#Vendanthe last time I tried to optimize php fastcgi, I ended up writing the fastcgi serverloop in php itself, cause php, even with opcode caching and all, still read the php code off the disk for every request
#aaronpkcoulda just put the php code in a ramdisk :)
#Vendanstill, you are instantiating the objects, just to tear them back down few milliseconds later, and then instantiating them the next request
#VendanI was handling 6000+ requests per second using a simple framework when <?php echo "hello world"; ?> was only doing 2000 on the same hardware
#aaronpkjeez. I hope you actually had a good reason for that kind of optimization
#Vendansomething in my is still like "oh, I could do so much better" and the other part is like "What's the point, it's 100000x more then I need at this point"
#LoqiThe WordPress Outreach Club is a group of active IndieWebCamp participants who reach out to individuals already running WordPress to add IndieWeb functionality to their existing sites https://indiewebcamp.com/WordPress_Outreach_Committee
#Vendanone of my thoughts, for when broadband starts picking up, is to package up a small image for something like CHIP or a raspberry pi, that'll be your indieweb
#GWGAnyway, I'm still, to get back on room topic...
#Vendanbut yeah, I can respect the whole convert users that are on wordpress already, just nowhere near willing to install it on my server or deal with PHP outside of work
#GWGI'm trying to design better display code. So, basically, I've built a system to store and retrieve information about a post...I want to display it with the appropriate markup.
#Vendanso if I want a different style of post rendering, for stuff like reply-context, I can just use the same structs and pass a different template object
#Loqigives Vendan a different style of post rendering
acegiak, shiflett and nloadholtes joined the channel
#Vendanoh, and bret, feel free to ping me if you don't understand anything, I'm horrid at commenting and documentation
#VendanI'll probably start going over the parser, just cause it's most likely to get used elsewhere
#bretVendan: i've never written a parser (other than regex stuffs), and still learning go. ill be constructive promise :)
#atomiculesHello. Just dropping in to say that since starting with Indieweb at the start if this year, I've now implemented POSSEing links, photos and notes
#voxpelliatomcules: you can always export all of the mentions from my endpoint, so it could be a way for you to get started, and then you can move to your own when you have one
#atomiculesvoxpelli: Good idea. That might be a plan.
#voxpellibeen thinking to maybe add a passthrough options for the webmention pings so one could have my endpoint forward a ping to another endpoint – to ease the transition
#LanceyWorkatm i've just checked if the atref is a valid domain name, did a get on the site, searched for rel-me links to twitter, and pulled the handle off that
#LanceyWorkthere is probably a better way to do that
#tantekLanceyWork - there's a bunch of work and previous sessions on the wiki that folks have discussed re: auto-completing names/URL of people including @-names
#tantekmarcthiele++ for co-organizing IndieWebCamp Germany and helping with venue and all the other things that helped make IndieWebCamp Germany awesome!
#tantekis seeing if he can ++ someone twice within minutes
#GWGI am back on trying to figure out how to get around WordPress and it's single content hook being inside e-content.
#tantektommorris: I don't really see the usefulness of such a page /standards_efforts as it is now - and frankly, I've seen far too many such random lists of projects/standards on wikipedia that are nearly useless (but I have no desire to get into a wiki-edit fight with anyone there about)
wolftune joined the channel
#tantektommorris: the #microformats wiki is a more appropriate place to just do arbitrary documentation of prior standards work/attempts/proposals like that.
#GWGtantek: Are there any microformats that can be nested inside e-content?
#tanteksure - they all can - you can even put properties inside properties like that and it will work fine
#tommorristantek: the existing list was more focussed on the standards that indiewebcamp was using. I kind of wanted a list of standards we'd looked at, so we could document them. if it's out of scope, happy to delete/revert or move
#tantekI think it's out of scope for #indiewebcamp because it's not helpful for building
#tantekthat kind of general documentation of past efforts is much better suited to microformats.org/wiki
#GWGI think I may have to give up and move the in-reply-to back into e-content.
