#dev 2018-11-22

2018-11-22 UTC
[eddie] joined the channel
#
GWG
What's going on tonight?
#
jacky
going to write a post about what I got going so far in koype and some next steps
#
GWG
Great
#
jacky
then to make and set aside all of the prep work for food tomorrow
#
GWG
I'm not sure if I'll get to Indieweb tonight
#
GWG
I'm on furniture moving
#
jacky
we got a vHWC tonight?
#
GWG
jacky: Not that I'm aware of. I think next week. But I'll be out regretfully
#
[tantek]
jacky, any thoughts for a /2019-01-01-commitment?
#
[tantek]
fixes singular to redirect to plural
benwerd, mblaney, the_hoom_, dansup, the_hoomanist, dansup_, jjuran and eli_oat joined the channel
#
jacky
I'm working on a few at the moment
#
[tantek]
I mean listing them explicitly, or at least an easy one, here: https://indieweb.org/2019-01-01-commitments#Commitments
benwerd, tantek, snarfed, [tantek], ben_thatmustbeme, the_hoomanist, mblaney, cweiske, eduardm and swentel joined the channel
#
Zegnat
sknebel: I saw you added a TTL question to autoauth. I am wondering if this is up to autoauth to solve, or if we should look into introducing this as a separate spec? I feel like different people have at different times asked about expiring tokens and the possibilities for refresh tokens
#
Zegnat
So far we’ve always just said that you, as endpoint owner, can issue auto-expiring tokens. But we have never specified how clients may realise this is happening
petermolnar and leg joined the channel
#
sknebel
Zegnat: "may realize" - it appears to be not mentioned in the indieauth spec (I thought it was), but in OAuth there's a recommended property of "expires_in" in the token response, whereas my question is about *requesting* a certain TTL
#
Zegnat
What is the reason to request a specific TTL?
#
Loqi
It looks like we don't have a page for "reason to request a specific TTL" yet. Would you like to create it? (Or just say "reason to request a specific TTL is ____", a sentence describing the term)
#
sknebel
making sure the client doesn't get a token that's valid for longer than their token with the auth endpoint?
#
sknebel
maybe that's not so important/a mechanism to re-validate that might be more important
barpthewire joined the channel
#
sknebel
I'm not really liking the revocation story yet. client token gets revoked at users token endpoint. users token endpoint tells auth endpoint that's happened, auth endpoint goes and posts individual revocation requests to all the sites? relatively short-lived tokens then make a lot of sense, should the auth endpoint be able to ask for those?
[kevinmarks] and jeremych_ joined the channel
#
Zegnat
Ah, right, yeah the revocation story doesn’t flow well.
#
Zegnat
It is easy enough for the client to go and revoke, but requesting revokation through the main token is a hassle
[grantcodes] joined the channel
#
[grantcodes]
After building this directory, I'm wondering if there has been any thought into how to determine a preferred name for a person? Specifically in the context of mentioning them on your site?
#
jgmac1106
grant search the chat for what [eddie] did with his nickname cache
#
[grantcodes]
Ok will do
[jgmac1106] joined the channel
#
sknebel
[grantcodes]: as in, how they'd like you to link them?
#
jgmac1106
people are working you drop in their url and if you syndicate to X and Y the correct nickname will go to each
#
sknebel
that's not a "preferred" name, is it?
#
[grantcodes]
sknebel not the link but more if I should display their name, nickname or url
#
[grantcodes]
In my mind it would go nickname > name > url
#
sknebel
hm, that's tricky. and might also depend on the context
#
swentel
hmm, I'm constantly getting parameters don't match if I try to login, weird
#
jgmac1106
i think it would be name > url> nickname on syndication or just <a h-card u-url href=“$>Name</a> ty
#
sknebel
i.e. on my blog I'd typically mention people here by nick or first name, if I were to write something to share for a wider audience I might go for full names
#
sknebel
displaying URLs is probably only last resort when talking about a person
#
Zegnat
[grantcodes], for the directory I would not prefer nickname from the h-card to actual name. The nickname property may specifically point at an alias or IM handle, and I do not think those are always going to be the most obvious (or even reflective of what people are called)
#
Zegnat
But for your own personal site, I would call people in writing the same as I would call them when talking about them or to them offline. Which will mostly be first names (or shortened first names in some cases)
#
[grantcodes]
swentel: Must be something to do with your auth endpoint returning a me parameter. To be honest I'm not sure I even need to make that check
#
[grantcodes]
Zegnat agreed for the directory, but I was thinking more in the case of writing a blog post and nickname
#
[grantcodes]
In some advanced future I am hoping to have a micropub client autofill mentions. I suppose it would might be possible to autofill the full name and the user can change it to whatever they want as long as the html with the link is preserved.
#
[jgmac1106]
@zegnat spending the morning going back to our citation hell...trying to find source code of this: https://library.unc.edu/citationbuilder/ to then add on what you and I already did
#
Zegnat
He, have fun [jgmac1106] ;)
#
jgmac1106
If I they let me play with their tool be relatively easy
#
[grantcodes]
Also oops yes Zegnat the directory doesn't resolve urls, it uses exactly what you type in as your url as the key, which maybe isn't perfect. But you should be able to log in without your slash to delete that extra one again 😅
#
Zegnat
Aah, gotcha, will do!
#
Zegnat
petermolnar: just had another go with Safari on your theme switcher, this works really nice and basically exactly as expected
#
Zegnat
[grantcodes]: on second login I am getting State does not match. I guess this will have to wait until after work.
