Text extraction. See Typescript Archive
09:34:49
09:34:49 From Jeff Miller (I've promised to go grocery shopping this morning, but there are things that need to happen first)
09:36:38 From Jeff Miller small-community scale software seems neglected even within organizations; departmental-level "Story Craft"
09:38:10 From Jeff Miller pop wide?
09:40:19 From Jeff Miller "quantity of" ? / "quality of" ? (variables)
09:43:26
09:43:26 From Jeff Miller Nice examples on the Miro board, showing the patterns, including the quantities / qualities. fear of the blank page? prototype based practice
09:44:18 From Peter templates / examples always help
09:44:25 From Jeff Miller 100% I love having multiple models tha are all valid but slightly different examples.
09:45:17 From Peter Interesting that latest "AI" copilots really end up generating / dreaming very customized templates (that may or may not be correct) Rather than a good textbook / lookup system with vetted good patterns
09:45:43 From Jeff Miller right, what is essential and what is incidental are not distinguished "there's grass in the picture, there must be a dog in the yard"
09:47:35 From Jeff Miller I was just reading Lorin Hochstein's recent blog entries about focusing on the wrong things, which sounds like Eric's template question problem. "If the outage cost $$$, how much should we paying to make sure it never happens again?" <- too particular a solution.
09:48:31 From Marc Pierson http://marc.relocalizecreativity.net/view/when-do-you-need-the-numbers Eric take a look.
09:48:53 From Jeff Miller Ward points out that part of incident response might be the salience of how an incident was recognized as being in progress.
09:49:22 From Peter Relating to software design, the relative value of a "Detailed Design Document" vs. "Operating principles Document" . Too many details require too much work, seem unchallengable by people not involved in those details.
09:50:07 From Jeff Miller complex systems and overspecific corrections: https://surfingcomplexity.blog/2023/12/22/the-courage-to-imagine-other-failures/
09:50:21 From Brian The template thing reminds me of different audiences talking past each other. The "top" person cares about $$$ and the bottom person cares about tech aspects.
09:50:30 From Peter But the existance of such a document just gives other (largely non-technical) people warm fuzzies that you know what you are doing, helps with trust Although might just be a snow job
09:50:51 From Jeff Miller uh-huh: also productivity based management vs. capability based management.
09:50:52 From Peter Generated by AI nowadays
09:51:21 From Brian "remote" Anna
09:52:12 From Jeff Miller The Copper Corridor in middle Arizona
09:53:54 From Brian I like that there is a single shared state and there are two evaluation functions on that state.
09:54:35 From Jeff Miller https://www.discovercoppercorridor.org/explore (tourist site map; Oracle, Arizona through Winkelman, Hayden, Superior, and Globe. Also an interesting place for geologists and paleontologists.
09:54:48 From Ward Cunningham https://scalar.usc.edu/works/boundary-objects-guide/boundary-objects#:~:text=A%20boundary%20object%20is%20any,409).
09:55:08 From Brian In Marc's EIP, I think the center column is made of boundary objects.
09:55:23 From Jeff Miller "Separate Ways' pattern in Domain Driven Design, when you need to not align things closely. Boundary object, can be pointed to and named. But the workflow and significance can be very different across the boundary.
09:56:20 From Brian A divinding river
09:56:28 From Jeff Miller "a hosiptal visit" different to each participant
09:56:40 From Paul Rodwell Boundary Objects: A Field Guide - https://scalar.usc.edu/works/boundary-objects-guide/index
09:56:50 From Jeff Miller "a hospital visit" <- spellcheck (a boundary object)
09:57:47 From Paul Rodwell Replying to "Boundary Objects: A ..." Links to same doc as Ward’s link above
09:59:33 From Jeff Miller Marc's VSM/Sofi model (including the environment and relationships) appreciated by the VSM community; a possible general boundary object for examining relationships from the point of view of various participants. (?) :lightbulb:
10:01:32 From Jeff Miller Ward's heuristic example about helping groups at Tektronix research: "draw a picture of your system" -- "the problem is always at the center of the diagram" -- becuase the team is thinking of that first.