#tantekalso - that kind of framing with "implied moral of this list" will likely just invite more defensive talk/conversations about standards no one is actually building on for their own sites - and thus be a distraction / noise amongst the rest of our more focused / productive discussion in IWC
#tantekso I guess two things: 1) I don't think such a general (no specific category/focus) list of past efforts is useful on IWC. 2) there may be some instances where it is worth documenting *specific* such efforts on their own pages, e.g. by listing the modern replacement that supersedes the prior effort
#tanteke.g. see /OpenID which evolved from something we depended on, to something that was dying, to something we replaced with a simpler alternative
fourtonfish and snarfed joined the channel
#LanceyWorkok, that just leaves converting a reply url to a twitter context if possible
#tantekfor 2) instances of documenting *specific* such efforts on their own pages, there should be some real world reason for doing so
#tanteke.g. one or more of 1. indieweb community folks support it live on their site (thus non-empty IndieWeb Examples section). 2. indieweb community folks *tried* supporting it, and gave up, for *specific* reasons (must be documented). 3. indieweb community folks looked it, but didn't bother to even try supporting it for *specific* reasons.
#tantek4. indieweb community created something that *supersedes* the prior effort, and thus it helps to create a page for the prior effort just to at least redirect people to the new/modern thing that supersedes it.
#tantek5. more than one person in IRC has asked about the past effort, and there was sufficient response discussing why the past effort is not in use by the indieweb community to document a specific Criticisms section for that specific past effort
#tantekbut the larger point is, simply yet-another-unfocused-list of previous efforts is not helpful to anyone and may actually incite more noise than signal
#tantekLanceyWork what do you mean by "converting a reply url to a twitter context "?
#LanceyWorkif it's a tweet permalink, or something that's been tweeted (ie posse'd from another indiewebsite), then specify the tweet as in_reply_to_status_id
#LanceyWorki was just gonna use the regex at the link i posted, but thanks!
#tantekyeah I wrote most of the /Twitter page incrementally while implementing the functionality on my site (and in my software /Falcon ) - though at some point I started extrapolating so now the /Twitter page goes beyond what I've implemented so far (but serves as a checklist for my future coding)
#tantekI wrote tw_url_to_status_id semi-defensively so you could throw any URL at it and have it return 0 or a valid status ID, with nearly no chance for false positives
#tantekmarcthiele++ for co-organizing IndieWebCamp Germany and helping with venue and all the other things that helped make IndieWebCamp Germany awesome!
#tantekaaronpk - at some point you way want to consider forking the PubSubHubbub spec, since specs work best when their editor(s) are actively selfdogfooding what's in the spec.
#aaronpki'm reaaaalllly trying to avoid that if possible
#Vendanrealistically, it needs to be 2 specs. One spec for hub-sub, one spec for pub-hub
#tantekaaronpk: it is good to avoid forking a spec until you feel you must (i.e. disagreement with nonselfdogfooding editors), or when your contributions approach ~50% of the spec.
#tantekone of the good reason to do so, if you're willing to rewrite all the content, is to license it more liberally - since according to Harry the existing editors / Google didn't want to contribute the spec to W3C or sign an IP agreement to allow its re-use.
#Vendanthen are you going to say that my site isn't PuSH 0.4?
#Vendanmy hub doesn't, and isn't likely to support publish via any form of http request
#aaronpkmy contributions definitely are nowhere near 50% yet
#aaronpkVendan: your website supports PuSH 0.4 in that subscribers can read the spec and subscribe to your site. but no you have not made a 0.4-compliant hub
#aaronpksuperfeedr and switchboard and phubb are all meant to be generic hubs
#aaronpktantek: wrt IP, is the spec primarily the written description of it, or is the spec the mechanism?
snarfed and julien51 joined the channel
#aaronpkwith food recipes, you can't copyright a list of ingredients, but you can copyright the written description of the recipe
#tantekaaronpk, with IANAL disclaimer, for copyright primarily the written description, for patent the mechanism
#aaronpkit bugs me because at the top of the 0.4 spec it says "To dramatically simplify this spec in several places where we had to choose between supporting A or B, we took it upon ourselves to say "only A", rather than making it an implementation decision."
