Text extraction. See Typescript Archive
09:11:06
09:11:06 From Jeff Miller some Shakespearian phrase about business models there, probably from Hamlet (mainframes as data-efficient platforms hobbled by licensing models)
09:12:12 From Paul Rodwell fond memories of CODASYL DBs
09:14:36 From Jeff Miller (an earlier discussion about a personal wiki devoted to a list of programs and what was interesting about them an what was learned)
09:15:50 From Paul Rodwell https://en.wikipedia.org/wiki/CODASYL and https://en.wikipedia.org/wiki/Content_Addressable_File_Store
09:17:21 From Jeff Miller COBOL "Move Corresponding" https://www.mainframestechhelp.com/tutorials/cobol/move-corresponding.htm using COBOL for clear and beautiful code expression
09:18:18 From Brian Murach's Mainframe COBOL?
09:18:45 From Jeff Miller 1978/9, about the first time I saw COBOL, when I was an office assistant in 7th grade, running a plugboard phone switchboard
09:18:46 From Paul Rodwell clear until ops drop the card deck.
09:19:46 From Eric Dobbs 1979ish https://en.wikipedia.org/wiki/Atari_8-bit_family which I chose over an Apple II ‘cos I liked the BASIC editor better.
09:20:00 From Jeff Miller <3 TRS-80 here, with the fancy dot graphics
09:20:50 From Peter Dimitrios Rectangular pixels
09:20:56 From Chris C https://youtu.be/tKP-FrBi3UI
09:21:00 From Peter Dimitrios Character graphics
09:21:44 From Eric Dobbs Replying to "1979ish https://en.w..." Atari 400 specifically.
09:22:04 From jan d works excellent!
09:24:49
09:24:49 From Jeff Miller afk
09:25:05 From Paul Rodwell Replying to "1979ish https://en.w..." think I still have an 800 somewhere in storage
09:39:35
09:39:35 From Jeff Miller Chris shared a community progress report from Regenerating Sonora / Leo's / Superior, Arizona.
09:39:47 From Brian I'm just starting (and really enjoying) "You're invited: The Art and Science of Connection, Trust, and Belonging" and might be interesting to a few of us here. https://www.amazon.com/Youre-Invited-Science-Cultivating-Influence/dp/0063030977 The upshot so far is that your network defines you and your success in nearly every facet, including your weight, influence, wealth, etc.
09:41:45 From Jeff Miller Eric considers a model of skills for sustainability at his workplace. How can we connect with new technology appropriate to possible opportunities and skill-building around those new technologies? what skills are foundational knowledge for sustainable community interactions?
09:42:20 From Brian Or manage their own website... :/
09:42:42 From Jeff Miller Taking a leaf from Chris and Regenerating Sonora -- what skills are important to build within a community?
09:44:16 From Jeff Miller Marc reflects that Eric's considerations call out Kerry's thoughts from fifteen years ago about getting systems dynamics into schools - a heavy lift. "Make virtue easy, make sin hard" - from Dave Snowden -- make it easy to do the right thing, "Lear gardens" (?)
09:45:53 From Jeff Miller (in my last California house I had Earthbox containers which were easy to use for growing food)
09:46:40 From Brian Gift economy vs ?
09:46:57 From Jeff Miller Marc reflects on skill transfer at Chris, where the tools and skills were one of the neighbors -- "Bob's skills and tools" -- and using casual apprenticeship, a year later, the skills were transferred and being made use of. at Chris's Regenerating Sonora community
09:48:00 From Jeff Miller https://regeneratingsonora.org/ Chris says: "make virtue easy" by compensating community members for work that supports those community efforts.
09:49:18 From Jeff Miller Chris reflects for Eric: have your skills transfer process be an ongoing effort. "If it's worth doing, do it on an ongoing basis". Without these ongoing processes, things look good for a moment and then go away. (Jeff considers the every-Tuesday Silicon Valley Patterns Group community)
09:50:39 From Jeff Miller "we've got to go look at those numbers and see what's happening, do we know if the new customers are happy?" Aiming for sustainability in a growth-oriented system. growth-based versus income-based?
09:51:49 From Chris C It’s also a paradigm shift to see the world always changing vs a world that is frozen and fixed
09:51:59 From Brian There used to be tax incentives for reinvestment in the company. Basically if the $ stayed within the company, the money was retained and if it was extracted outsite the company, then it was taxed...
09:52:29 From Jeff Miller Has software-based enterprise forgotten how to nurture ongoing efforts?
09:52:39 From jan d scaling: I like "On Nonscalability: The Living World Is Not Amenable to Precision-Nested Scales" by Tsing, it is sadly a bit hard to read
09:52:59 From Brian I think there is a future where companies should be <200 employees and if there is enough scale, then form more companies and incentivize the companies to be cooperative to be efficient rather than being a monopoloy.
