dev-diary/mix

2018-10-xx

prepped and did some in person pairing sessions with @dan hassan, resulting in this nice tutorial module : https://github.com/mixmix/ssb-client-basic

2018-11-02

recorded a tour of scuttle-dark-crystal with dan, posted about it over here : %srfbULm...

2018-11-05 > 2018-11-10

Solving patchbay problems, working on standalone app pattern, and trying to integrate scuttle-shell.
See here in the grant dev diary for detail

2018-11-12 > 2018-11-13

Making installers for infinite-game
Reviewing a PR by Kieran

2018-11-14

Reviewed a PR by Peg,
Proposed some changes to Kieran's PR
Started reading through all the recent thinking on the coconut death roleplays, crypto inheritence best practice, .. didn't get to looking at others designs yet, want to think for myself and not be primed by others.

2018-11-15

Had a call with @dan hassan, discussing what we should focus on.
Wrote up a proposal that we should focus on the theft case (not ssb identity recovery)

here : %0S/DsQN...

2018-11-16

Made this https://github.com/ssbc/modules.scuttlebutt.nz (not strictly dark-crystal, but an ecosystem contribution which will help improve documentation which will help us in the long run)

Did some thinking about what the flow could / should feel like.

This is my home page. It's all about seeing the status of my secrets and creating new ones.

• looking at a secret you can see if shards have been returned
• you can click on a secret to dive in
• there's a section for "spent" secrets because a secret that's been returned is dangerous because it's recorded in it's entirety in a single place now.

There's another page about "others shards". What matters here is the friendships and trust, so I show peoples faces large and prominent, with shards I've got below (these should look quite different to the complete crystal form on the my secrets page). -

• for Dan one of those shards is greyed out because it's been spent / sent.
• for Cat, there's a shard that is summoning my attention by moving or pulsing. Clicking on it will give me actions
• for Dan old, I've either seen a message that suggests this account is dead, or I've marked it dead. I'm at a juncture where I can see there are live shards associated with that account, and I can decide what to do.

This is what I got stepping way back and it's not complete. I think it could be good to make a cheklist of things a person should be able to do. These designs would fail "I know easily if I've been summoned to perform some action"... maybe that would be in a notificatins tab... maybe that would be on my home page there's a Bat-signal or a glowing shard that takes me to people that need actions.

One of the things I noticed with our v1 is it's easy to get swamped in shards and not see that some have been requested. etc.

I also had a daydream about when you're on a screen and there are actions that when you hove over an action it gives you a preview animation. This is an idea straight up stolen from Starcraft 2...

e.g. imagine these as being a moving image

me centric

shard a secret

recover a secret

friend centric

return a shard OR fwd a shard

The idea of movement showing where things are going excites me

Love this.

2018-11-19

I've realised that this week the design work is less about what the views look like and more about what paths we're building. Scamps (rough paper sketches) are really just to help us reveals what matters on those journeys.

Ok wow here's some paths : %4TpLHYN...

Organised weekly gathering and invoked a design meeting (to replace the standard retro and planning)

Tried to draw up visual calendar of when everyone is away but blocked because don't want to install keybase D:

Started a thread about pay tracking

2018-11-20

pulled design thinking together, prepped for weekly gathering, coordinated pulling together all the loose ends with Dan

did 2-hour design call with (most of) team and got a minimal set features agreed for the next release of DC, which is super exciting!

2018-11-21

researched varios possible changes to sharding, wrote demo code to learn more about cryptography
threads : %m6A6R6K... %2qEtYbJ... %2MWm+AM...

2018-11-22

met with @peg and hammered out details from sharding research, deciding what would be "in" this release

pulled the tuesday meeting decisions out into really nicely spec'd trello card. This made me surprisingly happy and settled feeling <3

god damn it @mikey just told me about sodium-native and sodium-friend who have tutorials in general ... cc @peg @kieran @dan hassan

AND I ordered dan thai food, which felt super great. I hope it arrives at the right time - ordering food across timezones is a trust exercise!

