Text extraction. See Typescript Archive
This typescript includes SVG diagrams, both as source and rendered, from Eric Dobbs' Sofi diagram import as SVG with enrichments.
09:34:13
09:34:13 From Jeff Miller a Genghis Khan maneuver there (traceable in the Y chromosome) or fission them is it https? paniolo ?
09:35:54 From Jeff Miller https://www.nps.gov/havo/learn/historyculture/paniolo.htm the NPS on Hawaiian cowboys competing in a Wyoming rodeo in 1908
09:39:06
09:39:06 From Jeff Miller Gerald Weinberg, the Law of Strawberry Jam; the Law of Raspberry Jam.
09:39:11 From Brian Jam analogy is really stretching... Maybe it's an instance of Raspberry Jam.
09:41:26 From Brian I wrote an email to a friend at work, how everything that management was doing could be mapped onto Demmings 7 deadly sins...who then forwarded that email to a large group of influential people...I got no fall out and no idea if anything changed...
09:41:44 From Jeff Miller Eric reminds us of W. Edwards Deming's Total Quality Management, where it's possible that he's misinterpreted by 2nd or 3rd-order readers and exponents.
09:42:32 From Brian I find that for junior staff, focusing on performance and tech progression is useful. For senior staff, focusing on how they adhere to values is a better metric. Both are subjective.
09:42:46 From Jeff Miller "evaluating individuals on individual performance defeats the principle of systems thinking"
09:44:39 From Marc Pierson Might be interesting to read Peter Robertson’s Always Change a Winning Team.
09:45:14 From Brian There is a difference between roles and individuals...I find that thinking about roles can be more objective, but seems like everyone focuses on individuals.
09:45:36 From Marc Pierson The Scrum Master seems to be the magician.
09:46:05 From Jeff Miller (discussion of evaluating components and evaluating feedback loops)
09:46:20 From Brian I've had way more negative experiences with Scrum Master's than positive ones...Scrum master was seen as a power position for micro managing... :/
09:46:33 From Jeff Miller Bohmian dialogue rules -- 2 1/2 times around the group discussion, I always learned enough to blow my mind - Marc P.
09:47:40 From Paul Rodwell https://en.wikipedia.org/wiki/The_Tyranny_of_Structurelessness ?
09:48:15 From Brian And reviewed against the values of the community. to provide the accountability.
09:48:29 From Paul Rodwell Text -> https://www.jofreeman.com/joreen/tyranny.htm
09:48:32 From Jeff Miller Paul, yes, that one.
09:49:11 From Brian Makes it written down too. No closed door decisions.
09:50:12 From Jeff Miller although it may loom larger in my mind as a type example where an incisive critic points to the nature of a group of equals -- equality of access and knowledege must be maintained so that all participants participate equally.
09:52:14 From Brian "better way". Continuous improvement through iteration, feedback, reflection, etc.
09:52:30 From Jeff Miller Marc reports from Superior AZ, a slow "syntegration" committee of five people around each of about six issues / topics of interest for the community that Chris Casillas is working with. Ward recalls a blog post about an offsite where there are writers and talkers present and the talkers felt they had reached consensus while the blog author left uncertain that anything at all had been agreed.
09:53:32 From Brian The value of those intense workshops is the relationships you form with the people. The tasks/topic is pretty irrelevant.
09:53:35 From Jeff Miller "I'm a listener" says Ward. I'm not convinced that the relationshps are really established unless you've created some work product together. (to Brian's observation)
09:55:16 From Brian Another similar system has 2 axes. One is Influence by asking or by telling. Other one is task oriented vs people oriented.
09:55:19 From Jeff Miller do people study a problem deeply before committing computer code? / or do people start working with the ideas as code, (like a potter with clay) ? and also the question of: do you REALLY need that horizontal scaling and all its failure modes?
09:56:25 From Jeff Miller GitHub Copilot is reputedly really good about remembering boilerplate
09:56:39 From Brian I find that spiking parts of the solution as early as possible is really beneficial. Need to get feedback from the rubber meets the road part of the problem. Especially the parts of the program that I know the least about (hence the most uncertainty and possible risk).
09:56:40 From Jeff Miller like iterator-in-Python vs. Ruby vs. Go vs. ....
