2013-11-14 UTC
benprew, ozten, andreypopp, _6a68, tilgovi and bnvk joined the channel
bnvk, benprew, josephboyle and scor joined the channel
# 01:14 KevinMarks_ "There’s a feed in the middle showing all your friends’ posts, and some other random stuff along the sides that no one has ever looked at" made me laugh aloud # 01:41 _6a68 KevinMarks_: that blog post made me laugh and laugh # 01:42 _6a68 but really, I didn't think it was fair to blame their troubles on the database # 01:43 XgF I think lots of people have ran to (Mongo|Redis) because RDBMS are hard (but powerful) and OODBMS are rare # 01:43 _6a68 sure, but if you don't foresee that you'll need to manage consistency and relations among denormalized data blobs inside your app, you are in for a world of hurt # 01:43 _6a68 regardless of which implementation of non-relational key-value store you pick _6a68 joined the channel
paulcp joined the channel
# 02:22 KevinMarks_ someone in Sarah Mei's comments pointed out that MapReduce is how you fix things in Mongo, which is one way of looking at it pfenwick, paulcp, snarfed, bnvk and b0bg0d joined the channel
tpinto joined the channel
tantek and b0bg0d joined the channel
# 03:33 tantek kevinmarks, sounds like a patch to a problem approach # 03:50 KevinMarks_ well, mapreduce is a different way of thinking - you construct the DB you need from the unstructrued data you have # 04:05 Loqi Ok, I'll tell him that when I see him next tpinto, b0bg0d and tantek joined the channel
tpinto, vrypan, neirpyc, caseorganic, skinny, cweiske, LauraJ, Jihaisse, pfefferle and tantek joined the channel
tpinto joined the channel
XgFgX, skinny and lukecyca joined the channel
andreypopp, b0bg0d, tpinto and barnabywalters joined the channel
# 09:57 Loqi barnabywalters: tantek left you a message on 11/13 at 11:16am: very cool! looks like it's picking up the u-url u-uid no problem. However the dt-published and dt-updated is "2013-11-13T" missing the time from: <span class="dt-published published dt-updated updated"><time class="value">6:01</time> on <time class="value">2013-11-13</time></span> which should work per: http://microformats.org/wiki/value-dt-test-YYYY-MM-DD--HH-MM # 09:58 barnabywalters tantek: indeed, turns out that was an undiscovered bug from ages ago, value-class time components with single-digit hours weren’t being matched correctly. fixed. tpinto, josephboyle, skinny, melvster, jancborchardt, pfefferle, bnvk, scor, brianloveswords, LauraJ, singpolyma, ozten, friedcell, snarfed and tantek joined the channel
LauraJ and Acidnerd joined the channel
tantek joined the channel
# 15:41 Jihaisse barnabywalters: hey, I've juste send a webmention to you, and, it look strange on your blog # 15:42 barnabywalters did that get fixed? or does the wordpress plugin registry need to get updated or something? # 15:50 tantek I'm wondering if this bug is a sign of taking the wrong approach with existing wordpress blogs # 15:50 tantek that is, since existing wordpress blogs already have hentry, why not just use that (or even enhance it with rel=author) warden joined the channel
# 15:51 tantek point is, due to the backwards compat support in phpmf2, the existing class="hentry" element in WordPress posts should "just work" for comments, reply contexts etc. # 15:52 tantek and if it's not, then why not fix it instead of adding even more markup for microformats2 # 15:54 Jihaisse I'm not sure, but I think that SemPress is a theme developped for mf2 support # 15:54 tantek Jihaisse - I understand - however it seems like a lot more work # 15:59 pfefferle it is some work, but dealing with the current mf implementations is difficult too, because the always use hfeed for example... # 16:02 tantek interesting - so is the current wordpress uf markup is too problematic to bother fixing? # 16:03 pfefferle it uses hfeed for single-posts and there is no hook that let's you add some informations inside the hentry # 16:05 tantek pfefferle - so how did such problematic implementation make it into "core" and not get maintained? ozten joined the channel
# 16:06 pfefferle the hfeed thing is not in the core, but it was part of a popular theme which everyone copied # 16:07 tantek hmm… search engines and sites like readability parse the hentry markup # 16:07 pfefferle most developers stayed with it because of seo stuff... nothing more # 16:08 tantek but even for that your markup has to be correct # 16:08 tantek maybe there wasn't a tight enough feedback loop for people to notice? # 16:08 pfefferle it is correct if you ignore the surrounding hfeed on the permalinks # 16:09 pfefferle Jihaisse that's too late... it is part of nearly every theme including the automattic ones # 16:10 tantek submit the patches - never know what will happen - maybe others will help fix it # 16:11 Jihaisse I'm not good at mf2, so I don't know if it's ok or not # 16:13 tantek pfefferle - the hfeed container itself should be harmless for any consuming clients looking for an hentry # 16:13 pfefferle only on overview pages so no themer had to add it by hand # 16:14 tantek so how do you go about fixing things like that in WordPress? _6a68 