butteryou'd have to run mbox2md or something to spit out individual messages.e it indexes on an as-seen basis, primarily symlinks used to reconstruct threads, and some per-address index-dirs
shaners_Maybe there could be a two column comparison table (like when a company is compary its product to another company's product. [or different pay levels of a product]).
shaners_but my main point / concern here is that the name of the page (and the content of the page) doesn't really acknowledge the trade offs. OTHER PAGES on the wiki do (to a degree), but A: need more attention (from me and others) and 2: are other pages :(
gavincalso, I'd worry a bit about ballence... have you seen the responce? there is a frighting number of people that think web applications or sites HAVE to have databases
tantekkylewm - you yourself remarked at how surprised you were at the performance of my site based on dynamically building pages from a flat file storage.
tantekgavinc - the response has been 99% emotional. "Is this for real?" "Words cannot describe" "wut" "stay away" etc. and thus indicates likely presence of Cargo Cult Programming.
tanteksince people defend with emotions (rather than reasons) when they've adopted something for Cargo Cult reasons, rather than actual thoughtful rational reasons.
shaners_but mostly, I don't care enough (about this db/flatfile debate) to care, which is why i've not put any effort into the db side of the conversation on the wiki. i've always thought that it wasnt worth my/our time to even talk about.
tantekshaners - I only found it worth documenting because I noticed a pattern of problems (around using databases for personal site content) that no one else had actually collected into an article - so I collected it.
tantekkylewm - I only brought it up because a) you remarked at last HWC how flat files were giving you perf problems and then b) I noticed you switched (back) over to using a DB for your site content storage (edit today)
kylewmso my rationale was: i was thinking about how not to have 9 zillion small files, while still making it easy to make small atomic edits (like adding a reply context asynchronously), and keep indexes searchable by type and tag
gavincDealing with databases is exactly why gavin.carothers.name back running wordpress on wordpress.com and not home rolled :( Databases are enough fun at work all day
tantekgavinc - fascinating, it's due to your data being stuck in a database that's preventing you from hosting your own site on web hosting instead of wordpress.com?
gavincnope, it was when I thought about setting up paggerduty for mine (and my wife's) blog when they went down when I screwed up a postgres upgrade :D
gavincand NOW, yeah getting back out of the database with comments etc is much more annoying then getting everything into wordpress xml. So likely just do it again via a wordpress xml export
acegiakso I made a font of my handwriting and then made a webpage with the "handwritten" message on there with pictuers and had a song fom when were teenagers playing
shaners_One could have an app for CRUD of a post type (or all posts), and a separate app that does just in and outbound webmentions, another one for posse sydication etc.
shaners_acegiak: Meh. It's cool that WP has a plugin architecture to add new features / functionality. But ultimately they all get loaded into big app, at least from a URL sense of it.
shaners_I.E., I can't drop a WP plugin into any other kind of app to use it. I have to use WP. Bc it's integrated at the app level, not the URL/HTML/HTTP level.
shaners_one of the (many) knock on benefits of this approach is that we can use other people's software (as long as it reads or writes HTML+uf2 over HTTP). and vice versa. someone making php software could run our ruby app/s alongside theirs and interop at the HTML/HTTP level.
KevinMarks_RDBMSes, though much more reliable than most user-written transaction processing code, are not nearly as reliable as a basic Web server pulling static files out a file system. Prepare to hire a half- or full-time database administratorÂ
butterit wouldn't shock me if a lot of google employees don't even like the products, but do it because it's how they can make a living. and want decentralized/p2p stuff for their own sites
petermolnarShanley 13 h "White men CANNOT BY DEFINITION lead the "independent web" movement to save us from the surveillance machines they built & profit from." Seriously? That woman is nuts.
butterit just happens to be a better databse for 99% of use-cases where you don't need extremely sophisticated querying-capability and want much easier backup/redundancy/basic-manipulation ability.. via all the existing UNIX tools, rsync/scp/nfs etc
KartikPrabhubutter: speaking as a non-dev, I think of file system and databases as different things. technically they both store things but from a usage pov they are quite different
tommorrisThe only thing I'd suggest we might learn from the database-antipattern twitter shitfest is that when there's a topic where people in the community differ (I use postgres, don't have an issue with it), it's probably best to, say, post the bulk of the content on our own sites and then link out. so instead of a page called "database antipattern", we just have
KartikPrabhuthese negative aspects might be obvious to people who deal with them regularly, for beginners even a subjective POV on this helps (speaking from personal experience)
tommorrisI think if we're to be a positive community, to make sure our rhetoric doesn't distract from the overall goal of encouraging people to build stuff.
tantekand I documented it there as I saw a bunch of problems happening with database-backed indie web sites and didn't see *any* collection of such problems
butteroh, ORACLE forked/bought your db, devs raged, forked a community verison which won't eat your conf files that are a 0.0.1 behind so the server didn't start?
petermolnarMySQL is not a bad db; the MyISAM engine was, InnoDB is way ahead; also keep in mind that MySQL is for speed, not for exceptional stability, like Oracle
tommorris"The Blackhole engine is a no-op engine. Any operations performed on a table using Blackhole will have no effect. This should be born in mind when considering the behavior of primary key columns that auto increment." is possibly the greatest documentation anyone has ever written.
petermolnartommorris: the blackhole will not write anything locally but it will send the replication down, so the slaves can store the data; thus you can offload write to a slave, for example, for an archive table
alanpearceKartikPrabhu: yeah, the wording is a bit ambiguous. It's more like an 'object store': attributes and child objects are stored, but it's all binary JSON or similar.
