#social 2015-07-09

2015-07-09 UTC
the_frey, jasnell and nicolagreco joined the channel
#
melvster oh i just realized i've already implemented direct messaging
#
melvster and we also demoed it at the cambridge F2F \o/
jasnell, tilgovi, bblfish and the_frey joined the channel
#
slvrbckt melvster: indexeddb is accessible via web workers now, so you can get even better performance that way.
#
melvster slvrbckt: oh! that's great to know!
#
melvster slvrbckt: im amazed how lightning fast my explorer is ... its the first time ive seen "semantic web" and "fast" in the same sentence
#
slvrbckt yeah, it's such a huge improvement
#
melvster slvrbckt: im going to have to do that ... i was previously using a function daemon() in JS on a set interval ... can I just swap that out to a WW?
nicolagreco joined the channel
#
slvrbckt you'll need to pass the data to the WW, but luckily you can do that without any overhead as it'll just pass the reference
#
melvster sweet
#
melvster slvrbckt: what about same origin prisons (i mean policies) :)
#
slvrbckt you'll need to make sure it's an ArrayBuffer, then you can just postMessage it
#
slvrbckt not sure I understand, you want to fetch the data from within the web worker itself?
#
melvster slvrbckt: well my indexeddb could potentially be used by a number of apps (consuming social data) ... would I be forced to make them all on the same origin, or is there a way web workers can talk to multiple things?
#
slvrbckt aha
#
slvrbckt no i think you still have the same restrictions as you would using indexeddb in the primary thread, so you'd need to figure something additional out
#
slvrbckt but the webworker+indexeddb improves performance when reading/writing data by a huge amount
#
slvrbckt huge amount being a scientific measurement
#
slvrbckt :)
nicolagreco joined the channel
#
melvster slvrbckt: i just run them all as unhosted apps on github for now so they share an origin ... but thanks yeah maybe we can figure something out, thanks for the tip!
nicolagreco joined the channel
#
slvrbckt np
#
melvster slvrbckt: sent you a PM, in case you want to do lunch ...
the_frey and jaywink joined the channel
#
melvster rhiaro: I tried to write up my findings on user story implementations : https://github.com/w3c-social/social-ucr/issues/7#issuecomment-119661234
nicolagreco and the_frey joined the channel
#
melvster 'Reading a user's recent posts'
#
melvster what is 'recent'?
#
ben_thatmustbeme implementation detail
#
melvster ben_thatmustbeme: what do you use?
#
ben_thatmustbeme i use a reader that just orders all entries from those i follow by time code. so it starts at most recent for anyone i follow and moves backward in time
#
ben_thatmustbeme can continue back as far back as they share
#
ben_thatmustbeme this is the same as twitter, G+, FB (though FB curates it for "best")
elf-pavlik joined the channel
#
ben_thatmustbeme generically for a single user its just "view the most recent and N previous posts"
#
melvster ben_thatmustbeme: how many posts does your reader display?
the_frey joined the channel
#
ben_thatmustbeme infinite scroll
#
ben_thatmustbeme If I want to view an individual person I'll look at their site. Which is definitely implementation detail
#
melvster ben_thatmustbeme: what about users that want to go from a certain date?
#
melvster im not sure if #posts or #date or #postid is the most useful here
#
melvster "last 10 posts" seems quite limited from a client point of view
#
melvster for example for financial stuff im doing, you need monthly activity
#
melvster and for my personal diary ive needed daily activity
#
ben_thatmustbeme melvster: exactly why I say it's an implementation detail. There are places where you may have to just fetch a bunch of data and then limit to a set of them (specific date range)
#
ben_thatmustbeme Last 10 posts but way to get more like rel=next
#
melvster ben_thatmustbeme: yes I agree it's a detail, but it's going to be a VERY important detail for the API, dont you think?