10:03:51 From Jeff Miller When the discussion was getting lost in details, the lead engineer could quickly draw a diagram of the system they were working with.
10:04:29 From Brian There is a similar thing that happens when you ask people what they value. Your working relationship can become signficantly more effective. For example, one person really likes to learn new things and another person wants to leave by 3:30 each afternoon. Easy to work with and plan, *if* you know those things.
10:04:43 From Jeff Miller Ward's example of a parser and a spec, tinkering with the spec until it could by parsed. Brian: asking my manager for his dashboard was very informative in the "what do you value?" sense.
10:05:24 From Brian It was the consultant, not the consulting...
10:05:28 From Jeff Miller (not necessarily helpful, but useful for understanding)
10:06:29 From Brian A solid team (and leader) will succeed despite the process...the process can help or not, but is the lessor factor...
10:06:41 From Jeff Miller "the customer demonstrates the workflow weekly to engineering" - Ward. "sometimes a dorky thing that the weakest programmer put in as a lark ends up being the most popular thing to the customer" -Ward
10:07:46 From Jeff Miller "you finally got around to the thing that sparks joy" (color customization, for example) "Oh, we could have done that anytime..." / "So why did you wait?"
10:08:36 From Brian Some of the most user friendly programs I've used were TUIs. :) UXR...
10:09:26 From Jeff Miller Marc as engaged customer using FedWiki, demonstrating workflow and usage for purposeful patterns.
10:09:57 From Peter Chat / LLM interfaces and AR/VR interfaces are now where we have lattitude for doing really new things in UX. WIMP interfaces have stratified since Nelson's Mother of All Demos fossilized?
10:11:06 From Jeff Miller Brian considers a warm hand-off for usability in that the dev team having to continue to use their software project alongside the customers; to see awkward usability issues by using it like the customers do. "I wouldn't have had that insight without working in the same way the customers did."
10:11:55 From Paul Rodwell Replying to "Boundary Objects: A ..." A single page from key concepts in Intercultural Dialogue https://centerforinterculturaldialogue.files.wordpress.com/2014/10/key-concept-boundary-objects.pdf
10:15:27
10:15:27 From Jeff Miller Brian, as an operator of the system alongside the team, built trust and rapport and gained insights; there's an "official script" and an "here's how we actually get the job done."
10:17:03 From Jeff Miller "a boundary person" , "an ambassador" , "a translator" (the glue work?) "I've had fun being in the middle." - Brian
10:19:06 From Jeff Miller "if we just had this ONE THING" - a way that a consulting developer can focus on the work most valuable to the customer. - Ward
10:21:22 From Jeff Miller Greg Wilson / Software Carpentry ?
10:22:34 From Jeff Miller "no computer scientist writes languages with units as a first-class thing!" - Eric
10:23:39 From Jeff Miller "and then you want to have quantities, which have magnitude, units, and uncertainty." - Brian
10:24:40 From Jeff Miller I love examples with particular named nodes and relationships.
10:24:57 From Paul Rodwell History of software carpentry - https://software-carpentry.org/scf/history/
10:25:56 From Jeff Miller (to Marc's demonstration of the EIP column diagrams, the hierarchy of ecosystems, the hierarchy of political governance structures, and the interest groups, "institutions" between) "where are we on this diagram" ? :)
10:26:25 From Paul Rodwell I’m reminded of Tim Ewald’s talk from 2013 - https://www.youtube.com/watch?v=ShEez0JkOFw
10:27:01 From Jeff Miller "an association of neighborhood centers" versus "a specific neighborhood center and a town council"
10:30:45
10:30:45 From Jeff Miller Eric describes interacting alongside other volunteers with school board members related to computers provided for the school. "If you bring in members of the larger community on your side, the board members can't argue that your priorities are narrow and cannot be advanced in favor of the broader community". ~ paraphrase
10:31:56 From Jeff Miller "cost of hardware" / "burden of retraining"
10:34:08 From Jeff Miller Brian reflects that groups of senior citizens in his local community are good at networking and co-organizing priorities.
10:36:20 From Jeff Miller Marc connects Leo's in Superior with Brian's "community party" example as a way of making human connections for networking and later working together. Marc provides an example from Kerry telling Chris about a British pub - a "public house' as a community living room - to make this sort of connection. "Third Place Books" <- an example of a deliberately intended community living room.