joined the channel
# 16:15 pfefferle i remove it via regexp and hook into the body_class() filter # 16:16 pfefferle tantek the most implementations use barnabywalters php-mf2 parser and check only on the highest level for h-entries... that is a problem # 16:17 pfefferle i tried it without adding the fix to SemPress and my mentions doesn't work on most indieweb pages # 16:18 pfefferle will add a page about wordpress-theming to the wiki if i have some spare time # 16:19 barnabywalters personally I suspect that mf support built into wordpress is a bad thing — it should be added by hand into themes # 16:19 barnabywalters whilst reader-building I have found plenty of wordpress blogs which have almost-complete hentries, or the shells of hentries # 16:20 barnabywalters but always missing things, most of the time probably due to people changing the theme without knowledge or awareness of mf classnames tantek joined the channel
# 16:23 pfefferle the history is, that there is only one method that returns class-names and it is for the post-container # 16:25 pfefferle and it wouldn't be that bad if the returned the hfeed via this function # 16:25 tantek pfefferle - I'd say yes, try submitting as patch to core. We should at least incrementally cleanup WordPress as we can. # 16:26 tantek barnabywalters - my guess is that changes were made to the core to split responsibility between it and themes and that may have been a mistake. # 16:26 pfefferle but that would break every wordpress theme out there so I don't know if they would accept it... will try it anyways... # 16:28 pfefferle if theme developers doesn't change their themes it would show up twice ozten joined the channel
ozten2, tilgovi, caseorganic, b0bg0d, benprew, ryana, josephboyle, ozten, bnvk, paulcp, snarfed, squeakytoy, squeakytoy2, paulcp_, npdoty, _6a681, jgraham909, thatryana, tpinto, _6a68, barnabywalters, vrypan, singpolyma, poppy and friedcell joined the channel
abrereton and tantek joined the channel
# 21:36 tantek perhaps he can help with the wordpress conundrums discussed this morning # 21:37 Loqi Ok, I'll tell them that when I see them next paulcp and josephboyle joined the channel
# 21:58 tantek aaronpk interesting. I've written code specifically to address a few of those concerns but not made a big deal of it. It's in CASSIS, let me see if I can document it more explicitly. # 21:59 aaronpk yea, but still involves wiring up an interface and javascript and figuring it all out # 21:59 tantek well, sorta. my interface gives me the character count from the server - no JS # 22:00 tantek anyway I have a CASSIS building block function you can use in PHP or JS for that. # 22:18 tantek aaronpk, there you go - I think that even does more than what you asked for # 22:18 tantek I for one find the $note exactly fits $maxlen characters when retweeted functionality quite handy - and often hit that specific limit # 22:19 tantek like craft my tweets to use 140 characters *when retweeted* # 22:19 tantek yeah I figured these were tricky / useful enough functions to share as CASSIS building blocks for others to re-use server-or-client-side # 22:20 tantek I've done a lot of debugging of them for sure # 22:20 aaronpk so the context of that quote was a conversation with dreeves of beeminder, they are currently publishing a feed of user-visible improvements to their app on twitter at https://twitter.com/beemuvi and would prefer to have that hosted on their own site # 22:20 tantek and I value posting notes that are losslessly retweetable # 22:20 aaronpk the challenge is making an interface that is at least as functional as twitter's # 22:21 tantek (not counting all the cheesy furniture blocks) # 22:22 aaronpk and it's especially challenging when you're talking about taking time away from developing core product to set up an interface like that # 22:22 tantek well for the indieweb, this kind of interface challenge *is* core # 22:22 aaronpk a much shorter path for them is PESOS, continuing to publish on twitter and pulling down a copy of their tweets # 22:23 tantek PESOS is always easier upfront, more vulnerable/fragile long term # 22:23 aaronpk that's kind of a common theme in general... do the easy thing now and fix it later, or spend longer on it now? # 22:24 tantek so that methodology fails because the "fix it later" starves due to distraction by too many "easy thing now" # 22:24 tantek it's empty/hollow/unreliable reasoning/planning _6a68 joined the channel
# 23:02 tantek has anyone implemented posting in the future on their indieweb site? # 23:03 tantek (like the feature in WordPress, I think, to schedule future posts) # 23:03 tantek I have some ideas around it that I've been meaning to explore # 23:04 snarfed aaronpk, tantek: the common term for that is "urgent at the expense of the important" # 23:04 tantek e.g. storing it in "future time" in my flat file structure, and then having my index/stream code only show posts that have been "published" in the "past" # 23:04 tantek snarfed - good to know - that's probably worthy of documenting as an anti-pattern # 23:05 snarfed eh yeah, maybe. it's obviously