2018-11-23

I was overcome by a rebelious urge to build features for threads... this is what came out. I didn't do anything substantive directly on DarkCrystal.
Call it a wellbeing day perhaps?

2018-11-24

reviewed the development branch of scuttle-dark-crystal and started in on some changes:

• changed out chloride for sodium-native
• refacored the the secrets-wrapper
  • seperated our versioned parts
  • changed the MAC to be at the front of the shard (as dominic recommended)
• factored out shard/async/build.js
  • to isolate some logic, make testing cleaner ...
  • TODO : add folder structure to make v1 and v2 tests clear?
• remove SCHEMA_VERSION as a floating variable
• check a minor possible error

2018-11-27

• finsihed reviewing development branch
• pushed refactors, and wrote up work still needed to make shard forwarding stable

2018-11-28

all of these are beta testing shards, don't get excited about me revealing holding actual secrets

• reviewed some of peg's work
• started a re-write of the "Others Shards" page (pictured above)
  • see work so far on others_shards_v2 branch
  • still needs a way to return shards
  • need to replace font-awesome diamond with shards

Any input on what the design for returning shards in this context could look like would be appreciated. Some options :

• when you hover over a shard you see more info, and can click actions (hard)
• you click on a shard and get a modal popping up with all the detail about it

p.s. @dangerousbeans you are the bain of my life when it comes to fitting your name into designs. Thank you (it's annoying but also very good for me)

2018-11-29

built a minimum "show shard" view, and way to scroll through them

2018-11-30

decided in the shower that that design is clunky, and am implemented something different

really want some mini crystal shard icons ... could be little watercolors...

2018-12-03

had a really great call with @kieran this morning. couple of takeaways were:

• I'm going to make sure I (or peg) spend an hour each week lining up clear work for Kieran (because spending 1 hour to unlock a smooth 2 days of contribution is just working smarter. without this kieran spends 1.5 days getting up to speed).
• we're going to experiment with noticing and voicing our emotions earlier. e.g. "hey kieran I notice I'm feeling really annoyed, and I think it's about how the meeting went. I don't know what it's about, but lets figure it out"

I also merged https://github.com/blockades/patchbay-dark-crystal/pull/18

MOST EXCITING - I discovered a bug in the code I'd written in v1 which was the cause of live-updating not happening on new messages. It's fixed by a version of scuttle-dark-crystal I published.

I just did a test across computers and now as soon as a request comes in shards start blinking !!!!!! (and new shards from friends immediately appear). This is gonna be awesome for workshops.

TODO : need to get better shard logos than that font-awesome diamond ...

2018-12-04

had a bad time getting patchbay installers working. was meant to be a quick build test of scuttle-shell, but didn't get anywhere near. needed work for being able to build isntallers for new version of dark-crystal, but not fun.

2018-12-05

started into a refactor which will hopefully allow for more comprehensive testing of version support for all the shard and replyies etc.

2018-12-06

swapped in fetch+mend for recombine and watched it work. fixed some minor bugs that it hadn't caught but all tests passing. feeling a little insecure about how much work it's taken and waiting to see if it's a net win payoff. I'm hoping some of the investment will come into its own with things like ephemeral keys.

it's certainly helped clarify requirements for me. e.g. we need versions on reply messages like :

• version - the layout of the message (as dictated by invite-reply)
• shardVersion - what is the content I'm receiving
• replyVersion - what sort of reply is it. Is that that shard with shardVersion X just unboxed and bare (like our v1 replies) or is it an ephemeral return I should by trying to unbox with some ephemeral key.

