#indieweb 2017-01-17

2017-01-17 UTC
#
KartikPrabhu
database--
#
KartikPrabhu
database--
#
Loqi
database has -1 karma
#
Loqi
[indieweb] "Interview mit Aaron Parecki" by Matthias Pfefferle on 2015-07-23 http://notiz.blog/2015/07/23/interview-mit-aaron-parecki/
#
Loqi
[indieweb] "10 Jahre notizBlog" by Matthias Pfefferle on 2015-12-07 http://notiz.blog/2015/12/07/10-jahre-notizblog/
#
Loqi
[indieweb] "IndieWebCamp 2014 – This is a movement" by Matthias Pfefferle on 2015-07-25 http://notiz.blog/2015/07/25/indiewebcamp-2014-this-is-a-movement/
#
Loqi
[indieweb] "IndieWordPress – WordCamp Frankfurt" by Matthias Pfefferle on 2016-09-23 http://notiz.blog/2016/09/23/indiewordpress-wordcamp-frankfurt/
#
Loqi
[indieweb] "IndieWeb – Die Daten sind wir!" by Matthias Pfefferle on 2015-04-02 http://notiz.blog/2015/04/02/indieweb-die-daten-sind-wir/
#
Loqi
[indieweb] "IndieWeb im @DKultur/@breitband" by Matthias Pfefferle on 2015-05-01 http://notiz.blog/2015/05/01/indieweb-im-deutschlandradio/
wolftune and Lana joined the channel
#
GWG
miklb_: Any chance to make notes?
#
miklb_
GWG where do you want them?
#
Loqi
miklb_: GWG left you a message 4 hours, 44 minutes ago: Uploaded missing change from my phone
#
GWG
miklb, Github.
aegibsonme joined the channel
#
GWG
I am pleased with the progress, but know that I have things to iron out.
#
miklb_
I need to update with the new file.
#
GWG
miklb, stupid mistake on my part. It was late
#
miklb_
GWG I'm not seeing the file in the commit
#
miklb_
or am I not following
#
GWG
The file I forgot to commit a change to was bridgy_publish.php
#
miklb_
oh, I just modified the Class to use admin_init instead of init and it worked
#
GWG
I didn't make any modifications today, but you said you had more thoughts
#
miklb_
well, when I tested with micropub, it did show the syndication targets, but it didn't post to Twitter via bridgy.
#
miklb_
when I looked at the post on in WP admin, nothing was checked.
#
miklb_
goes to an issue in GitHub
#
GWG
But did it generate a tweet?
#
GWG
I'll be working on improving that code.
#
miklb_
GWG no. I had to check the option and update the post manually
#
GWG
I got it to go, oddly. But I will work on further testing
#
miklb_
I'll test again later this evening.
#
miklb_
Also, even though I have "disable back links" set to "maybe" I'm still getting just site title and link. When I looked at the parsed post in bridgy it seems to be marked up correctly, so not sure what need to do to get it to post the content and not link.
#
miklb_
added that to issue also. Need to get some $work done before I do anymore on my site.
wolftune, tolerablyjake, Kamelreiter_, tantek, snarfed and [shurcool] joined the channel
#
[shurcool]
I'm wondering, how do people generally handle users (user data) on their personal sites. Do you have a users table in a db, or something equivalent? Other ways?
#
aaronpk
Mine is all stored completely denormalized. Each reply stores all the data it needs in that row. For replies, the reply context is stored in the post file. One of my recent blog posts talks about that.
#
aaronpk
im in the process of turning that into a nicknames cache but I'm not there yet
#
[shurcool]
nice, thanks :slightly_smiling_face:
#
[shurcool]
> In this implementation, I store the contents of the post I'm replying to in my reply post itself, rather than in some external database or file.