09:57:03 From Paul Rodwell do…while or while…do or something totally different
09:57:23 From Jeff Miller I've forgotten what it feels like to be confident in a language, and "The Programmer's Brain" is chiding me that I should be doing my exercises. (discussion of how to compose various languages related to subprocess or function call exit status)
09:58:12 From Brian intellisense and hit the . key..?
09:58:33 From Jeff Miller yes, that or whatever Eclipse or IDEA calls it. microclippy as a hotkey
09:58:47 From Paul Rodwell http://www.clojurekoans.com/
09:59:40 From Brian The web is getting increasingly less useful...not sure if that is because the way I use it isn't aligned, or if the information is diluted and I need to use other methods/techniques.
09:59:43 From Jeff Miller (sounds like when you look up a recipe for making food -- they're terrible user experiences) Cory Doctorow seems to have put his finger in it.
10:04:12
10:04:12 From Paul Rodwell Like the cooking travelogues - thinking of Stanley Tucci series on Italy
10:04:59 From Jeff Miller at Pivotal Labs! (alas, I didn't get the chance)
10:05:05 From Brian My recipe http://paniolo.xyz:8085/view/welcome-visitors/view/pancakerecipe
10:05:30 From Jeff Miller It sounds like a Jim Weirich-ism, Ruby Koans.
10:05:48 From Brian Docker is great option for that. OpenJDK isn't too bad... npm is the worst.
10:06:53 From Brian I tried to remove python from Linux...and that didn't work out.
10:08:11 From Brian You'd have been fed up way before then, don't worry.
10:10:42
10:10:42 From Brian Prior to Spring?
10:10:48 From Jeff Miller Apache Struts ~ CGI for Java Spring is super late in the game SUPER late
10:12:15 From Brian I was forced to write java for a couple of years...Not the funnest project I've ever had. I think JVM/Java/Scala is in a much better place today. A version of the Google's problems are not most people's problems.
10:14:05 From Jeff Miller EJBs - the "Tyrant" from which RPC-over-HTTP was a rebellion (first things like WSDL, and then just pushing and pulling JSON) "write once, read never" - the Copybook approach to protocols
10:16:44
10:16:44 From Eric Dobbs I’m glad for my Java experience. It’s where I got introduced to distributed systems and concurrency by a DBA who explained why transactions were important and got us to use a JMS message broker. I’m also glad for early exposure to service oriented architecture and JAR and WAR package and deployment schemes. Also very glad to have that early exposure to Inversion of Control as an idea. All that stuff has been re-created in containers and Kubernetes.
10:18:43 From Jeff Miller +1 (to Eric)
10:20:14 From Jeff Miller The challenges of documentation, especially documentation intended for incident response, are such that the docs don't themselves highlight the most relevant points of orientation.
10:22:19 From Paul Rodwell and documents that don’t get updated cause there never any time put aside for it.
10:22:27 From Jeff Miller Pete recounts the typical scenario where if you don't flip datacenters frequently enough, then things can drift such that the evolution doesn't go smoothly. (the longer between migrations)
10:24:37 From Jeff Miller (Eric recounts a comparison between Ward's WyCash experience -- where the business managers were unable to micromanage the developers -- as long as the developers were producing incremental updates responsive to business needs, that was good enough)
10:26:53 From Jeff Miller (a comparison to Eric's experience pair programming with a programmer who had come from a very different point of experience; his disputes on the code with the pair on the code were strongly positive, in that there was good results in the message flow code)
10:29:19 From Marc Pierson I am having fun integrating Meadows and Ostrom in the EIP framing.
10:30:21 From Jeff Miller "with a small amount of refactoring, wrestled between Eric and his pair for design, the central message system could then be taken forward by other team members, and improved to land a big contract using the same core message flow tuned for a higher speed and flexibility" paraphrase
10:31:02 From Marc Pierson Ward, I just stumbled upon your Start Playing with Wiki and saw that that format could should will be used to have fun while learning about the Ecological Recursions, the Political Recursions, and a particular Institutional network bridging the E and P worlds.
10:31:20 From Jeff Miller building domain language (Eric Evans' DDD) the origin of the DDD patterns for Eric Evans was being able to express the model as different sub-models
10:32:53 From Jeff Miller "bounded context" in DDD "context map "
10:33:27 From Brian I've found that before refactoring, I should always have 2 or 3 examples/use cases first. THis often results in that I copy and paste 1 or 2 times before I refactor. And depending on how it's used and evolves, I may never refactor. Another rule I generally follow now is that functions can only have 1 level of flow control in it. This tends to result in pretty factored code.