10:38:36 From Jeff Miller Brian recalls a mentor advising: "You have to go to where the players are if you want to be part of what's going on. Or maybe you organize the community gatherings yourself."
10:40:20 From Jeff Miller Paul gives an example of a local pub run by the local builder, who, if he comes back late from work, the locals know where the key is, and where the tin for paying is.
10:41:26 From Jeff Miller Brian relates a story about the local coffee shop where a group of farmers would come in and talk together every day; this is a place where Brian might have started to connect with the community.
10:42:52 From Jeff Miller Innovations in terms that people can understand (Brian's story) - "a firefighter can stay in a fire site twice as long, or carry half as much weight" - target communications to an audience and use their jargon, not "we can store twice as much oxygen".
10:44:12 From Jeff Miller Or address documents to the people who will be working with - (Brian's thought about Eric's incident report document) - that the document was written by high-level execs and managers ("cost of the outage") but aimed toward having the operations team fill it out ("so, how do we figure this out?")
10:46:03 From Jeff Miller Brian considers the center of interest for the FedWiki zoom and chat groups - broadly, FedWiki; but amping it up toward more specific project deliveries would likely narrow and fragment the members of the group.
10:47:24 From Jeff Miller Marc reflects that it's easy to get people to pay for coordination (not stepping on each other's toes) and difficult to get people to pay for cooperation (making specific choices to work together, as a commitment).
10:49:25 From Jeff Miller (Marc and Brian discuss the limitations of compatibility, and the consequences of workflows which depend on new and non-general item types requiring plugins) [I'm thinking that the security model is already a point of shattering around FedWiki's federation]
10:51:06 From Jeff Miller "consortium" of standards?
10:53:04 From Jeff Miller Eric nods to Brian's suggestion about standards as somewhat timely -- that Eric has been surveying various basic protocol specifications which have appeared as important, on reflection.
10:54:53 From Jeff Miller Eric reflects on the environment of standards for how the Java environment communicates with external resources - JDBC, etc -- does Spring still use JDBC? Is it still at the core? Yes. And how about JNDI (service discovery) and JINI (local network discovery)?
10:55:04 From Brian Marc, Eric is talking about resolving names, ie.g urls or fedwitki sites.
10:55:36 From Jeff Miller Eric also relates the various ontology languages in the world, like JSON-LD, a formalized update of the Microformats approaches to things like blog syndication.
10:56:09 From Brian https://json-ld.org/ https://docs.oracle.com/javase/jndi/tutorial/getStarted/overview/index.html
10:56:49 From Paul Rodwell https://atproto.com/
10:56:52 From Jeff Miller "We gave up on XML -- some of us -- and picked up JSON as the format to use" - "I went back to looking at things like SOAP and CORBA, as previous examples of good and iffy/painful attempts in the same area."
10:57:14 From Brian https://fedimeister.onyxbits.de/blog/bluesky-at-protocol-vs-activity-pub/
10:57:53 From Jeff Miller "The At Protocol Roadmap" from Paul's link: https://atproto.com/blog/2023-protocol-roadmap (authentication, account migration, federation)
10:58:55 From Paul Rodwell Paul Frazee
10:59:14 From Jeff Miller Paul relates some of the history / pre-history of the At Protocol related to Paul Frazee and the Beaker Browser". IPFS -"Inter-Planetary File System" - aligned somewhat with At Protocol and Bluesky's federation work.
11:00:56 From Jeff Miller Eric considers "what is the specification for how Federated Wiki interoperability and federation?" - because Eric has considered alternative ways of organizing the content: so what is the essence which allows interoperability? Eric relates: "Ontology doesn't seem to be the right path for me -- I have to have everyone participating maintaining the JSON-LD to a compatible spec level"
11:02:48 From Jeff Miller Eric nods that, say, FOAF dot org has to be maintained by someone for ontologies to be useful during transition to a JSON-LD scheme more broadly. Using all of these global ontologies depends on groups like Google Auth, like Twitter Auth, like the certification authorities. [Did Matrix do it right-ish? Somehow?]