#
Loqi
[Aaron Parecki] Day 25: Automatically Fetching Reply Contexts for p3k #100DaysOfIndieWeb
#
[shurcool]
a potential problem is if someone changes their name/nick, that change won't be reflected on old content. but that's probably fine/not a big deal, maybe even good.
tantek, wolftune, KevinMarks, guest175 and KevinMarks_ joined the channel
#
Zegnat
shurcool, I don't think it's a big deal when old content refers to old nicknames. I believe this is the case almost everywhere, including on Twitter and GitHub where @-mentions could theoretically be updated
KevinMarks and [kevinmarks] joined the channel
#
[kevinmarks]
When you export your twitter archive, it gives you people's current name and photo, not the one at the time. So it isn't an accurate archive.
snarfed joined the channel
#
[kevinmarks]
Also, if you do a dump in October, you get the Halloween names and avatars for everyone you ever retweeted
cweiske joined the channel
#
Zegnat
KevinMarks, so it updates the names of people you retweeted but not of people you at-mentioned?
guest175 joined the channel
#
KevinMarks
It depends when you @ mentioned them I think. If it was autolinked I think it changes
guest175 joined the channel
#
KevinMarks
Looks like they don't change the @'s based on that
#
KevinMarks
So if you switch and don't grab the old handle, it could mislink
guest175 joined the channel
#
Zegnat
Yeah, that's what I seemed to recall. I think mentions are just stored as text-of-tweet and made clickable on render, meaning you don't get username changes, while retweets are stored by tweet ID and will always render with the latest username on it.
aegibsonme, pindonga, jihaisse, pfefferle, rrix and adactio joined the channel
#
pfefferle
good morning
#
Loqi
guten morgen
#
@ncollig_net
Let's create the web of tomorrow, together! #IndieWeb, #selfhosting, #encryption, #decentralization & many other topics @HwcBrussels soon!
(twitter.com/_/status/821298219200159744)
nitot, Pierre-O, DanC, Agadir, jihaisse, userXYZ, Kamelreiter, mlncn, hs0ucy, guest1751 and [jeremycherfas] joined the channel
#
[jeremycherfas]
OK, thanks to a bit of encouragement from tantek and aaronpk I have started trying to describe my current mess (https://www.jeremycherfas.net/blog/putting-my-house-in-order-phase-1) in the hope that I can first, get some suggestions from more experienced people here as to how best to arrange things and, secondly, to give myself a clearer idea of what I need to do. Comments here welcome.
jonnybarnes, hs0ucy and thebaer joined the channel
#
Zegnat
jeremycherfas, could you elaborate on "Indieweb Slack channel, where I'm both amazed at what people are doing and increasingly convinced that it is beyond me"?
#
Zegnat
We want to try and get technical talk out into our IRC dev channel and make this main channel more approachable. But we need people to keep reminding us of that!
#
Zegnat
What is Grav?
#
Loqi
Grav is a flat-file CMS built on PHP, with Twig templating, and YAML + Markdown for storing articles (YAML for metadata, Markdown for the content) https://indieweb.org/grav
#
[jeremycherfas]
Well, to be honest, I sometimes see things here where I don’t really understand what people are talking about. But I figure that’s because I am ignorant, and that as I hang about here and listen and read, I’ll understand more. I don’t want to whine every time I don’t understand anything.
#
Zegnat
To be fair, you should whine! Or at least just write "What is X?" and have Loqi explain the term.
#
Zegnat
If Loqi doesn't know what something is, or the definition doesn't help you, prompt people to update it.
#
Loqi
says if loqi doesn't know what something is
#
Zegnat
We want to make things as easy as possible.
#
[jeremycherfas]
OK. Well, I’m making a start on jeremycherfas.net and I have an h-card on the home page. I don’t actually want people to see it because there is a separate About page, so I have set `display:none;` I believe that as the content is still in the html, I should be ok with that.
#
Zegnat
Parsers should be fine with that. But hidden metadata is always a pity.
#
Zegnat
If you read the brainstorming (and issues) here, you will see that you are not the first to have this /about problem: http://indieweb.org/h-card#Brainstorming
#
Zegnat
If you want to elaborate on the use case, and maybe dream-up a solution, add it! No solutions seem to have been documented so everything you say is as valid as anything else
#
Zegnat
(again, don't feel like others are too experienced for you to contribute anything.)
#
[jeremycherfas]
I’ll read that now. In the meantime, I have established that indiewebifyme does indeed check out for on of the rel-me links.
#
[jeremycherfas]
And the h-card validates OK
#
Zegnat
Nice job!
