aaronpk, you mean you didn't blog/tweet about it when you implemented it?
#
aaronpk
not always
#
aaronpk
and even if I did it's usually harder for me to find those
#
barnabywalters
I don’t have permissions to edit the parsing spec, otherwise I’d have added it already :)
#
aaronpk
more often I drop a note in IRC
#
tantek
harder to search your own site than search github?
#
aaronpk
not github, local git history
#
tantek
barnabywalters - darn it - I need to just see about proposing making you an admin, or rollback some of the changes protections since it seems like we have a handle on the spam/vandalism.
#
acegiak
quick question about the long web principle: what happens when folks die? morbid but I always thought that was an advantage to silos
#
tantek
acegiak - silos have killed off more permalinks than personal sites when people die
#
hadleybeeman
Valid question, acegiak. How long do you think it'll be before we see "20 years of ongoing hosting!" added to death insurance?
#
acegiak
hadleybeeman: oh that's a good idea
#
tantek
so no, it's not an advantage for silos (per silo-deaths data)
#
acegiak
tantek: oh yeah I don't doubt that
#
hadleybeeman
:) Go for it. Hope you make a bundle, acegiak.
#
tantek
average silo lifespan is much shorter than average use of personal domains
#
acegiak
tantek: yeah, I'm just thinking that they at least give the illusion of permanence, which I think is part of the draw
#
acegiak
and the reason people put so much of their lives into things like facebook
#
tantek
acegiak - I still want to simply will my domain to the internet archive to keep it registered and redirecting to the last static copy they have
#
acegiak
tantek: that's a good idea
#
tantek
acegiak - no illusion any more: http://indiewebcamp.com/site-deaths
#
hadleybeeman
I suspect most people also don't think about permanence. Silos give the illusion of "managed", and it's easier to say, "I assume they'll have thought of everything I might need."
tpinto joined the channel
#
acegiak
hadleybeeman: ah ok. My reason for using facebook when I had it was partly the illusion of permanence
#
hadleybeeman
S/easier to say/easy for most users to assume
#
hadleybeeman
That makes sense, acegiak.
#
acegiak
but I spend a lot of time thinking about things like that which are maybe not as pressing for many other people
#
acegiak
different priorities etc
#
tantek
acegiak - nah, people in general are quite upset when all their family photos etc. disappear
#
hadleybeeman
In my view, that one comes up for average user when a Facebook friend dies. It's an odd and jarring experience.
#
hadleybeeman
(But I can see why the issue wouldn't occur to someone often for consideration.)
#
acegiak
I started thinking about it when a younger cousin at my grandmother's funeral a few years ago said "She doesn't even have a facebook page" as if they couldn't remember her without that and hten I wanted to know what their actual policy is
#
barnabywalters
handling deaths on FB is something which makes me angry at facebook (they could take the opportunity to have a real positive effect on people’s lives)
#
barnabywalters
but people tend to use the inadequate tools it gives them to mourn and celebrate anyway
#
acegiak
do they delete content?
#
barnabywalters
I don’t think so
#
barnabywalters
so good people make up for facebook’s inadequacy
#
tantek
all they need to do is add a "power of abandoned account" feature where you can specify one or more other users that you authorize to act on your behalf if your account appears to be abandoned for some duration (a week? 30 days?)
#
tantek
way easier than writing a will
#
aaronpk
that would be great
#
hadleybeeman
Acegiak: a friend of mine died last summer. Facebook waited for x number of death reports from us (corroborated with obituaries), and then they eventually "memorialised" his page — freezing some of the features.
#
acegiak
hadleybeeman: inlcuidng, hopefully, no longer posting "sponsored stories" or whatever it is they do
#
hadleybeeman
Right. And no longer accepting friend requests.
#
hadleybeeman
Also, oddly enough, his profile underwent a lot of traffic (mostly new tags in photos people were posting in their own accounts) and my timeline was flooded with them. Bug? Feature? Algorithmic result of the activity? I don't know.
#
tantek
such a feature would also enable an interesting way to walk away from a FB profile without deleting it
#
tantek
if you trust a friend or family member to look after it after you've walked away from FB.