11:03:46 From Paul Rodwell between Beaker and BSky, Paul was working on a personal cloud - https://github.com/atek-cloud
11:04:27 From Jeff Miller Eric notes that a thing which looks interesting about the At Protocol ... well, what we're wanting to maintain here in Federated Wiki is a distributed database of pages. The pages need to be able to evolve, but they also need to be able to talk to each other.
11:05:31 From Jeff Miller Eric considers David Bovill - he can do anything he likes for his wiki client, in whatever language he likes; but the common point needs to be in the structure of the page JSON - we have a fuzzy spec, here starting from Ward's descriptions; we also have a sitemap JSON, we have Assets folders; those are the core pieces of the data model.
11:05:44 From Eric Dobbs https://atproto.com/guides/overview
11:05:55 From Jeff Miller We're also pretty dependent on domain names, but everybody is. - Eric
11:06:18 From Brian In traditional client-server, those formats and protocols are enforced by the server. In fedwiki, there is almost no server, it's largely a file server (of json files) and everything is in the client. So then when everyone has their own clients, how to keep clients interoperable with each other as the various clients evolve.
11:07:35 From Jeff Miller Eric reflects on the At Protocol - "we'll use domain names, but also the W3C decentralized identities -- sort of a brand attached to cryptocurrencies, even if that's not the significant connection; there are signed repository records from the domain holder -- this is missing from FedWiki. It could be useful for adding a cryptographic signature for the pages I publish, to make them hard to forge as my words. Paul Frazee's description of these mechanisms on DAT before it became Hypercore points out to a way which some of this attribution could be worked out.
11:08:54 From Brian A fedwiki block chain for permantent storage is an interesting idea too.
11:09:33 From Jeff Miller Eric considers attribution of wiki content history by tracing them across history; maybe there are old copies of the page, but the domain is lost; the signature can be particular to the author (W3C decentralized ID) preserving the autonomy of the author. "You can take the local copy of your own data to a new server on a new domain."
11:09:59 From Paul Rodwell Earthstar has some of the same things - https://earthstar-project.org/
11:10:12 From Brian A fedwiki foundation could "own" that blockchain and the resource to serve it.
11:10:36 From Jeff Miller Eric hesitates to work too closely with the At Protocol, because its heart is to replace something like Twitter; the common factor is giving the author autonomy and power over their own content, and not go go to Twitter scale.
11:11:44 From Jeff Miller Search and neighborhood are enabled by fetching things from the federation. Including pulling over indexes from other servers, like (for example) Ralf's servers.
11:12:20 From Peter Thanks for the Earthstar link, hadn't heard of them
11:12:34 From Brian The utilization of assets and what to embed in paragraphs is another thing that a "spec" could address.
11:13:03 From Jeff Miller Paul points out that a certain percentage of the large indexing relates to things that are not text and should not be indexed in the same way. A spec could also say what items are subject to indexing, or what to do with less-obvious item types.
11:14:09 From Jeff Miller (Because embedded SVGs and diagrams are pretty wonderful, actually) Eric points out that the simplicity of Ward's data design offers opportunities for experimenting with different ways to compose wiki content which are not a long reach.
11:15:27 From Jeff Miller Eric reflects that the complexity of the federated links - the magic there - is the interesting point for how to preserve. (a discussion among Marc, Brian, and Eric about failure modes for federation) "a failure to whom?" (thinking of, say, could Archive dot Org render wiki content somehow?)
11:16:45 From Peter How the internet archive would show this in the (not-so-far) future
11:16:46 From Jeff Miller * archive-as-HTML (one sort of affordance) * enable forking by members of the community (interaction, get a wiki instance into the neighborhood) * how much support of the Javascript liveness is needed?
11:17:43 From Peter We're depending on browser ecosystem to remain viable (no Internet-Exploder or Chrome balkanization)
11:17:49 From Jeff Miller * losing the context required for dynamic operation and federation, losing the metadata outside the page content itself. I think of the University of Minnesota Geometry Center dying (in some sense) because several Mosaic features were discontinued, like the sketch/jot panel.