09:53:17 From Jeff Miller Does Regenerating Sonora's example of ongoing apprenticeship offer strong models for building ongoing skills and strangths?
09:54:18 From Jeff Miller to Brian's thought, W.L. Gore and Co. organizes units of about ~200.
09:57:17
09:57:17 From Jeff Miller Chris reflects on a virtuous cycle of work rewarded with results, challenges overcome to sustain the intentional effort, and ongoing success in that area of work.
09:59:45 From Eric Dobbs Don’t know if this helps, but the “near the headwaters” reminded me of a Colorado mining failure https://en.wikipedia.org/wiki/2015_Gold_King_Mine_waste_water_spill
10:01:42 From Paul Rodwell https://www.patagonia.com/stories/the-sacred-headwaters/story-18338.html
10:02:18 From Ward Cunningham Recent news: https://www.npr.org/2024/01/12/1224494403/klamath-river-begins-to-flow-again-with-dam-removal-project
10:03:31 From Paul Rodwell https://emergencemagazine.org/gallery/poisoned-beauty/ https://emergencemagazine.org/essay/saguaro-free-of-the-earth/
10:04:49 From Jeff Miller (Chris and Marc reflect on a copper mining proposal near the headwaters of Queen Creek, and approaches toward seeing the whole system and effects and risks)
10:05:51 From Jeff Miller structures as ongoing infrastructure costs - the Ruhr valley (subsidence over mined coal) just as much as the Netherlands. (Jan's example of "costs for eternity")
10:06:19 From jan d could only find the term in German: https://de.wikipedia.org/wiki/Ewigkeitskosten
10:06:22 From Brian There are all sorts of abandoned cities and amazing how fast the building deteriiate. For example, in Chernoybl.
10:07:10 From Jeff Miller the song Paradise -- "Mr. Peabody's coal trains have hauled it away."
10:09:05 From Jeff Miller Paul reflects on under-maintained infrastructure in the UK whose costs fall upon a future government - hospitals using aerated concrete construction with a short lifetime.
10:10:17 From Jeff Miller XVI The Tower as an ice block construction
10:11:33 From Jeff Miller Brian reflects on Eric's thoughts on knowledge of PKI as a basic skill -- Large Language Models are revealing humans as oversensitive to style over substance.
10:11:41 From Paul Rodwell Reinforced autoclaved aerated concrete (RAAC) - https://www.bbc.co.uk/news/education-66669239
10:12:18 From Jeff Miller Brian thinks: what will be the next good proxy variable, when gracefully expressed style becomes useless?
10:13:55 From Jeff Miller Ward reflects that media literacy in terms of a well-put-together site, or photography once tools for processing became cheap and easily available -- these tips became less useful when it became more easy to fake things which had previously been good signs of substantial content.
10:15:19 From Pete IMHO it all dovetails with continuum of human history - being good at the form of talking allows you to be a good salesperson / pitch-maker and get people to like you or get behind you. AIML can generate similar kinds of text and emulate some of that form
10:15:31 From Jeff Miller Ward reflects on a business plan review reading technique. Take the executive summary and write the assumptions present in the first sentence; the second sentence; etci. Write the assumptions in the margin. Cross the assumptions off as the business plan addresses the assumptions. If you run out of margin, the business plan is not a good bet.
10:16:52 From Brian auto tune...
10:16:55 From Jeff Miller If you want to avoid being scammed, you have to learn a lot of things, so that you can check them based your understanding.
10:18:11 From jan d "Prompt Injection"
10:19:07 From Brian And people have oversight...LLMs don't feel that so much... LLMs should be tools that help people, so really to use LLMs it's a person + a tool to do better quality...not a replace a person sort of thing. If a person uses an LLM, then they should be held to account...
10:20:21 From jan d Like the "no bird thing". Or wait, was not "Foucaults pendulum" by Eco about that?
10:20:44 From Jeff Miller "help a person, not replace a person" conspiracy theorizing as a disorder of the critical thinking function
10:22:02 From Jeff Miller (Eric's story: "the fact-checkers were clearly bribed") law and science dissolve if there's no common set of facts
10:23:07 From Brian Science is useful when it has repeatability... reproduceability.
10:23:18 From Marc Pierson Jan, re CEOs or the executive function. VSM suggest that there are three key functions: maintain identity of org, deontic (must, must not, may) and finally balancing investment in new vs current operations.