#
[grantcodes]
Ha oops 😄 mvp
[Vincent] and [schmarty] joined the channel
#
petermolnar
Zegnat: thank you!
#
Loqi
petermolnar: [grantcodes] left you a message 1 day, 16 hours ago: - re react website it's basically as Eddie described, the server does the initial render based on the initial state of the components, so basically just generates html + css for a page, then if a user has js enabled everything is loaded client side where you can do the real time interaction stuff
#
petermolnar
I'm glad it does.
tantek, [kevinmarks], iasai and [eddie] joined the channel
#
[eddie]
grantcodes: right now I use custom nicknames but I should probably move that to something that isn’t already used so I could store someone’s nickname AND what I want to call them
#
[eddie]
grantcodes: autofilling full name and allowing editing is similar to Facebook. It auto fills fullname and allows the removal of the last name through backspace.
#
[eddie]
So that’s good via prior art. But it would probably be helpful, to be able to modify any of the text as long as the html link stays intact like you said
#
[eddie]
You could also just provide options where the person could select the person’s first name, full name or nickname when they are autofillinf
[grantcodes] joined the channel
#
Zegnat
What is petname?
#
Loqi
petname is a local, typically private, nickname for a user, similar to how the entries in a phone's telephone book have user-defined names, e.g https://indieweb.org/petname
#
Zegnat
“What [you] want to call them” ^^^
#
Zegnat
I feel like that page may need some citations for the usage of the word
#
Loqi
[Christopher Lemmer Webber] Petnames for Self Sovereign and Human Readable Identifiers: https://github.com/cwebber/rebooting-the-web-of-trust-spring2018/blob/petnames/topics-and-advance-readings/petnames.mdAdvanced reading / topic paper I'm submitting for Rebooting Web of Trust...
#
Zegnat
petname << [http://www.skyhunter.com/marcs/petnames/IntroPetNames.html An Introduction to Petname Systems] by [http://www.skyhunter.com/marc.html Marc Stiegler] describes the envisioned different between petnames and nicknames, and some of its use-cases
#
Loqi
ok, I added "[http://www.skyhunter.com/marcs/petnames/IntroPetNames.html An Introduction to Petname Systems] by [http://www.skyhunter.com/marc.html Marc Stiegler] describes the envisioned different between petnames and nicknames, and some of its use-cases" to the "See Also" section of /petname https://indieweb.org/wiki/index.php?diff=54186&oldid=54185
#
Zegnat
I thought I read something about it published W3 also, but can only find the petname tool that allowed you to assign petnames to websites: https://www.w3.org/2005/Security/usability-ws/papers/02-hp-petname/
[jgmac1106] joined the channel
#
[jgmac1106]
@zegnat DID stuff usually comes out of the verified claims community group...or a lot of the people and vocab overlap
snarfed joined the channel
snarfed joined the channel
#
Zegnat
Sure, but I was looking at petnames stuff. Which can be mapped on DID but they themselves definitely are no such thing (not globally unique)
iasai joined the channel
#
[jgmac1106]
ohh I see, thx
#
aaronpk
[grantcodes] I get "missing auth parameter" when I try to log in
#
petermolnar
https://github.com/yaronn/blessed-contrib -> "Map", for whoever wants a truly retro website, but with checkins
#
Loqi
[yaronn] blessed-contrib: Build terminal dashboards using ascii/ansi art and javascript
#
petermolnar
although I may have misunderstood it, it's not a web toolkit, it's a terminal toolkit
#
sknebel
need to add a JS terminal emulator, done :D
#
petermolnar
aaargh :D
#
sknebel
hm, I occasionally toyed with the idea of doing a site in the style of the old Turbo Pascal (and other stuff) TUIs
#
petermolnar
for design + fonts
#
sknebel
petermolnar: yeah
#
petermolnar
the disturbing part of that site (the int10h) is that for me, it's really easy to navigate and read it, way easier, than most of the modern/material/flat/etc designs
#
petermolnar
the bad part of those fonts is the obvious lack of utf-8
#
petermolnar
which, in my case (thanks, ő and ű), is quite important
[grantcodes] joined the channel
#
[grantcodes]
aaronpk probably because the auth endpoint isn't required to return the me parameter any more is it? I think I was reading outdated info somewhere
#
aaronpk
correct
#
[grantcodes]
Cool, probably works for you now
#
sknebel
curious to see how well the microformats parser holds up
#
aaronpk
woo works
#
[grantcodes]
The node / js one is pretty out of date
leg joined the channel
#
sknebel
aaronpk Zegnat do you remember why in the autoauth proposal there's a callback to deliver the token to the authorization endpoint, when in normal indieauth roughly the same steps happen during a request?
#
aaronpk
not totally sure but maybe it had to do with not wanting any party to be waiting for a second request to happen within another request?
#
sknebel
might be.
#
sknebel
guess that'S not the case in the normal flow
swentel and DenSchub joined the channel
#
sknebel
should that have it's own "state" parameter or do we reuse the authorization_code?
barpthewire, iasai, the_hoomanist, tantek, jjuran, [mrkrndvs], [jgmac1106] and [kevinmarks] joined the channel
#
[kevinmarks]
Pet names are fine. Publishing them is weird.
[tantek] joined the channel
#
[tantek]
also, mostly handwaving is not helpful. if you've got screenshots to UIs that are using them then great, otherwise not sure how useful that page is TBH
cjwillcock joined the channel