I think @peg was gonna have a look into some of that? (or at least we talked about these ideas). I'm not sure what the ideal solution is or if there's a clearer place all of this info should exist. Maybe it should be referrer to in the request (invite) message ... but it would be nice for there to be a symmetry between forwarded messages and reply messages (IMO)

2018-12-07

didn't get any code done today so far.

had broken sleep, swim class with ziva, and running chicken soup to sick friends. I did manage to attend a call with Rich about crews + congregations, which I think will lead to stronger garden in th future. am going to put today down as wellbeing and hollistic and make up some hours where I can later

2018-12-10 mon

more versioning

2018-12-11 tues

finally merged the versioning work. kinda excited about how I could combine recombine and recombineForward. I think our testing is pretty comprehensive now, but it's kinda shocking out much it took.

it also seems like versioning of different message types should not all move in lockstep. but this will require more time.

2018-12-12 weds

started in to reviewing PRs

2018-12-13 thur

caught up with @peg about what our next priorities are. We agreed that I would spend a time-boxed amount of time on seeing about breaking the version lockstep. Peg is gonna have a go at building some front end view as a bit of a stretch + learning exercise

2018-12-14

connected @kieran's work in, and pushed the forward-shard button for the first time, which was exciting!

hmmm, there's some missing work days here. Rough summary:

• furiously reviewing and merging work
• documenting what needed to be published in order to hit milestone
• reworked how schemas work
  • refactored sscb-dark-crystal-schema to decouple versions - means we can release new versions of e.g. shard format without having to release new versions of all other messages redundently
  • re-worked ssb-schema-validation to help fit this more
• added UI ability to label secrets
• completes UI work to enable fwding shards
• tried to organise a weekly meeting
  • we didn't plan ahead, ended up trying to do it on new years day
  • decided for group we were meeting same time but a week on ):
• read all of the bits of code Kieran and Peg worked on in last couple of days before xmas
  • was surprised there was only 1 PR
  • found a bug and documented it on trello
  • added an action to revisit how we're merging + communicating

2018-01-07

STANDALONE DARK CRYSTAL

INSTALLER: github.com/blockades/dark-crystal-standalone/releases

I've only made a Linux build so far.

forked infinite-game work to build this alpha prototype. Really happy... it only took me 58 minutes to make this - earlier work has totally paid off.

It will run at the same time as the current Patchbay. No gaurentees about what happens when you try this at same time as current Patchwork - my best guess is if you're running from the master branch you might be ok... but if you're running the installer this is gonna be slow because you have to move between different versions of the database setup. (this is outside my control <3)

descended into the hole which is trying to build scuttle-shell based patchbay installers (so that we can get an app ecosystem going more). Immediately hit problems with being unable to run Patchbay, so rolled back and just tried to update Patchbay to use the same version of ssb-server as scuttle-shell. No fun there either. Started peeling the onion and had a cry. begged for help + support

2019-01-08

phew, a big day

• eth foundation call
  • tried to make a video, but linux was being a dog, so I got this demo video of me talking like a chip-munk : https://www.youtube.com/watch?v=kjhzHGqDMIk
  • had call, went really well
• got back into scuttle-shell work... the rabbit hole went deeper:
  • working on upgrading to ssb-server@13.5.2
    • https://github.com/ssbc/patchbay/pull/295
    • https://github.com/ssbc/patchbay/pull/299
  • improving docs, testing...
    • https://github.com/ssbc/ssb-client/pull/35
    • https://github.com/ssbc/ssb-config/pull/28
    • https://github.com/ssbc/ssb-server/pull/612

2019-01-08

• got a scuttle-shell based patchbay installer working (for linux)
• progressed a bunch of those PRs

2019-01-10

Design planning

We've recently added new shard-forwarding functionality to dark-crystal, but that has only been partially designed. We're at a point where we've agree it's important to stop and do some design work (and also clarify how we do design work!)

Past (where we are, and our journey so far):

• large parts of our current setup are just cobbled together (visually and UX wise)
• we had tension in this recent round because:
  • different people did different design, but we didn't have or make time to decide which path we were taking. In the end we jsut built something quite simple.