Agadir and snarfed joined the channel
#
Zegnat
jeremycherfas, you have put rel="me" on https://twitter.com/EatPodcast and I am not sure if that is completely correct. While you do *run* that Twitter, I would be inclined to say that "EatPodcast" is a different entity from you. That is why aaronpk and tantek were asking you to distinguish project and identities yesterday (https://chat.indieweb.or
#
Zegnat
g/2017-01-16#t1484587881304000)
#
[jeremycherfas]
Thank you. There are some problems with the About me plugin for Grav, but I have raised them with the plugin author. Now to see about the webmentions plugin!
#
Zegnat
There is a webmentions plugin? Didn't see one on /grav on the wiki
#
[jeremycherfas]
I know! That’s the issue with the About me plugin; it isn’t picking up some of the changes I have been trying to make!
hs0ucy joined the channel
#
Zegnat
Do you have a link to the issue? On GitHub?
EHLOVader joined the channel
#
[jeremycherfas]
[Yes](https://github.com/Perlkonig/grav-plugin-webmention), it is by Perlkonig and I thought I might give it a try and see whether I can offer any useful feedback.
#
[jeremycherfas]
The issue is at https://github.com/Birssan/grav-plugin-about-me/issues/7 and I’ve just seen that there is a response. Better look at that.
#
snarfed
[jeremycherfas]: you may be interested in http://indieweb.org/h-card#Issues
#
cweiske
when I see what effort goes into *parsing* html pages with microformats, how does anyone expect that people will prefer h-feed to rss/atom feeds? (see dev discussion and x-ray code and mf2-parsing-lib-code)
#
[jeremycherfas]
Thanks snarfed. Yes, I looked at that after @zegnat pointed me there, and I have added a line that I made my h-card invisible.
#
Zegnat
jeremycherfas, I just documented the webmentions plugin on the wiki, thanks for pointing it out :)
#
petermolnar
cweiske fair question, but parsing rss, atom, rss+atom, rss2, etc. is not much nicer
#
petermolnar
the difference it that that is kind of done
#
petermolnar
and is a fixed format in many terms
#
petermolnar
unline mf2
#
aaronpk
"done" for blog posts, but even simple text notes don't work well in those formats
#
aaronpk
not to mention other kinds of posts
#
Zegnat
That question should be split between consumers and publishers too I think. As a publisher you lose the need to support extra sidefiles for one.
#
Zegnat
cweiske, note that criticisms have been documented: https://indieweb.org/feed#Criticism
#
cweiske
I have a script that imports podcasts into my UPnP server
#
cweiske
no dependencies
#
aaronpk
XSLT sounds like a dependency to me ;-)
#
aaronpk
yes, podcast feeds are a pretty frozen spec, since there are so many consumers of them. so naturally the tools available to handle that are well established.
#
aaronpk
although ironically google added their own meta tags if you want to publish a podcast on google play although they also parse the itunes tags as well
#
snarfed
also cweiske simply downloading a media enclosure URL is a very small, narrow use case, much simpler than readers or other use cases here
#
cweiske
snarfed, it does not matter if it's a narrow usecase. just imagine I have that small narrow usecase, but with h-feed
#
snarfed
cweiske: sure. my point is that small narrow use cases may be the exception, not the rule, for both rss and h-feed
#
snarfed
(also for the same use case with h-feed, ie download each h-entry's url, i bet you could do it with almost the same XSLT and wget)
#
Zegnat
Anything that lets you write queries on the output of an html5 parser will let you quickly extract all u-audio links from an h-feed.
#
cweiske
i'll come back when there are html5 parsers available in common programming languages and on shell
#
cweiske
I think it's ok to have a tightly controlled output format if that makes parsing easy, but the meinungsmacher here don't
#
cweiske
s/tightly controlled/limited/
#
snarfed
python: htmllib, shell: xsltproc, etc
#
snarfed
cweiske you're definitely right more broadly. HTML can be painful to parse, even with a good library. smaller formats like JSON are often easier and feel simpler. RSS often isn't much better than HTML though.
mlncn_ joined the channel
#
aaronpk
I'm a fan of "limited" output formats, however creating that format needs to be done in a slow and gradual way to avoid the problem of too early defining that format and forgetting about some need and having someone else come along later and start the work over again. The way I avoid the headaches of thinking about HTML parsing is by offloading all that to the microformats2 parser so that I start my work
#
aaronpk
with a much less variable input.