10:34:58 From Jeff Miller "the hierarchy of the model comes from the computerization of the domain" (like Advancers in WyCash), an un-named term in the actual financial domain an Advancer lets you automate a common function that operates differently for different financial instruments an abstraction like an iterator for "the next, the next, the next..."
10:36:29 From Jeff Miller or more generally a container or collection (Smalltalk's Collection classes, later copied into Java 1.2 and Java 1.5+) The code base is obliged to name code objects which may not be distinguished with names in the non-automated practice.
10:38:07 From Jeff Miller Marc asks: "when you name something, and use it other places, is it pulling other parts of the world with it?" Ward says: yes, because there are layers built up that make the program work, and all the higher things assume a common lower layer.
10:39:40 From Jeff Miller Ward says: Smalltalk actually is composed in a way which allows making modifications at any level without breaking the relationships among them -- if done cautiously . When we discovered new things about money and dates, we might go back down to bottom layers.
10:40:50 From Jeff Miller Marc checks in: in discussions with the experts in the financial domain, concepts emerge that the programmers give names to, and sometimes there's a surprise which requires you to fix things down below in the multiple levels of supporting code. Ward: yes.
10:40:57 From Brian decomposing big problems into small problems is basically hierarchy. the hierarchy has DAG like properties, it's generally useful...I wonder if there is a better metaphor.
10:42:06 From Jeff Miller Ward says: it was three years into the product where we were told "we will buy the product if you handle multi-currencies". So we decided that there's money, your base currency (simple) and also Money, your here-is-a-basket-of-currencies; nothing broke as long as you were trading in dollars;
10:42:16 From Brian In programming, call tree is basically the hierarchy, which is enabled by the composition of all the factored parts (e.g. subroutines).
10:43:03 From Jeff Miller but if you wanted to make a report converting Money to money to draw a balance, you'd have to note the date of the exchange rates in order to make a report.
10:43:54 From Brian Money gets more complicated the deeper you dig. The common definition is woefully under defined.
10:46:43
10:46:43 From Jeff Miller Eric suggests that Ward's experience with WyCash was unusual in the field, that it was "just another feature" to support multicurrency where it was more typical for software teams to believe a feature like "support multicurrency" would have needed a new start to the entire system.
10:47:48 From Jeff Miller Eric checks in with Ward: the reports needing an "as of DATE" for reports to add up currencies, because of exchange rates, that was a feature that came in alongside multicurrency. Ward says: "you can set a default currency, like dollars or euros; you need a date, but you can default to today's date" (WyCash) did you want ${process}? I can sell you ${process}, enterprise mode!
10:48:57 From Jeff Miller (my notion about Eric's observation) Eric's observation: WyCash as a system which allowed resilient cascading of a change through all the layers, due to supporting multicurrency -- this is something that most designers and teams are unable to do.
10:51:18 From Jeff Miller WyCash had a supple composition of code (as a result of pairing and refactoring) which would survive a a cross-cutting design change such as multicurrency.
10:51:49 From Brian A bit off topic, but reminds me of this web framework in Ocaml. https://www.youtube.com/watch?v=h_e5pPKI0K4
10:53:58 From Brian the static vs dynamic is pretty mismatched in this case... If they embedded LISP, they'd have a chance...lol.
10:55:20 From Jeff Miller Ward describes a model of the Report object (aka the Report Writer) - that some views did need some special handling, things like the balance based on the application of each financial change, day by day. The concept of a "natural ordering" (advancing by date) was necessary for drawing a daily balance while also allowing sorting by the magnitude of the daily balance.
10:56:43 From Jeff Miller so: - you could have a Report - some computations are sequenced by natural order (temporal in this sense): - once this is clear, the effects on the way the code is is composed
10:57:52 From Brian Programming is 90% about understanding the problem and customer need, 10% about coding and syntax... :/ At least, that has been mostly my experience.
10:58:37 From Jeff Miller Late in the day, we introduced a Natural Order concept, which would fit gracefully into our models of the various columns; a cumulative column (the running balance) needed to be keyed to an order by date in order to be computed. The complexity was concentrated where it was needed, and you could reason about it.