11:19:37 From Jeff Miller Brian says: "could I create static-page versions of wiki content?" Eric says: "yes, though there are some wrinkles to work through to make the content consistent." Brian reflects: Pandoc as Markdown renderer, with my own flavor, I can generate all the content in the way which I prefer; it's easier than writing those plugins for FedWiki; my underlying simple file format is also Markdown rather than Json.
11:20:39 From Jeff Miller MASFWiki or Obsdidian Obsidian MASFWiki == Obsidian ("Markdown with front matter") MassiveWiki
11:21:23 From Eric Dobbs https://massive.wiki/
11:21:46 From Jeff Miller right, Peter Kaminski, I've interacted with him around MASFWiki. "GollumWiki" an example implementation Markdown Shared Versioned Files
11:22:55 From Jeff Miller (I am absolutely attracted to that for a personal wiki format) fine-grained paragraphs allow dragging-around affordances
11:24:26 From Jeff Miller Federated Wiki's structure and idioms of dragging paragraphs around ("page as data source" - Brian) plays with some contrast to Markdown as a lump of formatted text. Thompson Morrison's paragraph-formed wiki pages, a paragraph as atomic unit of ideas, "Separation of concerns with regard to prose" - Eric
11:26:21 From Jeff Miller "Not clear if the paragraph is the right unit of granularity." - Brian (relates an anecdote of how large a block of knowledge is, naturally to the form?)
11:26:27 From Marc Pierson It would be interesting to find a way to import FedWiki lineups into Ulysses.
11:26:52 From Jeff Miller "If you want the page to be an essay, it may not work to have each paragraph be separable as a tweet." - Brian. Eric reflects on GraphViz diagrams, with their clickable links, as high-density entanglements of ideas, which can summon many other pages.
11:28:31 From Jeff Miller Brian reflects on the form, where it's easy to digest three things at a time (Marc: and graphs let you digest five to seven things, using color or grouping to keep down to five to seven things) - and then you can compose the three things Once you have a use for an overview, maybe you do want to have a hundred nodes on the screen. - Brian.
11:29:35 From Jeff Miller Marc reflects on the center of attention needs a continuous triple awareness "the thing I care about, the thing it's a part of, and the things it's composed of".
11:30:40 From Jeff Miller Brian reflects on Tufte's discussion of web pages, like the National Weather Service's page arrangement for the public landing page, increasing in complexity and detail and scope as you scroll vertically. (sounds like the Inverted Pyramid of journalism)
11:31:46 From Jeff Miller (Marc suggest that Brian's insights on Tufte's methods might be useful to unfold and reflect on, for usability purposes) "ordering of the columns, ordering of the data rows, which don't necessarily need to be ordered by column one"
11:33:15 From Jeff Miller Kafka: The Trial ? :) (messaging infrastructure)
11:34:03 From Brian Maybe a paywall...for Christmas movie anaysis by NYT. https://www.nytimes.com/interactive/2023/12/23/upshot/hallmark-lifetime-christmas.html
11:35:55 From Marc Pierson Brian, I suspect that I should better understand Tufte’s work.
11:35:57 From Jeff Miller Paul Frazee, https://github.com/pfrazee and Beaker Browser (archived) https://github.com/beakerbrowser/beaker
11:36:45 From Marc Pierson Anyone else use Ulysses?
11:37:03 From Jeff Miller (though the cool kids are using DJot, from one of the Markdown core implementors, since it's unambiguously parse-able)
11:37:08 From Brian Looks like Ulysses is a Apple thing. So I haven't .
11:37:23 From Paul Rodwell Replying to "Paul Frazee, https:/..." avoid the beaker-ng follow on, from the same place as Thorium with some of the same problems.
11:38:56 From Brian I think tables and graphs are isomorphic, but not sure they are usefully isomorphic.
11:39:05 From Paul Rodwell Replying to "Looks like Ulysses i..." has a subscription model, so no as the trial period is short
11:39:08 From Peter Apple is the 800lb gorilla here and lots of secrecy / proprietary stuff hiding
11:39:29 From Jeff Miller I got tired of gorilla wrestling It's just tough on open source developers.