#
cweiske
snarfed, xsltproc does not work with html(5)
#
snarfed
backs away slowly
#
aaronpk
wanting to have a more "limited"/"restricted" output format is why I encapsulated a bunch of logic in XRay. so now when I'm consuming HTML pages I'm actually working with a fairly tightly defined JSON structure.
#
aaronpk
I based some of the design of XRay on that
tolerablyjake and [kevinmarks] joined the channel
#
[kevinmarks]
Is it worth iterating on jf2 again? It being a simple programmatic transformation of mf2 may not be the ideal, listening to this, as we all have written code to simplify mf2 markup variants
#
[kevinmarks]
Html5lib is useful, but poorly documented and fiddly to use
tantek and jmelesky joined the channel
#
aaronpk
[kevinmarks]: my iteration of jf2 has taken the form of XRay, which is more than just a json transformation, since it's actually vocabulary-aware and knows the /authorship algorithm
guest175, snarfed, wolftune and aegibsonme joined the channel
#
KevinMarks
Right, which in some ways is a bit more mf1-like
#
aaronpk
only from the consuming side, since multiple different ways of authoring the HTML will result in the same XRay output
#
KevinMarks
Right - it ends up more tightly structured
#
KevinMarks
Is there a case for a lingua franca between webmention.io, webmention.herokuapp.com etc for normalised responses?
#
aaronpk
i think so
#
aaronpk
also likely useful in readers
snarfed1, wolftune, tantek, leg and snarfed joined the channel
#
KevinMarks
!tell cwieske parsing feeds is not easy either.
#
Loqi
Ok, I'll tell them that when I see them next
snarfed, snarfed1 and arush joined the channel
#
Loqi
[Sebastiaan Andeweg] Day 2: 410 Gone
#
sebsel
How do I get those [indieweb] posts? Is that Websub?
#
aaronpk
yeah, via superfeedr
#
aaronpk
Loqi's subscribed to a Superfeedr search for "indieweb"
#
aaronpk
so there isn't actually an explicit way to get them here, you have to kind of wait for superfeedr to index your site
#
aaronpk
hm i should work on that
#
sebsel
Ah okay!
#
aaronpk
oh also you can post to https://news.indieweb.org/ and it'll show up
#
sebsel
Yeah, but since I'm doing this every day not that would be overkill
#
aaronpk
yeah :)
#
sebsel
So I'll go for an overview on each 10th day :)
#
sebsel
just using all the ideas
#
aaronpk
hm it might be time to spiff up indienews a bit
#
sebsel
yeah it's just a collection of links now
#
sebsel
But no e-content might be by design?
#
aaronpk
correct
#
aaronpk
it's supposed to be like hackernews/reddit style
#
aaronpk
just acting as a hub to collect interesting links, but not actually host any content of its own
#
Loqi
[indienews] New post: "Week in Review #100DaysOfIndieWeb" https://aaronparecki.com/2017/01/14/4/week-in-review
#
aaronpk
you reminded me i forgot to post that one :)
#
sebsel
haha nice :)
#
Loqi
sebsel: lol
#
Jeena
I don't quite get it, how is Micro.blog different to Known (despite the fact that it already exists and does a lot more) is it just the better PR?
jonnybarnes joined the channel
#
sebsel
Yeah, the site looks really nice. That's 90% of your project, having a nice font. :)
#
snarfed1
Jeena: one difference is micro.blog has much bigger reader plans
#
aaronpk
yeah it seems to focus much more on being a reader whereas Known's focus is on publishing
wolftune joined the channel
#
sebsel
reading is important too.
#
sebsel
a lot of non-dev, non-writing people I know just read on social media
KartikPrabhu joined the channel
#
sebsel
they only post if it's temporary, like snapchat or instagram stories.
#
sebsel
(that's a feeling I have, no data here!)
marc___ joined the channel
#
Jeena
ah that is an interesting point
#
Jeena
I'd like a geed reader too, woodwind is the best I've seen but it is really rough to use and the UI isn't very smart
marcthiele joined the channel
#
aaronpk
building a good reader is one of my long-term goals
marc_thiele and marcthiele joined the channel
#
sebsel
one of the problems of a reader is the MB's. I've got 43MB of tweets and posts in ±20 days, for just me. And I'm not caching images like I probably should.