Future (what I think we should do):

• take what we've built and do (more) in person user-tests (and record them)
  • give people the tool with no context, see what they think
  • see how far they can get with a little context
  • keep your mouth shut as much as possible !
• pause and reflect what was most important / confusing
• make scamps (paper drawings)
  • go and user test these scamps with humans
  • rework and repeat.
• write the code after we've tested designs

The advantage of this approach is that it's faster and less wasteful, and is more empirical than basing our direction on our individual preferences. I've seem this work really well before

cc @kieran @peg @dan hassan

2019-01-13

Added local peers visibility to make user testing easier for future iterations (test account on a local wifi style !)

Progressed a bunch of ssb-server config, and got dominic to help debug a mystery error. Many ongoing PRs

2019-01-15

User interview >> See here

More work on back end config PRs

2019-02-16

worked on an ssb-config PR which adds more stability and clarity to how connections are being set up. This is a piece of maintenance work which is gonna contribute in a more long term way to being able to build installers as versions of ssb-server are released.

Got a strategy meeting tonight

wow, my ability to accurately increment dates is suck! ... might be cool to build a tool for dev diaries huh!

2019-01-17

talked to @matt and started the process of extracting his suggest mentions work from patchwork - it's on the server and very fast + efficient

got a minimal setup kinda working, but was stumbling on a depject bug for a while. sigh

2019-01-18

A stable save-point! :

• https://github.com/ssbc/ssb-suggest
• https://github.com/ssbc/patchbay/pull/307

Really happy with how this is going. I've taken the best of the Patchwork work, but extended it to include mention suggestions for people out to 2 hops. There's been a regression in that this suggestion pattern does not support alternative names for the same person that have occured through history... this is a must have for me (e.g. mikey is the same as ahfinosaur) because scuttlebutt is all about plurality!

So still to do:

• add alternative names
• add mapping for Suggestions which enables rendering of primary name + secondary names
• modify sorting to handle all this new data
• coax Matt into using ssb-suggest

Really massive thanks (once again) to the awesome work @matt has gifted. I learn a lot from you guy, and you enable me to build so much more so quickly <3

2019-01-21

Low work day (was a regional holiday), but got the guts of mentions work in place

2019-01-22

Published new mentions work to patchbay master branch : see release notes here %QS86lcE...

Really happy with how this came out. Sure there'll be some tuning to do but seems to be pretty solid

2019-01-24

• tried upgrading patchbay to ssb-server@14 but was blocked
  • missing deps from the refactor, and inability to make changes because dominic holds ownership of many of the new modules. Have messaged him
  • work so far : https://github.com/ssbc/patchbay/pull/309
• fixed patchbay on windows
  • should enableto build new installers finally

2019-01-25

Started a scry for our second weekly meeting slot. (rest of the day was a sick day)

2019-01-27 (sunday)

Pinned the second regular meeting time for the week, started a gathering

2019-01-26

• got fixes in place for ssb-server@14
• released a new patchbay (based on new server)
  • installers release (linux + windows only so far ... Alanna is using her laptop!) : %FeXkqUq...
  • @kieran this could be relevant to user testing

that 26 is a 28...

2019-01-29 | tues

• built multi-channel support for patchbay (team tools?) >> see here
• prepped for and facilitated meeting

2019-01-30 | weds

• did some prototyping and comms about new on boarding ideas > read here

2019-01-31 | thurs

• Slot B call this morning, talked about grants
• workshopeed with Dan:
  • team application EHF
  • spreasheet forecasting runway with current funds
• Onboarding:
  • read peoples replies to my onboarding app proposal
  • started testing Dominic's new ssb-user-invites and left notes on some gotchas

... I've missed some entries, and I'm just gonna let them go because worrying about accounting them has been putting me off writing

....

put an edmund hillary fellowship application in with @DanHassan.
(ask if you'd like to see our 3 minute video)

...

did a bunch of financial admin / forecasting work

2019-02-13

Got a prototype onboarding thing built called ssb-ahoy : %KFnMyQ8...