#
ben_thatmustbeme No. So long as you can fetch all the data. Or not. If some system doesn't want to allow finding posts older than X. It hurts readability of their user posts but that's their choice
#
ben_thatmustbeme I don't know of any programming language that can't inspect a list to see how many elements in it
#
melvster ben_thatmustbeme: can agree to disagree on that one, depends on what functionality is wanted by people from the API time vs postid vs numposts
#
melvster thanks for the insight on your reader tho
#
melvster it would be helpful to know *exactly* how many posts your implementation gives, if you have that information
#
ben_thatmustbeme Huh? Didn't understand the second half of that
#
melvster ben_thatmustbeme: np, isnt too important, i was trying to ask about how many posts you get in your reader, in your implementation
#
ben_thatmustbeme My site itself returns 25 right now. I think. I changed it at one point
#
melvster thanks
#
melvster 25 seems like a good number
#
ben_thatmustbeme I don't have rel=next links right now (I have them working just never linked them. I really should)
#
ben_thatmustbeme I go back by skip=, count=
#
ben_thatmustbeme I know aaronpk does previous posts by post id offset (reddit style)
#
melvster ben_thatmustbeme: for my wallet ive been doing last 100 tx -- maybe that's a bit much, for my chat/blogging system it's the previous 2 days worth of messages e.g. today and yesterday ... or could be today and one month ago ... for my diary it's per day
#
ben_thatmustbeme Yeah it really depends on the data you are sharing
#
melvster for a commerce system we're working on (enterprise) we are thinking 90 days of memory, but that's a much bigger scale project
#
ben_thatmustbeme Shorter messages can have a lot more. Longer articles should have less
#
melvster exactly
#
melvster im just coding the inbox for the wallet now
#
melvster so happy ive added caching
#
melvster my app is so fast i could literally almost cry :)
#
melvster but it's important to know what to cache, how much, when to go back etc.
#
melvster ben_thatmustbeme: in case you were wondering the wallet is going to be used to implement the anti spam 402 code
#
ben_thatmustbeme Lost signal there. On the train to work
#
ben_thatmustbeme Antispam 402 code?
#
melvster http 402
#
melvster here's a writeup of attempts to implement SWAT0 http://ods.openlinksw.com/dataspace/doc/dav/wiki/ODS/ODSSWATOTutorial
#
melvster this is using SoLiD technology when it was called data spaces
#
melvster SoLiD is, in some sense, the evolution of that work
#
rhiaro Wasn't the idea with SWAT0 that there were supposed to be three totally different implementations of the spec interoperating, not three versions of the same piece of software?
#
ben_thatmustbeme oh they payment required stuff
#
melvster ben_thatmustbeme: yup for reducing spam in the inbox user story
#
melvster but i like payments in general, so it's nice to find a concrete use
#
ben_thatmustbeme rhiaro: yes, but it is a first step
#
ben_thatmustbeme as to it being implemented by others is another hurdle
#
melvster id be happier if swat0 was less focused on, and the pre requisites were completed first
#
ben_thatmustbeme as said before, SWAT0 was an ACID test, they are supposed to be difficult to complete
#
ben_thatmustbeme i don't see a problem focusing on it
#
aaronpk err_connection_refused with that ods.openlinksw.com link
#
aaronpk yes, the whole point of SWAT0 is that there are three implementations, not just one piece of software that can fulfill it
#
aaronpk and yes it's supposed to be hard
#
aaronpk and it's fine that it has dependencies, that's what makes it hard
ShaneHudson and tilgovi joined the channel
#
melvster aaronpk: if it's supposed to be hard, it should be in v1, imho ... it's really putting off implementors
#
aaronpk who is it putting off? a bunch of us in #indiewebcamp are happily making progress towards it!
#
melvster progress != implemented
#
aaronpk implementation requires progress
#
melvster afaict indieweb has only implemented 1 user story, the 'post a note' to date ... looking forward to more progress tho
#
aaronpk go search the user story page for "implemented" and you'll find a few more ;)
#
melvster i did
#
aaronpk RSVPs are also implemented
#
aaronpk as is posting a photo
#
aaronpk as is comments
#
melvster aaronpk: feel free to document your implementations at : https://github.com/w3c-social/social-ucr
#
melvster im sure you have made lots of progress
#
aaronpk but whatever, my time is better spent working on swat0 than arguing with you
#
melvster good luck!
#
melvster please do share your results if you get anywhere
almereyda joined the channel
#
melvster aaronpk: or if you dont have time to fill in the page with links, perhaps you can clarify on this thread: https://github.com/w3c-social/social-ucr/issues/7#issuecomment-119899208
#
ben_thatmustbeme huh youtube uses http 402 for captca needed
elf-pavlik joined the channel
#
melvster yeah
bblfish, jasnell, tilgovi, Arnaud, KevinMarks, elf-pavlik and nicolagreco joined the channel