10:23:51 From Paul Rodwell A few bits from the current Risks digest - https://catless.ncl.ac.uk/Risks/34/4#subj2 and https://catless.ncl.ac.uk/Risks/34/4#subj4
10:23:57 From Marc Pierson The CEOs generally micromanage and fail to do the essential three.
10:24:19 From Jeff Miller Jan notes: if you check every fact and assertion, you cannot make any progress; if you trust every fact and assertion, you cannot maintain what you have or make any progress. (hopefully the second half of that is close enough)
10:25:21 From Brian Settle that 10% return is enough and don't need to squeeze every last $$$ out of everything....
10:25:22 From Marc Pierson I wonder when on should abandon old tech and start with new. We have children rather than 2000 year old adults.
10:26:23 From Brian From sunken cost fallacy, if you wouldn't invest in the old today, then invest in the new....I reliaze there might be some weighting necessary to evaluate that though. With nano technolgy and some health advances, we might have 2000 year old adults...
10:27:29 From Jeff Miller Eric reflects about the maintainability of a system like Ward's Eldorado - that the shape of the larger system needed to accommodate changes, that the graph queries need to be tweaked - so the skills of maintaining the graph model need to be practiced and maintained. practiced, maintained, transferred.
10:27:37 From Pete Mel Brooks 2000 year old man becomes a reality ? ;-)
10:27:51 From Jeff Miller venture capital skill profile site
10:27:55 From Brian linked in
10:28:08 From Jeff Miller oh! LinkedIn every week I go fish LinkedIn for jobs skills asserted, peers endorsing skills web of trust?
10:28:54 From Brian I bet there were companyies that would boost your rating, so presume it's all gamed too.
10:29:05 From Jeff Miller What does LinkedIn (skill assertion) in a corporate environment?
10:29:26 From Brian There is glass door that has some related kind of info....
10:29:37 From Jeff Miller (corporate directories can include self-declared skills - employee profiles can include them, Microsoft had that) query the corporate directory
10:30:00 From Pete "badges" are the current way companies track 'skillz'
10:30:11 From Jeff Miller Salesforce is heavy on skill badges
10:30:23 From jan d Replying to ""badges" are the cur..." Mozilla used to have some initiatives around that.
10:30:53 From Jeff Miller Netflix had common-interest groups ("guilds") which appear to have connected common concerns and common skills
10:31:45 From Pete external user groups / MeetUps can kinda become Guilds (often clustered around open-source software)
10:31:49 From Jeff Miller Ward considers a skills inventory as supporting Eric's skills needed to maintain and evolve a system, and suggests keying the skills to a project which exercised those skills. (Jeff considers his father, who is the embedded skills source for SLAC's accelerator controls software)
10:32:51 From Jeff Miller make virtue easy, make sin hard - make it easy to connect with skill learning output-focus versus capability-building focus capability-maintaining focus, knowledge transfer, "truck number" increase ...
10:34:05 From Jeff Miller financialization leans against building capability?
10:34:24 From jan d what is Eldorado here?
10:34:29 From Jeff Miller (takes a while to socialize the skills across the company rather than identify a few experts?)
10:34:32 From Brian I think there is the "meta" inventory...things like "willingingness to learn a new technology, skill, or work with people" much more so than specifics about a given thing.
10:34:40 From Pete Keeping a Diary to help us reflect on ourselves, then how to share appropriate/comfortable pieces with others 'radical candor' can be too much
10:34:53 From Jeff Miller Eldorado is Ward's project at New Relic which connected a model of people and projects across the company radical candor can cause ringing in the social-emotional-hierarchy system that can get amplifed in awkward ways
10:36:18 From Brian Sort of "TDD" monitoring...
10:36:48 From Jeff Miller monitoring as scaffolding for development; remove the superfluous extra monitoring once the system was deployed
10:37:15 From Pete running with LEVEL=DEBUG or LEVEL=TRACE vs. LEVEL=INFO or LEVEL=WARN or ERROR
10:37:21 From Paul Rodwell a wiki about El Dorado - http://ddd.ward.wiki.org/about-the-el-dorado-project.html
10:37:21 From Jeff Miller Eric reflects on Kent Beck's approach to TDD, where scaffolding and architecture tests may not be the ones suitable for regression.
10:37:37 From Brian TDD is great for confirming your understanding of how functions work...for example, how the terms in a regex are grouped.
10:38:08 From Jeff Miller TDD can be a struggle if you're not good at refactoring - that in order to ask a question, you have to define an interface and then once the structure is suggested, to move the code into that structure.
10:38:37 From jan d Reacted to "Eldorado is Ward's p..." with 🙏🏻
10:38:58 From Brian If someone asked me that, especially my boss, I'd be pretty cynical... :/
10:39:02 From Marc Pierson What kind of world would accommodate zero maintenance? Then move back a step from that.