extremely broad, way bigger than indieweb or even tech in general # 23:06 snarfed known solutions are discipline, strong leadership decision making, clear prioritization processes (although getting those right is tricky, often can do more harm than good) friedcell joined the channel
# 23:08 snarfed basically prioritization and weighing the value of quick fixes vs long term investments # 23:09 tantek ironically this is related to the "at some point" antipattern as well andreypopp joined the channel
# 23:09 snarfed and prioritization is hard in so many ways. 1) making real priority decisions in the presence of non-obvious tradeoffs, 2) getting consensus, 3) actually following through and working on highest priority stuff first # 23:10 snarfed there's always tons of lower/medium priority stuff that's so tempting to do! # 23:11 tantek snarfed - in a community context I allow some amount of others' "urgent" items to dictate what I work on (especially if I can answer quickly) figuring that when I later have something *important* and ask about it - the likelihood of reciprocation is greater. # 23:11 tantek the dynamics of prioritization are more complicated in a community context # 23:11 tantek I look at it as more of an investment than tax # 23:12 snarfed also, urgent vs important obviously isn't a dichotomy for solving any individual problem. you can do both. # 23:12 snarfed it's only an antipattern when there are lots of things to do, and you're overdoing the urgent alternatives instead of the important/long term ones # 23:18 tantek snarfed - right - that first URL of the two above showed a 2x2 matrix of urgent vs important caseorganic joined the channel
# 23:22 tantek thanks for the "common term" snarfed - that was the key to further understanding this phenomenon and then documenting it. andreypopp joined the channel
# 23:34 tantek aaronpk - do you have a system for posting to the future? # 23:37 aaronpk I don't. I've never really felt the need to I guess # 23:37 aaronpk I have a hard enough time writing stuff in the past :) # 23:38 tantek however the concept does have some specific uses/instances that you *do* care about, e.g. posting future trips # 23:38 aaronpk yeah, I have both future events and future trips right now # 23:38 tantek which are just special cases of future posts in general # 23:38 aaronpk but I want both of them to be visible before their dates # 23:40 aaronpk yeah I'm not happy with the way they look right now, cause they are in reverse chronological order meaning the farthest future event is alwasy on top # 23:41 tantek ok I have some ideas I bounced off of benwerd last night, I think I'll try coding something up as a way to explore them # 23:42 tantek because it sounds like there isn't much other experience with such "future posts" - so time to get some _6a68, b0bg0d, caseorganic and _6a681 joined the channel
@dwhalen77 RT @peterarp: âMesmerizing story w/universal appealâžâTriumphant & delightfulâžâ #mustread âž GOOD LIKE THIS http://www.amazon.com/Good-Like-This-Peter-Arpesella-ebook/dp/B00GBN7RDI/ref=la_B00D9F9HTS_1_1_title_0_main?s=books&ie=UTF8&qid=1384368856&sr=1-1 #IndieAuth… (twitter.com/_/status/400788100042850304)
KevinMarks_ "There’s a feed in the middle showing all your friends’ posts, and some other random stuff along the sides that no one has ever looked at" made me laugh aloud
@kevinmarks @dhh if document centric apps are important, why use an ORM based framework like rails that forces documents into sql tables?
bret !tell tantek I heard back from harry halpin and he has since updated http://www.w3.org/2013/socialweb/report.html
Loqi tantek: bret left you a message 42 minutes ago: I heard back from harry halpin and he has since updated http://www.w3.org/2013/socialweb/report.html
Jihaisse http://wordpress.org/plugins/relsyndication/ new plugin for wordpress
barnabywalters tantek: indeed, turns out that was an undiscovered bug from ages ago, value-class time components with single-digit hours weren’t being matched correctly. fixed.
@BarnabyWalters #indieweb goal: by 2014-01-01, no longer be using https://twitter.com/ to read+reply to my friends’ content. (http://waterpigs.co.uk/notes/4T3FSd/) (twitter.com/_/status/401008278056235008)
pfefferle except with SemPress or a plugin we are currently working on https://github.com/pfefferle/wordpress-uf2
tantek pfefferle do you know Eric? https://twitter.com/ericnakagawa/status/370604732689027072
@ericnakagawa @t hey tantek, are the micro formats like h-entry a valid place needing work/support? Am considering building validators for wp themes.
tantek.com edited /Twitter (+1270) "/* Challenges */ note CASSIS building block functions for enabling implementation of better UI" (view diff)
aaronpk so the context of that quote was a conversation with dreeves of beeminder, they are currently publishing a feed of user-visible improvements to their app on twitter at https://twitter.com/beemuvi and would prefer to have that hosted on their own site
@cassie917 Successfully facilitated over 350 data chats w. scholars at Carver today! #ownyourdata (twitter.com/_/status/401115737953423361)