#
tantek.com
edited /webmention (+226) "/* How to Test Webmentions */ another caveat - some sites/tools only do discovery via HTTP Link header"
(view diff)
#
acegiak
ok back to more immediate practical questions: what do I need to fix with this h-entry markup?
#
acegiak
should it be one h-entry on the body and one on the content or just one on the content?
#
barnabywalters
acegiak: h-entry and hentry need to be on the same element
#
barnabywalters
doesn’t particularly matter which element, body is good
#
acegiak
I'm actually having trouble finding the code that inserts hentry :S
#
acegiak
does windows search work like grep or just like find?
#
tantek
"having trouble finding the code that …" is one of the reasons I gave up on messing with WordPress
#
acegiak
tantek: that's fair enough
#
barnabywalters
acegiak: I had a poke around in the sempress template and couldn’t figure out where anything was
#
tantek.com
edited /webmention (+489) "add subheads to brainstorming, add Discovery in PHP section"
(view diff)
#
acegiak
barnabywalters: a lot of the stuff is actually in the functions.php
#
tantek
acegiak - re: where to put which microformat class name, see: http://microformats.org/wiki/uf2#backward_compatible
josephboyle joined the channel
#
acegiak
tantek: I wrote a pretty decent basic cms back in high school but my lack of make things pretty skills meant that it was continuously ugly
#
acegiak
oh, cool
#
tantek
the key is to modularize it so that "make things pretty" does not depend on the content flow
#
tantek
then you can more easily get help with "make things pretty"
#
acegiak
and I'm thinking of moving him to wordpress so he can use the app on his phone to post updates
#
tantek
but then you'll saddle him with an update problem
#
tantek
or a spam problem
#
tantek
or both
#
acegiak
spam is only if he allows comments, which he doesn't have at the moment anyway
#
barnabywalters
okay, phpmf2 updated to parse in BC mode by default: https://github.com/indieweb/php-mf2/tree/v0.1.22
#
acegiak
update problem is only necessary if he's concerned about security and the system he's using now has all the security holes of something written by an highschool student
#
barnabywalters
goodnight all
#
tantek
acegiak - spam also meaning having his site taken over because he missed a zero-day wordpress vulnerability/update
#
Loqi
buenas noches
#
acegiak
barnabywalters: night
#
tantek
aaronpk: do you do webmentions synchronously in your posting UI? or do you do them asynchronously? either way, do you report webmention results back to your posting UI?
#
tantek
and by "do" I mean discover/send
#
acegiak
tantek: which is less likely to happen to a custom coded but less security aware system?
#
aaronpk
totally asynchronous and invisible
#
tantek
acegiak - absolutely
#
acegiak
fair enough
#
tantek
acegiak - when consider what to put friends/family members on - for me #1 priority is zero maintenance/headache on their part
#
tantek
which WordPress is far from
#
tantek
aaronpk - wow - I mean, I guess that makes sense? you don't care what the return values are then?
#
aaronpk
I would consider giving myself some feedback in my UI
#
aaronpk
but if it doesn't find a webmention endpoint then what really can I do?
#
aaronpk
I suppose if I knew it did or did not send a webmention, I could decide whether to ping that person in IRC or something
#
tantek
I'm thinking more of the error case in sending a webmention then getting an error
#
tantek
I have a feeling we'll be adding UI at that point once we implement whitelists or other spam fighting measures.
#
aaronpk
well we're also encouraging async processing of accepting webmentions, so most implementations will just send back "ok" even before they go and check whether the site links back
#
tantek
also - would be kind of cool to see a list of successful webmentions sent in the posting UI
#
tantek
that's kind of where I'm heading in my thinking
#
tantek
thinking about what might demo well
#
aaronpk
could show a list of all outbound links from your posts, and whether a webmention was successfully sent for each
#
tantek
showing a list of links and then a green checkmark saying webmention sent on each that accept it
#
tantek
right :)
#
tantek
then you could go check those links with checkmarks to see if they parsed your comment or did something else with it
#
tantek
also - you might discover sites that newly add webmention receiving support that way
#
tantek
when something you link to says it accepted the webmention and you see it in your UI. imagine if you saw that popup next to an app.net permalink.
#
tantek
I like UIs that provide feedback