#j12tEach of which has a bunch of steps, with screenshots and/or video.
#j12tGoal would be that even somebody only moderately confident in their abilities to complicated tech stuff can be successful simply by following detailed instructinos.
#tantekj12t - sounds like a pretty good vision, will try to read and understand and how I can help!
#j12ttantek -- what would be really useful (once I'm a bit further along) is to try out these trails, see what I missed / got wrong / didn't explain properly ... I'm going through all the steps myself right now, signing up for new accounts etc etc but experience shows I will miss substantial things :-)
#Loqi[IndieWeb WordPress Outreach Club] Description
This plugin is based on the idea and code of @snarfed:
One big IndieWeb raison d’être is using your own web site to reply,
like, repost, and RSVP to posts and events. You do this by annotating links on your site with simple microfo...
[miklb] and tantek joined the channel
#[miklb]it was moved to a plugin as I understand it snarfed, but I didn’t realize the IW press this would break.
#snarfedthe bookmarklet was the important part for the flow, and that's gone :(
#[miklb]in that case, we could fork the new plugin and just build the IW parts in.
#[miklb]since you have to install the IW press this plugin anyway, if we just merge the new plugin into ours same user experience
#[miklb]one time fork. I don’t think there is going to be much effort put into the Press This. I think they are going a different direction and possibly moving it into a Jetpack feature.
j12t joined the channel
#schmartyhow does Press This work without a bookmarklet?
#sknebelwhat specifically did they remove? I assume the press-this.php took URL params before?
#[miklb]how much change has occurred while it was in core?
#snarfed[miklb]: security patches and bug fixes are the parts i care about. but they also substantially redesigned within the last year or two
#sknebel(going by kraftbj comment "As it stands, copy an URL, visit press-this.php, paste the URL and click the button to parse it.")
j12t_ joined the channel
#snarfedsknebel: not sure of details, but yes, those url params no longer work
#tanteksnarfed, indeed those kinds of changes (substantial redesign) tend to be a reason people learn to avoid habitual upgrades, nevermind breaking key functionality like a smooth bookmarklet Press This user flow. kinda shocking
#snarfedeh change is the only constant, updates are important for security and bug fixes, and the new redesign in this case actually was significantly better.
#tantekTHIS: "Telling a user to copy a URL and then paste it is like telling someone to set the clock on their old-school VCR."
#[miklb]I didn’t realize they had redesigned it. I thought it had been pretty static for quite some time. I know there have been talks about doing something similar with the REST-API but not sure where that went.
j12t and j12t__ joined the channel
#sknebelyou guys don't think you can get the bookmarklet functionality added back as an "interim solution" until they get extensions figured out (which might take ages, but that's the reason to have a "temporary solution"? I get the issues with bookmarklets, but they still work mostly...
j12t_ and j12t__ joined the channel
#snarfedsknebel: sure we can. it's all code, anything's doable. it was just very low on my todo list before, and i'm grumpy about it maybe having to prioritize it now.
#tanteksnarfed, re: "updates are important for security and bug fixes", what happens when an update introduces a *worse* bug (e.g. breaking Press This bookmarklet) than what it may happen to have fixed, whether security or otherwise?
#snarfedtantek: we've had this debate a few times before ?
#tantekI suppose in open source the answer is to fork, take the "good fixes" and leave out the bad ones
#snarfedexcept that's a huge time and effort burden to ask of most users
#tantekwith closed source it's harder to make that trade-off
#tantekit's all about time and effort, and when the update *wastes* your time and effort by breaking something that made you more efficient, then it should be avoided
#tantekthere is no "updates are always best" absolute. it is about time & effort, and often updates are not worth the time & effort, or worse, cost you additional continuous time & effort
#snarfedre security, in general you're significantly more secure updating than staying on old, unpatched software. security is also herd immunity and community responsibility.
#snarfed...but you're right, security is only one of many factors. (an important one though.)
#tanteksecurity is also not an absolute as we've discussed before, there are (potentially, likely) ways of mitigating nearly any security "issue" that's resolved, often at additional layers (defense in depth etc.)
#tantekand the time & effort cost of those other layers may be *less* than the damage that an update does to your own productivity
j12t joined the channel
#snarfedeh yes but that's probably unusual. again, the vast majority of users don't have skill/time/effort to add either their own security *or* maintain their own features, etc
#snarfedeven arguably most of us, at least in code we didn't write ourselves
#snarfedso for everyone who's not highly technical, i try to keep the message simple: update apps and use 2fa. (yes, ideally non-sms.) there are downsides to both, but definitely more upsides.