10:39:15 From Jeff Miller "I used this skill on this project" -> information about the project's maintenance needs.
10:39:43 From Brian If those skills relate to the compensation structure, it would be a very worthless list for relevant skills real soon...
10:40:14 From Jeff Miller Ward reflects on a "tennis ladder" approach to anyone under your level could challenge you; also where an overmatched game could pause and become knowledge transfer (an example from a squash game playing ladder).
10:40:18 From Robert Best Leave the ladders down
10:40:23 From Marc Pierson Where does the cost of maintenance cross the benefit of continuing use?
10:40:52 From Jeff Miller a casual currency of contribution giving reputation
10:40:54 From Brian Maybe ask the peers for what skills the individual has. For example, to my peers, I'm a physics + programmer combo that can talk both fields effectively. That is not how I would have described myself.
10:41:08 From Jeff Miller (Ward's thought about compensation via recognition and the ability to do more of that thing) Brian: that's a bright reflection - "that is not how I would have described myself" - because skills emerge in context.
10:42:11 From Brian Because your success becomes an anchor very quickly and will prevent you from new opportunities.
10:42:12 From Jeff Miller what is salient - what stands out - depends on what your'e working on with whom
10:42:16 From Marc Pierson Skills drift ? = new skill or simply shared skills?
10:43:44 From Jeff Miller your successes can stereotype you? (to Brian) (others won't allow you to be bad at new things) (you won't want to be bad at things?)
10:44:31 From Brian For sure. To many of my collegues, my "identity" is that of me from 15 years ago. I've quite different with different skills from being a DBA now...lol.
10:44:36 From Jeff Miller (a discussion of skills transfer, LSI chip design) I've been advised to ask folks I know what they think I'm good at, in order to help me prepare my elevator speech.
10:45:51 From Jeff Miller "good at sense-making and summarizing" is not a highly paid skill except alongside other particular domains
10:47:15 From Jeff Miller Jan reflects - how far should a skill and badge system go? Universities; technology certifications; but let's say you exercise your skills within your team and company? How far should that badge travel as information?
10:48:02 From Brian I probably have more respect from a Google janitor than some of the other positions...lol.
10:48:24 From Jeff Miller Some of the other positions actually stood up for the janitors "they're keeping our workplace in good shape, and they are underpaid and have to travel long distances!"
10:48:45 From Brian I still haven't found a reasonable use for blockchain...
10:49:06 From Jeff Miller Pete reflects on a skill badging system in the employee profile at IBM. Ward suggests building a questionnaire by interviewing people - "what skills have you used in the last week?" - what do we need to know in order to do what we do?
10:50:19 From Jeff Miller Brian reflects on "and what skills do you rely on your teammates for?"
10:50:42 From Pete one-eyed man in the kingdom of the blind
10:50:47 From Paul Rodwell missing skills - like ‘explaining’ and ‘listening’
10:51:11 From Brian "Sarah" knows how to navigate StackOverflow and therefore the path to the solution...
10:51:43 From Jeff Miller Eric has an anecdote about the peers seeing in you a thing you don't recognize in yourself - "why does everyone go to Sara for database stuff?" - Sara knows a relational subquery pattern, let's say.
10:52:07 From Paul Rodwell and that often overlooked being curious
10:52:38 From Jeff Miller (holds up a write ring) "What is this thing that I'm holding?" Ward reflects: "This has to be an assessment of the work done, not your ability - because this is what you worked on this week."
10:54:21 From Jeff Miller "I"m skilled at async/await, because I did that this week" - but maybe there's a new library people might want to use next year? Query your records of actual use of async/await, and there's your evaluation committee. That's not your measure of expertise, but of experience.
10:55:21 From Brian Better than the pseudo promise versions before ... Jeff has an "async" badge now...
10:55:44 From jan d I like https://pragprog.com/titles/es6tips/simplifying-javascript/ (it does not explain DOM, though, just JS)
10:56:08 From Paul Rodwell https://developer.mozilla.org/en-US/docs/Web/API/HTMLImageElement
10:56:42 From jan d Reacted to "https://developer.mo..." with 😲
10:58:26 From Jeff Miller Pete Hodgson - "stop putting things in the FedWiki DOM, build models that you work with, and then you want an async library, here are three possible ones." Ward reflects on his Photo Rotator project - crashing the Chrome browser multiple times - "there's something going on in this image loading that I don't understand" - I needed .decode()
11:01:09
11:01:09 From Jeff Miller actively bad at learning the opportunity landscape (Jeff) pattern recognition
11:02:22 From Paul Rodwell https://developer.mozilla.org/en-US/docs/Web/API/XMLSerializer and serializeToString() to escape markup - no need to use those old regex replace
11:02:46 From Jeff Miller an anecdote about hoarding knowledge and status "retool on the new technology after 2 years" - Ward
11:04:08 From Jeff Miller (about the same cycle for industry skilled people as embedded consultants in government IT modernization) "Most of my clients asked me to come and work on Java, in order to tell their boss to go back to Smalltalk."