LoqiHaxxa: finalcut left you a message 1 day ago: - you might not have url rewriting turned on in apache if you are getting a page not found error with the begin page. I was just able to recreate it by removing url rewriting
tommorris"Host your own content. Don't let these silicon valley douchebags con you into giving your personal property to them to do with it as they please. If you want to post things on Youtube, Soundcloud, etc. Fine. But you should also host it on your own hosting account somewhere using a wordpress blog or whatever. Also, keep copies of everything in your physical
tommorris"The Internet Archive is also a safe place to store media long term. I have been posting media on line since 1997 when I had a live Real Audio stream of Yeast Radio. The only place that hasn't deleted some or all of my content during that time is Internet Archive."
bretis there one too many "syndicate-to" key names in this string or is that a thing? syndicate-to=syndicate-to=twitter.com%2Faaronpk%2Cfacebook.com%2Faaronpk
ooland, dietrich, alexhart_, paulcp, eschnou, brianloveswords and npdoty joined the channel
bearto be pendantic, urlparse.urlparse() that returns a tuple with the query as a string, it's urlparse.parse_qs() that returns a dictionary and yes, multiple items with the same name are returned as a list
kylewmcontext of the parameter discussion is we were wondering if micropub should be using multiple entries for the same key instead of comma separated values
tantek.comedited /manifesto (+326) "clarify no manifesto, add section header other manifestos, add Independents Day Manifesto as original art, note in summary no indieweb(camp) manifesto" (view diff)
LoqiA manifesto is "a published verbal declaration of the intentions, motives, or views of the issuer, be it an individual, group, political party or government" according to Wikipedia[1] - in summary there is no IndieWeb(Camp) manifesto http://indiewebcamp.com/manifesto
LoqiA manifesto is "a published verbal declaration of the intentions, motives, or views of the issuer, be it an individual, group, political party or government" according to Wikipedia[1] - in summary there is no IndieWeb(Camp) manifesto http://indiewebcamp.com/the_indiewebcamp_manifesto
tanteklet's see those wiki stub pages people! if you use an acronym / jargon word/phrase - please add a short stub to it! It's easy: http://indiewebcamp.com/start_a_page
binbastii tend to agree when it comes to the inline link format, but that's nice to write. discourse does a great job of taking it as input, but transforming it to the more readable page bottom notation
bret.iocreated /gem (+166) "Created page with "{{ stub }} <dfn>Gem is the ruby language package manager.</dfn> It can be used to managed ruby packages and provide access the the packages at https://rubygems.org."" (view diff)
tantekthe thing about HTML is that it has quite minimal use of punctuation: <..> and <../> to start with, then other parts are readable letters - tag and attribute names
tantekI think that's the key - minimal punctuation that you have to commit to memory, and then use of text (which is easier to recall than random series of punctuation)
bret.iocreated /npm (+240) "Created page with "{{ stub }} <dfn>npm is the "node package manager" and is can be used to manage javascript package for any type of javascript project (not just node).</dfn> It provides quick ac..."" (view diff)
binbastii think when you're writing it all day, typing less matters more to many people, which is why markdown and textile were invented in the first place
tantekbear, the thinking was to first design what would "make sense" and look reasonable in *plain text* and then see if it was something consistently parsabel
Loqitantek meant to say: bear, the thinking was to first design what would "make sense" and look reasonable in *plain text* and then see if it was something consistently parseable
tantekbear - what's sad is that I think the crappy (line-noise) parts of markdown were design with parsing considerations first, and plain text considerations second.
butteri agree with 95% of your wikipages, great decentralized/personal-sovereignty ethos, but the structured-data stuff as microformats? i'd take RDFa or microdata+Schema.org over that
tantekbutter - why? do you like typing more than you need to? (RDFa microdata) and do you like trusting big companies with overdesigned data models? (schema)
butternot really a fan of RDFa either, but at least property-names are 3rd-party/decentralized-extensible not decided upon by a couple google/yandex employees
butterKartikPrabhu: most tools i use consumer Turtle, which has RDF inside. for a basic 'finder'/directory-browser https://github.com/deiu/warp , and so on
tantekthe funnything about all this single-page-site development fad is that once it's long over, it will be difficult to research/understand what the heck happened, because none of the single-page-site apps will have any resilient permalinks to view in the future.