#snarfedalso on a philosophical note, change really is the only constant. if you try to keep things the same forever, you're bound to be disappointed. be clear eyed and realistic about how much external change you try to fight or prevent. :P
#tantekenough on updates, let's pick on non-selfdogfooding decentralization advocates instead, specifically blockchainers
#Loqi[Iiterature] @pfrazee @team_slava @jessewldn @Truebitprotocol @chain @kin_foundation @PROPSproject @VoltzRoad @zeligf @joincolony @relevantfeed @dat_project Let's be explicit about the benefits and tradeoffs. Also the potential for incentives that tokens introduc...
#tantekthree tweeters in that thread, only one has their own website, zero of them post to their own decentralized websites (or other discernible solution) and instead, ALL of them ONLY post to the centralized Twitter service
j12t and j12t_ joined the channel
#tantekDear those tweeting about "decentralization" on Twitter and presenting yourself as an expert and/or knowledgeable in anything decentralization, scaling, etc.:
#tantekYour decentralizaztion tweets have zero credibility until you are posting even just simple "notes" (AKA tweets) on your own decentralized solution (like your own website) instead of only on the centralized Twitter service.
#tantek(didn't even need to mention blockchain, but I feel like as the hypetech of the moment, it's the worst offender of that)
[eddie] joined the channel
#[eddie]Is there any place that details silo.pub besides https://silo.pub/developers ? Specifically I'm wondering if it returns the URL of the post after it's posted to use for syndication links But I can't seem to find any info on that
#tanteksnafed, (perhaps this should go in meta or chat), re: "change really is the only constant", I disagree from a framing perspective, as the important aspect is not change, but time.
#tanteknearly all (security or otherwise) defense is about *slowing* change, not stopping it.
#[eddie]Cool. I've been using bridgy publish, but that requires me to post, rebuild my jekyll site, and then do a webmention. silo.pub seems to have a wider span of support (specifically thinking about Goodreads and GitHub) AND I can have it happen during my micropub receive step so that the syndication URL can get baked into my posts. ? So I think I might transition over to that for my POSSE
#tantekthere are strong economic reasons to deliberately slowing the changes thrown at you (e.g. $ : fewer purchase over a period of time, less time spent waiting for updates on various devices etc.)
#tantekso no, change is not a constant, it's just non-zero, and it is heavily malleable at your discretion, for personally improving your freedom (time), and agency ($)
[miklb] joined the channel
#[miklb]eddie the syndication links part was awkward with brid.gy. I was storing the response from brid.gy as _data files and then grabbing the url for my posts, requiring another round trip through my CI/GH/server workflow
#tantekre: SMS TFA in particular, I just had to help someone recover their Amazon account that had a phone number setup with it, and there were actually *bugs* in their account recovery flow that only exhibited when you gave amazon a phone number
#[eddie]miklb: Yeah, it's tricky with Jekyll. I do manually webmentions right now through visiting the telegraph.p3k.io website so I basically don't have any syndication links happening when I syndicate to Twitter, I just do back syndication links to the original content. Definitely not the way I want to keep doing it ? haha
#tantekso I'll say it once again, I strongly recommend *against* giving your phone number to any silo, because 1) they default to increasing your vuln with SMS account recovery, and now 2) their account recovery may actually be *more* buggy when you give them a phone number
#tantekwith silos, be lazy, and give them as little info as possible, because you can assume more users are lazy than not, and therefore the "lazy user" user flows are more tested, ergo, more reliable, ergo, more efficient (time savings) for you when "things go bad"
#[miklb]eddie gotcha. I was using the webmention.io jekyll plugin and brid.gy in my Rake deploy script. Was less than ideal, but that plugin has gotten a lot of love since.
KartikPrabhu joined the channel
#snarfedtantek: agreed! please also encourage non-SMS 2FA also when you say that though. it's a huge step for individual security, and vast majority of silos support non-SMS 2FA.
#tanteksnarfed, agreed, and we should document those that do support non-SMS 2FA and *simple* instructions for setting it up
#tantekbecause those same silos have mis-designed UI that misdirects users into SMS 2FA including SMS 2FA recovery by default
[chrisaldrich] joined the channel
#[chrisaldrich]reads logs and can't help but think re: Tantek quote above: "The first rule of decentralization: We don't Tweet about decentralization. The second rule of decentralization...."