10:59:11 From Eric Dobbs Agree! And I think what Ward & team were able to do is actually /spend/the/time/ when they learned something surprising to convert that slice of understanding into changes in the code.
10:59:17 From Jeff Miller "We have a hierarchy of column computations, and there are six of them that we do."
10:59:32 From Eric Dobbs Over and over again I see teams who learned something and then never get around to changing their code to reflect what they learned.
10:59:44 From Jeff Miller (nodding to Eric's point about /spending/the/time/ except it's m complaint in large)
10:59:53 From Brian Sort of a version of YAGNI, you aren't going to need it.
10:59:58 From Jeff Miller my complaint in large is that the technical landscape is unreadable to me the discourse changes faster than I can learn how to navigate it
11:01:27 From Jeff Miller Ward observes: the CUSTOMER demonstrates the software to the implementation team That's what a weekly demo is about. "We can delight ourselves by doing these things in this way" - Ward, on the FedWiki approach.
11:03:14 From Brian I read a fair bit of that content, Eric. :)
11:03:35 From Jeff Miller (Eric describing his experience trying to run FedWiki in Docker on his own machine; and also trying to run it in Kubernetes; but mostly there's not a big win for K8S or Docker, beyond plain old Unix hosting)
11:05:19 From Jeff Miller Jeff reflects that Eric's demo was useful because seeing the pieces moving and connected provided insights which were not obvious in the web inspector. "Does the project manager not have requisit variety to direct development?" requisite variety to direct development
11:07:05 From Jeff Miller Marc replies to Ward: when you have a hierarchical management structure (that's not just about domination); the point behind a hierarchical structure is to match decisions to decision makers at the same level. "Does it matter that some people in the game know more than others?" [I can taste the edge of the Bikeshed versus Nuclear Reactor decision making issues]
11:08:14 From Jeff Miller Bikeshed: everyone thinks they understand a bikeshed, it can be limitlessly debated based on everyone's preconceptions.
11:09:15 From Jeff Miller Nuclear reactor: only the experts at the table understand the details, the decision-making committee cannot second-guess the details, but only evaluate based on heuristics and checklists (reputation of the experts and their organization). Ward says: because Smalltalk was a single bundle, we could solve problems at the right level of the hierarchy.
11:10:35 From Jeff Miller Marc says: the cool part of the cell, where the decisions are made, are more significantly at the cell surface, and not so much the DNA as central master code.
11:10:42 From Brian And now you have a queue with related rates and is Turing complete. :)
11:11:02 From Eric Dobbs Riffing on Marc’s story about cell membrane being the interesting part of a cell… Saotome Sensei describes skin as the original brain. (https://en.wikipedia.org/wiki/Mitsugi_Saotome)
11:11:11 From Jeff Miller 100% running-system tests, 0% unit-tests, lots of copy and paste reuse. (proteomics - what cells are doing based on what sorts of proteins they're using) Marc describes the way that cells use the same signaling mechanism in wildly different contexts.
11:15:37
11:15:37 From Jeff Miller (Jeff mentions a Portland-based chestnuts - as - main - calorie - crop project: "Build Soil - Plant Chestnuts) Marc asks what's new in FedWiki land
11:17:12 From Jeff Miller "The Hunt for ReLocalize" - why did Marc's site disappear from the FedWIki index? (if I got that correct as the account of the ReLocalize problem) "Is it a pile of old scripts with interesting behavior?"
11:18:07 From Brian I've seen things on the WebCrawler search, the registration on farm thing for students. Eric had a editor-Preview thing. I'm dabbling with what it means to move wiki's to another place, and possibly turn a fedwiki into a markdown/static html site (I'm thinking for permanent archive)
11:18:24 From Jeff Miller "domain names that had the word 'local' in it, like 'relocalize'"
11:18:50 From Brian I've been looking for an OpenAPI spec for fedwiki...might consider trying to make it, if it doesn't exist.
11:19:02 From Jeff Miller Eric's Editor-Preview based on Observable's message passing system and HTML item renderers was pretty dramatic.
11:19:08 From Brian Ralf has been working on an Elm powered client.