11:40:02 From Brian For example, a table with from, to, edge could encode any graph...but that turns arrows into paths, so may not be as useful for communication.
11:43:00
11:43:00 From Jeff Miller Wrap-up history of the conclusion of Beaker Browser: https://github.com/beakerbrowser/beaker/blob/master/archive-notice.md
11:44:39 From Jeff Miller (The challenges are interesting to compare with FedWiki's challenges -- some are similar [peer to peer state], some are just different) (locally built compact data indexing was a point of overlap)
11:47:11
11:47:11 From Brian Blockchain is cheap...it's the "proof of work" part tied to the "coins" part that is expensive.
11:47:22 From Jeff Miller nod "distributed ledger" - interesting "distributed ledger with proof of work" - eh? "don't copy that floppy!" - bits vs atoms
11:48:38 From Brian Also it rate limist the growth of the pool...
11:49:32 From Jeff Miller distributed ledger with proof of work == algorithmically fascinating, behaviorally and organizationally eh.
11:49:40 From Brian We care about the "authentication" part, not the secrecy part.
11:49:42 From Jeff Miller "signing my work" "signing my work" x "at a particular time" probably of interest "at a particular time" maybe shows up on the signed ledger
11:50:37 From Peter encryption is now fairly cheap ... curren fedwiki "distributed ledger" doesn't have any accountability via cryptography
11:51:46 From Jeff Miller public key / private key cryptography being a core for cryptographic authentication these days. "idling the old Buick to solve Sudoku puzzles you can trade for illicit drugs" == cryptocurrency mining
11:55:55
11:55:55 From Jeff Miller Eric points out some risks of fragility and future-proofing, if there are customizations of the wiki format and protocol (easy affordances, easy customization) -> create a book using FedWiki -- that these affordances might be difficult to preserve. "your page won't render' <- the Internet Archive "your page must interoperate" <- Wikipedia
11:56:36 From Brian I supposed a foundation blockchain would have a moderation issue... :/
11:57:01 From Jeff Miller Mastodon is already a protocol where the federation comes and goes for moderation reasons.
11:57:47 From Brian In my perfect world, we'd just be writing LaTex anyway....lol...
11:57:56 From Jeff Miller "the simplest thing that could possibly work" - one-day computer projects (in Eric's reflection)
11:58:02 From Peter aaaaaargh latex
11:58:16 From Jeff Miller Ward will write A PERFECT TINY GEM which is complete in essence.
11:58:45 From Brian Having a concrete real example of a problem is invaluable to solving the right problem.
11:59:18 From Jeff Miller "the code is simply and closely attached to the real use cases" - Eric's reflection - "which is why such a small group of people can keep FedWiki running" "It's more powerful, the practice of simplicity, from Ward, from his experience of tiny simple solutions." - Eric EJBs, I missed out on them happily
12:00:45 From Brian Great book for people that draw stick figures...https://www.amazon.com/Drawing-Right-Side-Brain-Definitive/dp/1585429201 Anticipate that there will be a version 2 and then a version 3, and be flexible for that change...
12:01:57 From Jeff Miller stick figures and stay there: Ed Emberley's books. or make a career of it: Tom Gauld is totally an Emberley style.
12:03:29 From Jeff Miller my very first web project: converting Vax Runoff docs to a static website. Migration, migration, migration. US-ASCII text was the general anchor in common with Runoff and HTML. export is BIG for images image editors are a big ecosystem
12:04:40 From Brian Exporting a page to Markdown might be pretty useful.
12:04:44 From Jeff Miller though there are a few favored formats and a few core binary libraries. we're at the hour
12:07:33
12:07:33 From Jeff Miller a nice clean format there FedWiki as flexible composing environment Export of Markdown and Markdown to HTML as publishing chain
12:08:57 From Jeff Miller Marc says that the combination of many sorts of plugins composable in FedWiki is a win over some of the niceties of Ulysses.
12:10:27 From Jeff Miller Brian's goal: sharing information as a website; not using all of the responsive aspects of FedWiki, the data manipulation of a live site. "wiki as a database to store information, a script to digest it into a static website" NeoVim - a programmer's text editor, nvim
12:11:56 From Jeff Miller EDLIN.EXE