#
sebsel
Woodwind seemed to 500 because of full disks too
#
aaronpk
43mb wow. is that HTML or JSON or DB format?
#
sebsel
eh, I do jf2
#
sebsel
and plain text files
#
sebsel
with a SQLite indexer
#
aaronpk
ah neat
#
sebsel
yeah, joining the earlier discussion: I also was thinking about only storing the fields I need, that would save MB's.
#
sebsel
Most of the time I have triple content because of p-name, e-content[html,value]
#
sebsel
(I actually edited my jf2 thing because KevinMarks' parser lets go of the html, don't know what's the spec)
snarfed and snarfed1 joined the channel
#
snarfed1
sebsel: ~2MB/day/user of reader data honestly isn't that much
#
snarfed1
disk is cheap these days. woodwind does indeed run out often, but that's arguably more an ops problem than a tech problem. aggressively GCing read posts and maybe rotating logs would probably help.
#
aaronpk
yeah most of my storage problems are caused by not properly deleting old log files. at one point Monocle's log file was like 30gb because it was in debug mode and was outputting every post it crawled
tantek joined the channel
#
sebsel
lol, okay!
#
sebsel
I'm on a cheap shared host, and although I can move myself, I kind of like it as a restriction, seeing what I can do with it
#
sebsel
turns out quite a lot
#
sebsel
but hosting my own reader is pushing it
#
tantek
sebsel, I agree re: shared hosting
#
aaronpk
that's great
#
tantek
I am also on shared hosting, and also like it as a restriction so that what I develop can be re-used by more people
#
aaronpk
i set up XRay on shared hosting to see if I could, and it turns out it works quite well there
#
tantek
(also, it's a good baseline for portability of your own stuff too!)
#
sebsel
yeah! that's the other end of eating your own dogfood: don't give your dog expensive steak because you like that more :P
#
tantek
what is shared hosting?
#
Loqi
shared hosting is a particular level of webhosting service that includes FTP access with at least PHP and/or Perl CGI on a server where a bunch of people have accounts that share the same filesystem, and the same Apache/nginx process(es) https://indieweb.org/shared_hosting
#
sebsel
aaronpk XRay worked for me on my server! And with the new folder-capable setup I might set it up under seblog.nl/somerandomstring, so I can use it while others can't
#
aaronpk
awesome
#
sebsel
still have to change my webmention-format first though. But I have enough days left :P
#
tantek
this is worth documents
#
aaronpk
hah i didn't even think about installing it in a folder that has a sufficiently random name to be a "capability URL" as a security measure
#
sebsel
lol, yeah, I can't even ssh.
#
aaronpk
that's like even easier than editing a config file to define a hard-coded access token
snarfed joined the channel
#
KevinMarks
What was the problem with my jf2 code?
#
Loqi
It looks like we don't have a page for "problem with my jf2 code" yet. Would you like to create it?
#
sebsel
Oh, there's not really a problem.
#
sebsel
If you have an e-content, that would make an `content: [html, value]` (if I may abstract)
#
tantek
jf2 sounds like an #indieweb-dev topic
#
sebsel
Here you take only the value
#
sebsel
Oops. Well that was my point.
#
KevinMarks
I mean if it's not suitable for your use case, that's a problem
#
tantek
KevinMarks, see #indieweb-dev
tolerablyjake and wolftune joined the channel
#
GWG
snarfed, I am close to having an all snarfed enabled posting experience.
userXYZ joined the channel
#
GWG
https://make.wordpress.org/core/2017/01/17/editor-technical-overview/ Wonder if this might solve my microformats 2 markup issues
hs0ucy and snarfed joined the channel
rrix and snarfed joined the channel
#
tantek
sebsel, aaronpk I documented a brief summary of the shared hosting conversation above here: https://indieweb.org/shared_hosting#Advantages
#
tantek
please feel free to add more or edit accordingly!
#
tantek
thanks!
#
sebsel
maybe I should delete unread tweets in my reader after 12 hours. 90% doesn't seem relevant at all anymore.
#
sebsel
why follow those people? well, it's nice too read them if you do have the time
#
sebsel
tantek I see! Can I list no SSH as a common thing under shared hosts?
#
sebsel
I have seen a lot of shared hosts, and none of them had ssh.