#aaronpkand then they didn't do that in a bunch of places
#tanteke.g. Hixie was able to rewrite HTML in his own words, and put a liberal copyright license on it
KartikPrabhu joined the channel
#tantekaaronpk - that sounds like that "bunch of places" are all issues to be filed, where you say, per your spec methodology, I say choose A (or choose B - whichever *you* aaronpk think is the better answer)
#aaronpkthat's what this thread is currently about
#tantekin face, just file pull requests that make the choice you think is best, and cite the design methodology
#gRegorLoveAnyone know of an h-event to iCal parser/converter?
#tantekvoxpelli: the spec should clearly cite the copyright and patent policies, preferably by linking to a generic license rather than making up their own
#julien51wowo, tantek, what I’m reading here baffles me! lots of lies! It also saddens me that you’d say things like “ since according to Harry the existing editors / Google didn't want to contribute the spec to W3C or sign an IP agreement to allow its re-use.”. Don’t you know how to reach me?
#julien51I emailed “Harry” a dozen times, he *never* got back to me :D
#tantekjulien51: sigh - very sorry to hear that - I trusted that Harry was thorough about this, and apparently I was mistaken. Apologies.
#julien51well, think about going to the source whe you know the source…
#julien51honestly, if you guys want to fork PubSubHubbub , go ahead
#julien51I’m in no way interrested by the publisher side of things
#tantekjulien51: I'd rather not fork if not needed
#julien51tantek: yeah? they wrote previous versions of the spec
#tantekjulien51: it's possible that Harry asked them first (since they're at the top), got a "No" answer, and then gave up because all editors would have to agree to contribute / openly license
#julien51I did not think it was fair/right to remove them when adding to their work, rather than forking
#tantekcontinuing: however, if 0.4 is a rewrite of 0.3 + new features you wrote, then it's more appropriate to list them in an informative Acknowledgments section as authors of the previous version
julien51 joined the channel
#julien51ragequit : lol :) sorry I my english AND my internet connections aren’t as good as yours :)
#tantekaaronpk - I'd suggest just keep filing specific pull requests for specific issues, and help move them along in parallel
#julien51as for 0.4, there’s a lot of new things I wrote (you can see the diffs easily on the GH repo), but in any case *they* fostered the field and the spec, so no, I won’t take them out.
#tantekjulien51: it's not clear from the header that you're the primary author/editor of 0.4
friedcell joined the channel
#tantekalso if they're not actively editing / writing the spec, listing them as Previous Authors (or Editors) is more accurate
#voxpelliregarding the PuSH-issue – seems to me like the current disagreement is mostly around where to add a standardized way rather than if there should be a standardized way? rather than just making a draft first and deciding where it belongs later on?
#KartikPrabhunow I wonder if adding Push to IndieMark was a good idea cc: tantek
#voxpelliespecially the "Notify the hub of new content" section
#tantekLet's act on the assumption that the PuSH editors are reasonably amenable to receiving such clarifying pull requests and will incorporate accordingly
#aaronpkokay so here's an example, currently superfeedr and google support publish requests like this: "hub.mode=publish&hub.url=X". That bugs me because "url" is ambiguous, and is in fact the topic URL which was named "hub.topic" in other parts of the spec. I would like to change it to "hub.topic" for the publish request
#aaronpksince that is not part of PuSH 0.4 I can do that on my hub and still have it be 0.4-compliant. Assuming making any recommendation of payload won't be accepted into the spec by julien, is my next best option to file a feature request on superfeedr to accept that parameter name?
#voxpelliprobably, but is it really worth breaking compatibility with the 0.3 spec (which still is the most established PuSH-spec) just for a little more consistent naming?
#Vendanis it that hard to accept both, at least during transition?
#tantekaaronpk: can you specify handling of "hub.url" in some way that makes sense and is consistent with past treatment?