11:19:46 From Jeff Miller Elm: https://guide.elm-lang.org/ "a functional language that compiles to JavaScript" It's pretty exciting that alternate clients look like they're within close reach.
11:21:18 From Jeff Miller oh
11:21:33 From Paul Rodwell http://ward.dojo.fed.wiki/view/search-index-logs
11:21:51 From Jeff Miller is that the same "found" as "found.ward.fed.wiki" ?
11:22:41 From Brian Did you pair program that script? ;)
11:22:59 From Eric Dobbs More like code review than pair programming.
11:22:59 From Jeff Miller a.b.c.local <- don't index this localhost <- don't index this relocalizecreativity <- index this!
11:23:42 From Paul Rodwell a.b.local:3000 <- don’t index this either
11:24:23 From Brian Don't underestimate the value of having a "progress map", wrapped in a metaphor, for story telling or encouraging an activity.
11:24:24 From Jeff Miller EIP diagram Environment - Institutions - Politics Paul: right! that .local can be not at the end of the server:port confection
11:26:20 From Jeff Miller Brian "a progress map wrapped in a metaphor" ? - like Marc's discussion of the EIP diagrams, particularized to individual things that the person already knows - the example reveals what the EIP diagram is doing / showing. ("Brian" there was "a question for Brian")
11:32:22
11:32:22 From Jeff Miller Brian says: "before you get into a story like The Hobbit, there's a map where you can refer back from the book into the map -- so having the eight-story guide going from floor to floor up the escalators, what Ward put together; our pointers go one way, like escalators do, but we'll find ways to go back down floors (etc)"
11:33:58 From Jeff Miller (Eric demonstrates linear SVG gradient colors done in the Observable / FedWIki preview explorer) view-source:http://aristobit.com/svg/radar.svg (me fooling with a SVG)
11:37:12
11:37:12 From Jeff Miller Eric's punchline with his SVG linear gradient demo: this will allow enriching SVG import nodes with some statistics, like the hard-edged linear gradients from another visual rendering ... ... that you can take the SVG linear gradient mechanism and use it to render a Sofi diagram with an aggregate roll-up of votes from assessors
11:38:24 From Jeff Miller The trick in the programming is that each linear gradient needs a separately named and ID'd gradient and if you have more than one Sofi diagram, then you have that many times eleven
11:39:53 From Brian Do you want the gradient aspect? If not, a rectangle with a circle mask might do similar thing.
11:40:15 From Jeff Miller Voila! (Eric shows a colored circle with green, yellow, and red horizontal stripes representing the Sofi vote roll-up for that node)
11:41:22 From Jeff Miller Eric considers whether the Sofi roll-up SVGs should go into the Assets page, carefully coordinated with the Journal, because the SVGs could otherwise bloat the Journal of a page.
11:42:36 From Jeff Miller Eric observes that the ability to preview SVG rendering in this way is a happy side effect of the prototype work on using Observable as a FedWiki client framework.
11:46:23
11:46:23 From Eric Dobbs
<svg viewBox="-100 -50 200 100"> <style> .one {stop-color: hsl( 0deg 90% 50%);} .two {stop-color: hsl( 60deg 90% 50%);} .three {stop-color: hsl(120deg 90% 50%);} .four {stop-color: hsl(180deg 90% 50%);} .five {stop-color: hsl(240deg 90% 50%);} .six {stop-color: hsl(300deg 90% 50%);} </style> <title>10% urgent 45% worried 45% happy</title> <linearGradient id="grad" gradientTransform="rotate(90)"> <stop offset=0% class=one /> <stop offset=16% class=one /> <stop offset=16% class=two /> <stop offset=65% class=two /> <stop offset=65% class=three /> <stop offset=100% class=three /> </linearGradient> <circle r=50 fill="url(#grad)" title="fizbuz" /> </svg>
_source below is pasted in as Markdown_ SVG source for the enriched Sofi node with green, yellow, and red indicating Sofi voting or scoring of this function. ``` <svg viewBox="-100 -50 200 100"> <style> .one {stop-color: hsl( 0deg 90% 50%);} .two {stop-color: hsl( 60deg 90% 50%);} .three {stop-color: hsl(120deg 90% 50%);} .four {stop-color: hsl(180deg 90% 50%);} .five {stop-color: hsl(240deg 90% 50%);} .six {stop-color: hsl(300deg 90% 50%);} </style> <title>10% urgent 45% worried 45% happy</title> <linearGradient id="grad" gradientTransform="rotate(90)"> <stop offset=0% class=one /> <stop offset=16% class=one /> <stop offset=16% class=two /> <stop offset=65% class=two /> <stop offset=65% class=three /> <stop offset=100% class=three /> </linearGradient> <circle r=50 fill="url(#grad)" title="fizbuz" /> </svg>7 ```
11:47:00 From Brian can wrap it in ```html markdown block...