11:05:21 From Jeff Miller (Ward's reflection on the evolution of relevant technology) async and await - "it still has to work when an operation isn't fast"
11:06:01 From Brian Effiicient for some things...
11:07:16 From Jeff Miller Erlang's model of parallel computing using mailboxes and communicating processes
11:07:35 From Brian OTP I think makes it the success that it has been.
11:07:52 From Jeff Miller BEAM / Erlang / hot reloading / WhatsApp as an early amazing demonstration of operating at high usage on the network and many users. explicit coordination via message handling
11:08:19 From Brian cooperative is really nice sometimes. Keeping cores busy is about memory management and cache utilization, which is a specialized skill that most of us don't have...
11:09:51 From Jeff Miller Pete reflects on loading up your jobs at a process and pod level by saying "this process needs about 80% of one processor" or "this process needs 20% but can need a lot of headroom") eviction and rescheduling
11:10:18 From Paul Rodwell back to the future with TSOS
11:10:32 From Jeff Miller cache usage observability is a skill most of us don't have
11:11:30 From Brian Replying to "BEAM / Erlang / hot ..." https://www.erlang.org/doc/system_architecture_intro/sys_arch_intro.html
11:11:34 From Jeff Miller (different considerations when hoisting a well-known system into a cloud provider)
11:12:36 From Jeff Miller the cost structure, the knowledge, the dynamic behavior from datacenter to cloud is just new
11:12:39 From Brian When cloud venders were willing to make computation a loss-leader, then it was useful to offload more system management to those systems...but not they are charging reasonable prices...it's often better to do your own sys admin...
11:15:00 From Jeff Miller "we understood how to deal with the dynamics of our databases in an old model -- but we're having to relearn that, even though the underlying hardware leakage issues are still present -- but we don't have the tools to see the detailed problems in a cloud platform" can't apply kernel level optimizations L1, L2, L3 cache localizations, compiling the inside loop to optimize the hardware related characteristics
11:16:17 From Jeff Miller a thing you can do if you're running natively on a particular architecture - but you have to guess and experiment when you can't get to the platform's lower level (and can change out from under you)
11:17:01 From Paul Rodwell https://en.wikipedia.org/wiki/DTrace
11:17:42 From Jeff Miller [Next week I'm talking with Apptio, an IBM subsidiary whose value proposition is "we'll help you understand your cloud running costs mapped onto your business-related computing services"]
11:18:14 From Pete Yup....got lots of that - Turbonomics gathering stats and giving recommendations
11:18:58 From Jeff Miller so maybe I'll be hiring in? early days yet.
11:19:15 From Pete Some people turn it loose on their actual infrastructure and let it twiddle the knobs
11:19:53 From Brian maybe there is a contrast of testing vs "feedback" as a form of testing. Both are really "talking" to the inquisitor to tell information/feedback to either confirm or inform, etc...
11:20:06 From Jeff Miller https://pnsqc.org/ - 42nd Annual Pacific Northwest Software Quality Conference, Oct 14-16 2024
11:20:08 From Pete can save money most of the time, but can cause issues if it overoptimizes (30-day average CPU is x, but once a month you need 10x just for that day etc)
11:20:23 From Jeff Miller "Tell me something you do routinely that others are surprised you can do." (I think of "solutions engineers" as a surprising set of skill kits)
11:20:47 From Brian "Exploratory Data Analysis" https://www.itl.nist.gov/div898/handbook/eda/eda.htm
11:21:45 From Jeff Miller Ward says "I think that the thing which I do more routinely is using GraphViz for visualization" - "here's how to integrate GraphViz by formatting output with a few particular tricks."
11:22:28 From Brian i.e. every line in it's own NPM module...
11:22:42 From Jeff Miller Ben says, "Every system has boundaries where things get simplified as they cross that boundary. You only have to read enough of the code to find the boundaries, and then watch the behavior at the boundaries" - Ward's interview with Ben.
11:22:54 From Brian Micro-services usually have an advantage that way.
11:23:12 From Jeff Miller instrument the boundaries - Ben learned it at the release process preparation group at Apple distributed tracing is pretty interesting for that
11:23:32 From Paul Rodwell maybe Open DTrace - https://github.com/opendtrace
11:24:02 From Brian The Doxygen references and referred by graphs are helpful sometimes.