#tantekshould be doable without breaking backcompat
#voxpellitrying to skim the old docs to see why it was named like it was
#shifletttantek: What phrases does the bot search for? (I want to pay attention to the same stuff.)
snarfed and julien51 joined the channel
#julien51aaronpk: the reason why the spec uses hub.url vs. hub.topic is legacy… but not only
#julien51the distinction is that the publisher tells the hub that a resource (hub.url) was changed… when people could have susbcribed to a different one
#julien51the most common use case would be: the publisher pings with a blog url when the subscriber has subscribed to a feed url related to that blog url
#julien51one of the key design principle of PuSH was to push complexity down to the hub, and the hub alone should ‘link’ hub.topic(s) with various hub.url(s)
#Vendanissue there is that the hub would then need to know how that one update relates to all the places that it may update
#julien51I think hub.url vs. hub.topic was a neat idea, but in practice it probably createsmore problems than it solves
#julien51one of the most “frequent” ones was that feedburner URLs are case sensitives… and it pinged the hub with different url (just a case difference) than the ones used by susbcribers… and The Google hub was not great at dealing with this
#kylewmaaronpk: why did you say earlier that Vendan's push implemetnation is not PuSH 0.4 compatible?
#aaronpkkylewm: i said he doesn't have a push 0.4 hub
#julien51… also, one of the reason why specifying the relationship between the hub and publisher is out of scope I believe.
#aaronpkbut from a subscriber perspective his implementation is 0.4 compliant
#kylewmand since the spec only specifies the subscriber perspective, it's a complete implementation of the spec
#aaronpk!tell cweiske would you consider updating phubb to accept "hub.topic" for the publish request? if so, then all 3 public hubs support it and I'll change the how-to-push page to recommend it
#cweiske!tell aaronpk: phubb supports hub.topic now
#Loqicweiske: aaronpk left you a message 1 hour ago: would you consider updating phubb to accept "hub.topic" for the publish request? if so, then all 3 public hubs support it and I'll change the how-to-push page to recommend it http://indiewebcamp.com/irc/2015-05-19/line/1432056978895
#KevinMarks_Medium had an initial goal of timeless articles
#gRegorLoveSocial share icons, followed by tags beneath that, followed by "Follow channel/author", then finally the published date in small light text.
#tantekseriously, namespacing and XML have cluttered up so many specs/protocols/formats
#tantekunnecessary garbage in 99% of the use-cases
#tantekand in the 1% experimentation/extension use-case, the CSS approach to vendor prefixing has worked ok in practice
#aaronpkokay make an argument for hub.url in the publish again. my main concern is that it is inconsistent with itself because subscribers use hub.topic
#Vendannamespacing is when you separate groups of items based on common usage or creation, commonly used in programming and protocols, but can be over applied
#gRegorLoveI'll let others update /namespace accordingly. The <dfn> will need to be "namespace" to work with Loqi prompts
#aaronpkVendan: did you also change the parameters your hub sends when verifying subscriptions?
#Vendanhrm, that's an interesting one, I have not yet
#Vendanit should match the params the subs sent, right?
#aaronpkpresumably if you get a request that sends "mode=subscribe" instead of "hub.mode=subscribe" you should not use the namespaced params when verifying
#kylewmarbitrary combinations of alphanumeric characters
#Vendanit's a named set of "things". in PuSH, we're refering to a set of parameters, in programming languages, it's a named set of classes, in xml, it's a named set of tags
KevinMarks__, glennjones, snarfed, tantek and KevinMarks_ joined the channel
#kylewmtantek: my understanding of the timeline is like 1) they chose "topic" to be generic, imply that you could subscribe to anything, 2) much later superfeedr added support for search queries and chose to implement them URLs
#tantekURL is not universal enough, let's invent a new term "topic" to be *more* universal. Sigh. some combination of xkcd 927 and /architecture-astronomy
#kylewmso amusing semi-related aside: java.util.File has as method toURL() which is deprecated in favor of toURI(). so now you always have to call file.toURI().toURL()
#kylewm(because every actual consumer takes a URL and not a URI)