11:47:31 From Jeff Miller right, although that would be a manual operation -- copy and drop it which I can do right now
11:47:56 From Brian ```html
11:51:00
11:51:00 From Eric Dobbs Andrew, Brian, this live editor is out on the ‘Net over here if you wanna kick the tires. https://css-spikes.dbbs.co/pwa/ And source code is https://github.com/dobbs/wiki-spike-css/tree/main/pwa
11:52:01 From Eric Dobbs I think the source code is probably still pretty tightly coupled to my own brain and not quite readable.
11:52:46 From Brian Thank you Eric. Figuring out plugins and the concept is one of my future goals, but not there yet.
11:53:58 From Andrew Shell So long folks! Gotta go. Hope to see y’all again soon. :-)
11:55:05 From Jeff Miller (Marc describes saving JSON from a wiki page and dragging it up into a browser tab, in the same way that you can drag in an isolated SVG to zoom into it or provide a wider view)
11:56:27 From Jeff Miller (a discussion follows of a wiki story as a document type which the browser doesn't know about, though we could use a browser plugin) (a browser plugin embedding wiki client rendering logic if it recognizes a wiki page JSON document)
11:58:19 From Jeff Miller Eric shows how, in the prototype Observable browser, how to re-set the width of a particular wiki page column rendering to allow what Marc was interested in -- you could use the Firefox or Chrome web inspector to temporarily re-size a page in the lineup.
12:00:34 From Jeff Miller Eric describes the process by which he expands a given wiki page column in the web inspector by editing the definitions of the panel. The example that matches up with the demo is the Supercollaborator, a narrow edit column on the left and a wide rendering canvas on the right.
12:01:42 From Jeff Miller Marc describes a "Tom Sawyer" UI for rendering graphs in a similar way to left-as-navigation, or left-as-editor-source, and right-as-rendered view could be useful for some Wiki cases
12:02:23 From Brian I like the idea of the toggle button, but maybe with defaults based on context.
12:02:52 From Jeff Miller Eric speculates whether an expanded page should be controlled by a widget on the column, or in some other way, like data within the page: "here's a generated ghost page of double width, do you want to fork it in?"
12:04:14 From Jeff Miller Marc reflects that the EIP (environment institutions politics) diagram synthesized Donella Meadows and Elinor Ostrom's work into a single diagram.
12:06:14 From Jeff Miller Paul compares his editor-plus-renderer prototype as a double-width column in the Observable framework wiki client. "you should be able to drag both of these together, once dragging is available"
12:08:41 From Jeff Miller "save chat!" (Marc and Paul discuss how close they are in England; Marc may map out where he goes on a wiki pin)
pasted addenda
SVG rendering using Eric Dobbs' prototype of an enriched SVG node representing a node on the Sofi diagram of a system, colored according to the Sofi scoring of the participants.
SVG source for the enriched Sofi node with green, yellow, and red indicating Sofi voting or scoring of this function. ``` <svg viewBox="-100 -50 200 100"> <style> .one {stop-color: hsl( 0deg 90% 50%);} .two {stop-color: hsl( 60deg 90% 50%);} .three {stop-color: hsl(120deg 90% 50%);} .four {stop-color: hsl(180deg 90% 50%);} .five {stop-color: hsl(240deg 90% 50%);} .six {stop-color: hsl(300deg 90% 50%);} </style> <title>10% urgent 45% worried 45% happy</title> <linearGradient id="grad" gradientTransform="rotate(90)"> <stop offset=0% class=one /> <stop offset=16% class=one /> <stop offset=16% class=two /> <stop offset=65% class=two /> <stop offset=65% class=three /> <stop offset=100% class=three /> </linearGradient> <circle r=50 fill="url(#grad)" title="fizbuz" /> </svg>7 ```