11:24:31 From Jeff Miller Ward says, for PNSQC - "here's how to annotate the program to understand what process is going on - there's a category of activity, why is this happening and why?" - the annotations create nodes and relationships
11:25:00 From Brian Annotatating sounds like the Illinois model rather than the New Hampshire model...
11:26:09 From Jeff Miller "here's a GitHub action which scans your code for annotations, and builds a picture of the code via what the programmers have intended in the function of the code in various places" - a tool for the quality team for understanding the software and asking informed questions.
11:26:44 From jan d Felienne Hermans (https://www.felienne.com/ ) had/has an code reading club. Would be curious what their preliminary methods were to read and if it involved diagrams etc. (I use a lot of diagrams when code reading, recently hotdraw)
11:26:49 From Jeff Miller "here's a line from here to there, is that a read or a write operation?" - you can revise it separately from the code I am a fan of Felienne Hermans but have been slow to adopt her methods.
11:27:58 From Jeff Miller "Ben how do you read code?" "I don't, really, I look for the interfaces and watch those closely." - Ward's example of a learning context which prompted annotation for processing into a diagram the revival of Javadoc? a model of the code generated as the skeleton of documentation
11:29:04 From Brian Do you get a lot of "assign x the value of 5, for int x = 5" type lines?
11:29:08 From jan d I like Javadoc (not as great as a custom guide doc but still) I'll leave, getting late in Berlin and I need a second dinner, cause I just rushed in at my 18:00 :)
11:29:28 From Brian Bye Jan,
11:29:55 From Jeff Miller Pete considers the generated code model as a useful artifact for examining outside the context (sometimes trade-secret/NDA) of the actual code.
11:31:07 From Jeff Miller prompted remembering of annotation context (Ward reflects on adding annotations to federation search code) - two different diagrams for indexing and querying
11:32:29 From Jeff Miller "Show me where indexing meets querying" - that's the interface that Ben pointed to - "so why's Relocalize Creativity missing from the index?" - and it's that "*local*" domains got skipped, reLOCALize creativity. "Look on the diagram where the thing was lost, it was on this side of the seam, it isn't on that side of the seam, what's this code here?"
11:33:36 From Jeff Miller an annotated model of code which tells you where things happen, so that you know where to look for what happens (reminded of the event storming model of data flow, and things happening in the domain)
11:34:40 From Jeff Miller "Paul figured this out a different way, by running the indexer on localhost, and getting nothing indexed." What GraphViz does well is promote an intuition of the system connections, and it's easy to deal with as a format for graphs. "I have a format for trees, for lists -- what do people use as a format for graphs?"
11:35:44 From Jeff Miller .CSV for tabular data (etc.) JSON for tree-structured data text with carriage returns for lists
11:37:03 From Jeff Miller What is a good format for graph data? A list of nodes, a list of relationships connecting nodes. input to Neo4J, for example what's a good format for a property graph (nodes and connections have properties)
11:38:39 From Jeff Miller Marc reflects that a grand plan generates blindness, "Let's focus on [] for now", blinders to exclude other issues. DeLorean car company - stainless steel body, gull-wing doors.
11:40:17 From Paul Rodwell time machine installed behind on the back seats
11:42:08 From Jeff Miller Marc's translation of Chris's Regenerating Sonora report in terms of a Miro board using Ecological, Institutional, Public variables What are points of leverage on the proposed copper mining project? Who cares about what points of influence?
11:43:45 From Paul Rodwell https://www.savetheboundarywaters.org/risks-human-health-sulfide-ore-copper-mining and https://www.savetheboundarywaters.org/sulfide-ore-mining-and-human-health-minnesota
11:43:54 From Jeff Miller Regenerating Sonora, because it isn't part of the copper mining boosters, was left out of the loop of meetings, plans, etc. "publicly traded company" - by definition an absentee landlord which cannot care directly about the details of its activities on the ground.
11:45:22 From Jeff Miller Marc reflects: if you want to get disparate parties to work together, you need to engage in conversations ahead of the project, about what people care about If you can get a third of the time of planning into talking about the whole which each party cares about
11:46:33 From Pete "A Particular Set of Skillz" - in Liam Neeson voice
11:46:48 From Jeff Miller then you have a chance to use that knowledge productively, to give a chance to deliver more of what each interested party wants, to prevent damage to their interests within the whole particular to them.
11:48:15 From Jeff Miller practical technical ethnography what is an unusual skill how did you learn it what are tips from that? Ward reflects on the decompostion of a set of tips for creating a clickable graph representation.