#
snarfed
sebsel: try making a twitter list with a subset of accounts that you actually want to read. lets you still follow people socially without having to see everything they tweet
#
snarfed
i've been doing that for a while, and love it. e.g. i follow >500 people but only read a list of ~100.
#
sebsel
yeah, I believe I can't follow people directly outside of twitter...
#
sebsel
that sounds nice! I have been maintaining a <100 list for a while, but that became 150 and then I lost it
#
sebsel
I'm actually using Granary to subscribe to my feed.
#
sebsel
looks for lists
#
snarfed
https://twitter-atom.appspot.com/ generates atom for twitter lists. nicer UI than granary :P
#
sebsel
ohh :o
#
sebsel
snarfed++
#
Loqi
snarfed has 248 karma in this channel (252 overall)
#
sebsel
thanks :)
#
sebsel
I will have to create RSS/Atom support for my reader then :)
#
snarfed
ah, or just stick with granary twitter list => h-feed
#
tantek
snarfed I tried that for a while, but found that the larger problem was people I wanted to read just getting more ranty on Twitter in particular. it's like Twitter reinforces rantiness </rant> :P
#
snarfed
heh yeah. user-defined filtering within individual accounts is an open problem. maybe an indieweb opportunity.
#
sebsel
posting tweets on my blog first made me more aware of it's contents
#
sebsel
a part of the tweeting brain forgets that tweets live longer than two scrolls :P
#
tantek
true. not how it feels when I post notes on my own site (they feel more "permanent", more findable)
#
tantek
also the rantiness is a Twitter problem, not a silo problem. In contrast, Instagram seems to cause people (the same people!) to post more positive things, personal achievements or moments of personal appreciation
#
tantek
if anyone can figure out the specific differences in UX design between Twitter and Instagram that cause that disparity, I would be very interested in hearing any theories!
#
@cswordpress
Curating comments. In other words, Google's introducing a Twitter/FB-like algorythm. #indieweb https://twitter.com/jkphl/status/821472642775580672
(twitter.com/_/status/821483291115225089)
miklb and Lana joined the channel
#
Loqi
[indieweb] "Putting my house in order: Phase 1" by Jeremy Cherfas on 2017-01-17 https://vaviblog.com/2017/putting-my-house-in-order-phase-1
#
KevinMarks
With instagram you're posting in someone else's comment thread, with twitter you are shouting into the void
#
KevinMarks
Also, instagram has historically been performative - image creating, curating the world
#
KevinMarks
!tell sebsel unmung.com does atom/rss to h-feed
#
Loqi
Ok, I'll tell them that when I see them next
#
tantek
KevinMarks: I don't understand your first point - posting in someone else's comment thread? On Instagram, all posts are top level.
#
tantek
if you mean replies/comments then yes, those go under others's posts - but the flipside is you lose track of them (classic silo comments problem)
#
tantek
also Instagram provides MUCH better comment/reply moderation tools than Twitter
wolftune joined the channel
#
KartikPrabhu
i think it also depends on how well you manage your Twitter feed
#
KartikPrabhu
I follow very few people and it does not seem ranty
wolftune joined the channel
#
tantek
KartikPrabhu: it's a tough balance. I want to follow people that care about things beyond technology, but sometimes that comes with rantiness
#
KartikPrabhu
yeah, it not easy
#
KartikPrabhu
also people post all sorts of things not just one category
#
KartikPrabhu
I have that problem in my /reader . Posts from people I follow but not the category of posts I want to see
#
tantek
basically, no one has really "solved" the reader problem
#
tantek
despite people claiming Twitter and FB have - they haven't, they've in some ways made it worse (algorithm filtering, so-called "bubble" reinforcement, etc.)
wolftune, userXYZ, CherryPuffs and [keithjgrant] joined the channel
#
[keithjgrant]
I've always wanted to be able to subscribe to select topics/tags from people I follow. Unfortunately it puts the onus on the publisher to tag consistently
#
[keithjgrant]
I could conceive of a future where various readers have algorithms & machine learning to handle it. It would make for an interesting aspect of choosing a reader
#
tantek
right, I think any such filtering must NOT depend on the publisher explicitly tagging
#
tantek
a search filter might work
#
tantek
(either filter-in, or filter-out)
#
aaronpk
i tag my posts pretty thoroughly but even i miss some tags that people might want to be reading