#Loqisnarfed has 21 karma in this channel (320 overall)
#Loqi[tantek] Re: "A solution using pure CSS for categories might be feasible if we could somehow layer JavaScript content over a common JavaScript and non-JavaScript CSS base."
I think that's a good approach. The categories themselves can be fairly mechanicall...
[manton] joined the channel
#[manton]So, thought I'd redirect /next-iwc on the wiki to the Austin page. But now the "Berlin" link in the sidebar goes there. ? Is it possible for me to update that sidebar for events, or does someone else need to do that?
#tantekright, this is just to produce something that other existing yml consuming tools could "just" re-use
#ZegnatI think it is pretty OK as far as JSON->[Object in your favourite language]->YAML goes.
#tantekthat way more existing tools can also consume the output of parsing an HTML+mf2 page
#tantekbiggest questions I'd have are, 1 can you nest objects, and 2 are plural valued properties as easily listed as just multiple lines with the same key or do you have to add more punctuation?
#sknebelIf JSON isn't valid YAML directly it's pretty close to being so
#tantekI think the point would be to produce something more "clean" than just a direct mapping
#ZegnatOh, right, I think there are some JSON documents that are valid YAML, that’s true!
#ZegnatI always found YAML to be a bit of a mess because of the huge variations in how you can write things. Have also had trouble finding proper libraries to handle it in PHP, most only implement an older version or a subset of the YAML spec :/
[miklb] joined the channel
#ZegnatReading the Wikipedia article, any valid JSON document should also be a valid YAML 1.2 document
#ZegnatSo in a way, mf2 parsing already creates YAML ;)
#ZegnatAs I said, YAML is actually a pain. It only looks nice in front matter because people mostly stick to a flat associative array with the occasional nested array ;)
#aaronpkI actually use a relatively simple version of yaml for my storage. I store things as single values unless there are multiple, and I have code that knows how to deal with the difference when I retrieve values
#aaronpkso my storage files end up looking very close to the jekyll frontmatter yaml
#ZegnatHuh. Reading some random parts of the spec. I *think* YAML can have arrays as associative array KEYS... i.e.: {['this','is','a','key']:['for','these','values']}
#ZegnatI doubt the parser I use for PHP accepts that
#Zegnat“YAML places no further restrictions on the nodes. In particular, keys may be arbitrary nodes”
#Zegnat“A YAML node represents a single native data structure. Such nodes have content of one of three kinds: scalar, sequence, or mapping.”
#ZegnatAnyone ever used an associative array as a key within another associative array? :P
#ZegnatI remember some of this stuff. I wanted to make my file storage 100% YAML. Most parsers already had issue with multi-YAML documents (a file can have several YAML documents split by “---”), I also had some differences between different libraries exports and imports. Gave up on YAML.
#ZegnatI seem to recall YAML 1.0 was a bit more what people think YAML is. But I have never looked into their actual parsing rules so no clue how wild the nesting can get.
#aaronpkI *think* YAML has always had the ability to name different parts of the document and refer to those named sections, but not sure
#ZegnatI guess it has more lenient escaping rules for strings, which arguably makes it easier to author than JSON. And writing arrays as an actual visual list is nice. But otherwise? Write JSON.
#ZegnatEven better: don’t write serialisation formats by hand. Ever. Not JSON. Not YAML. Not XML.
loicm joined the channel
#ancardaWhat do you mean, Zegnat? By hand, as in never write yourself, always use a function like json_encode()?
#ancardaI’d appreciate feedback on my h-card, or anything else, if there’s something I should change? One validator I ran it through did suggest adding u-email, I’ll do that soon. My domain is https://markdain.net/
#ZegnatI’d include a link to https://markdain.net/ within the h-card too. To make it clear that this is the h-card that goes with the domain. (In case you ever add another h-card to the front page, e.g. of a company or in a person-tag)
#tantek.comedited /POSSE (+153) "provide current working link for diso interview, and internet archive link for original" (view diff)
#tantek.comdeleted /FHIR "content was stub that did not relate to indieweb in anyway, "better left to Wikipedia" per [[wikifying#What_should_not_go_on_the_wiki]] (enwp.org/FHIR). Recreate only if going to define it in terms of its indieweb specific relevance"