11:49:38 From Jeff Miller what are the steps which might be common to processes that give a useful view? "what is common across these examples? are there tips which emerge from these?" - and different people might see different things
11:50:35 From Robert Best We also figured SQLite could be used to hold the node and edge lists on tables... Not sure how that would be better or worse than in arrays, and being represented in JSON trees.
11:52:11 From Jeff Miller what are examples of how you want your identity and its authority to be recognized? (PKI)
11:52:12 From Robert Best ^^ for storing and using graph data.
11:53:17 From Jeff Miller Eric reflects on personal data leaks as an example; therefore, it should be more routine to have personal data stored in encrypted form controlled by the individual
11:53:27 From Ward Cunningham http://xpdx.org/view/leveraged-activities
11:53:34 From Jeff Miller and to be used in limited contexts ("the analog hole" for DRM tech) "why can't I screenshot my video player?"
11:54:40 From Jeff Miller Facebook is theoretically an institution localized which can be attacked by a class-action lawsuit
11:55:13 From Paul Rodwell an aside - have you looked at the list of root certificates in the browser? have you heard of many of them? or even trust them?
11:55:13 From Jeff Miller therefore they have some interest in protecting their reputation right, HTTPS as a deferral to "authority"
11:56:00 From Ward Cunningham http://wiki.dbbs.co/graph-a-structure-with-graphviz.html
11:56:47 From Jeff Miller Eric reflects: having fluency, having tools which make it visible when you're transferring trust - that's a thing I want to see. (Scams are actually better if they're detectable by the average person, but undetected by a more credulous person)
11:57:43 From Robert Best I recently saw a headline claiming Google is losing the SEO spam bot battle.
11:57:53 From Jeff Miller the evolution of imperfect scams as more efficient search quality is one of Google's crown jewels and it's getting tarnished
11:59:32 From Jeff Miller Ward reflects and demonstrates about finding the appropriate information and wiki pages related to skill interviews. "tips.xpdx.org" - that site has disappeared - so let me search more generally for "Graph a Structure with Graphviz" - search for similar titles - so here are a handful of pages - Eric Dobbs has that page!
12:00:41 From Jeff Miller https://wiki.dbbs.co/view/graph-a-structure-with-graphviz
12:01:43 From Jeff Miller "Graph a Structure with Graphviz", Things to try, How it Works, Triggers" (triggers from a vocabulary of when habit behavior occurs)
12:02:11 From Pete http://xpdx.org/view/leveraged-activities/view/search-from-graph-a-structure-with-graphviz/gd.fed.wiki/graph-a-structure-with-graphviz/gd.fed.wiki/dump-a-structure-into-graphviz
12:02:16 From Jeff Miller "Dump a Structure into Graphviz" LOCKSS in action! lots of copies keeps stuff safe
12:04:21 From Eric Dobbs http://gd.fed.wiki/welcome-visitors.html
12:04:32 From Jeff Miller oops, "search failed: error" (SEARCH SLUGS ...)
12:04:58 From Ward Cunningham http://wiki.ralfbarkow.ch/dump-a-structure-into-graphviz.html
12:05:51 From Jeff Miller "Guiding Diagrams" http://gd.fed.wiki/ - Ward's notes for a class on diagramming. turning down XP from 11 to 7
12:07:08 From Jeff Miller (evolution of XP users' groups to "Agile" users' groups)
12:09:06 From Jeff Miller http://gd.fed.wiki/view/guiding-emergent-structure/view/lineup-diagram
12:11:28 From Jeff Miller Eric to Ward, asking about code annotations for visualizing structure "a thing is happening" "it has some purpose" "it connects these pieces" ?
12:12:57 From Jeff Miller Ward is interested in getting a co-speaker from someone who lives that problems that the graph annotations are aimed at
12:15:10 From Jeff Miller a link for Jan on "El Dorado" http://found.ward.bay.wiki.org/view/el-dorado/view/about-the-el-dorado-project
12:16:13 From Jeff Miller (Ward describes the result of annotating federation search code) http://ward.dojo.fed.wiki/view/search-index-logs http://search.dojo.fed.wiki/
12:19:32
12:19:32 From Jeff Miller http://search.dojo.fed.wiki/view/welcome-visitors/view/federation-search-explained/view/first-trial-segmentation/view/search-services/view/search-script-annotations/view/search-nodes-and-relations some related pages annotations to diagrams http://search.dojo.fed.wiki/view/drawing-automation distinguish scripts and files by color (blue for scripts, green for i/o files)
12:22:10
12:22:10 From Paul Rodwell getting this talking back to wiki needs some work on message passing, so we don’t have different uses of messaging clash.
12:22:30 From Jeff Miller http://search.dojo.fed.wiki/view/drawing-automation/view/solo-super-collaborator http://search.dojo.fed.wiki/assets/pages/solo-super-collaborator/README.html (the supercollaborator, but empty)
12:23:47 From Jeff Miller JSONL file ("insecure download blocked, click to confirm") a trick from C2 wiki with non-ASCII separators
12:24:58 From Jeff Miller Ward uses a Unicode triangle character as an annotation punctuation, making it easy to identified the annotations in source code
12:25:35 From Pete Tags
12:25:43 From Paul Rodwell Need to leave to grab some food…
12:25:48 From Jeff Miller "how did I choose these Categories or Aspects?" (or Tags - Pete)
12:26:23 From Pete Labels in some sytems (could become label of a box / line)
12:34:01
12:34:01 From Jeff Miller search graph generated from annotated code https://github.com/WardCunningham/search/blob/master/graphs/Index.svg "2 months ago" - or maybe pre-annotation?
12:36:06 From Jeff Miller Ward describes coaxing an extracted diagram by making small corrections (spelling of "Kafka", for example) Eric reflects that production incidents reveal that two teams' systems are connected and that a post-incident activity might be annotating where those connections exist teams are often reluctant to take on work which reduces autonomy
12:37:17 From Jeff Miller "no promises!" annotated code reveals connections at low overhead
12:38:49 From Jeff Miller here's how we represent the seams between the teams GitLab CI that was GitLab's competitive advantage for a while
12:39:52 From Jeff Miller (might be a portfolio project for me? code annotation processing for understanding)
12:40:29 From Pete Dynatrace collects some of this in interesting ways but there is so much 'programming by config file' out there it can't keep up without lots of special modules to parse it all out
12:42:56 From Jeff Miller example of extracted annotations https://github.com/WardCunningham/search/blob/e82ca51bc6e834d6ea04becae16b8774aa6b462a/xref.txt "You know these diagrams are unmaintainable ... unless we have a process around them." (Eric and Ward) with El Dorado as the working example.
12:45:33
12:45:33 From Jeff Miller example of an annotated file related to wiki federation search https://github.com/WardCunningham/search/blob/e82ca51bc6e834d6ea04becae16b8774aa6b462a/cron.sh
12:46:56 From Jeff Miller the despair of trying to maintain working documentation is something I've seen a couple of times code style check? "is there an annotation? did new code have an annotation?" maybe not, sigh.
12:48:06 From Jeff Miller "what if incremental work allows the updated diagram to be self-generating?"
12:49:10 From Jeff Miller tags and folksonomy? Pete reflects on an incident - it's not just the annotated source code, but the configuration of the deployed system; the integration points happen at configuration of collaborating services.
12:50:20 From Jeff Miller Pete considers Dynatrace (see earlier comments) Dynatrace injects a request property
12:51:31 From Jeff Miller (multiple implementations of this in action - I think New Relic might have supported distributed tracing via injecting a GUID per request, showing fan-out of queries, etc.) Moldable Development - does this sort of tracing using a live-coding technique
12:53:06 From Jeff Miller "an on-demand composable diagram from a collection of parts" "with a recommendation engine based on shared keywords / nodes / vocabulary" "get it close to the artifacts which you're deploying" - whether those artifacts are in source code, or in config code, etc.
12:54:08 From Eric Dobbs http://exploreddd.com/
12:54:57 From Pete Wriring down intention / aspects in comments / javadoc etc
12:56:36 From Jeff Miller examples of ways to use the extracted data with different output types and visualizers? (so you're not bound to a specific tool, once you have annotations and annotation parsers to extract and cross-reference them) a navigable inventory "Do you understand the inventory problem you've had for a long time?" The "Backstage" framework was about inventorying microservices.
12:58:18 From Jeff Miller (September 2023, San Francisco, Developer Productivity Engineering Summit; Backstage was a system prevented for microservice inventory) *prevented -> presented for microservice inventory
12:59:38 From Jeff Miller Product and service opportunity discovery * what do customers want? * what might provide that? * how do we investigate that? (stop, do that later) "how do we know when we can turn off a microservice" ? (related to inventorying)
13:01:09 From Jeff Miller the system pre-existing to El Dorado - "Treasure Map" there was too much structure to easily extract it to a spreadsheet and ask "how many microservices do we have?" "How about we try Neo4J to visualize the service graph?"
13:02:31 From Jeff Miller "the microservice infrastructure is complicated, but we can browse through the structure to help understand it" a process of product discovery yielding a surprising outcome that was positive
13:03:38 From Jeff Miller (thinking that there's a parable about scattering seeds and being unclear about when they might sprout) canned queries for the first treasure map graph in Neo4J evolving into El Dorado