Text extraction. See Typescript Archive
09:21:07
09:21:07 From Paul Rodwell talking around the creation of a zip file to migrate a wiki from one farm to another
09:21:20 From Jeff Miller Brian's wiki migration case
09:22:30 From Jeff Miller having regular or admin access in the starting wiki server; having regular or admin access in the destination wiki server; rewriting asset references to re-home them in the destination wiki.
09:25:00 From Brian I missed most of what marc just said....
09:25:08 From Jeff Miller Brian's reflection of Assets as a key-value store; if Asset lookup is done by a wiki instance's server, maybe they don't need to be absolute URLs? Marc says: when we include photos from Wikimedia in the page, will the Assets collection make copies of those?
09:26:10 From Jeff Miller Brian: my solution didn't do that, but "wget" as a spider approach will pull copies of them into the archive. Paul observes that the new Image plugin makes a (reduced) copy of the image into the Assets store.
09:27:18 From Jeff Miller phototelling with TOO BIG photos squeezed down for wiki browsing
09:27:48 From Brian I think the serve move and the "deep" copy are two separate concepts.
09:27:49 From Jeff Miller deduplication of image copies when moving multiple wikis to a new common server host
09:28:52 From Jeff Miller Marc reflects on when having specific Assets folders versus a shared Assets folder for a given server.
09:29:16 From Brian Largely, disk space is cheap...so keep an eye on it, but don't worry about it until it becomes a problem...
09:29:19 From Jeff Miller (when it makes sense to have multiple Assets folders versus a shared images folder) Paul says: the references within a shared wiki server for images are links to a common copy of the bits on disk. Eric and Paul point out that video and audio can be heavy disk users.
09:30:21 From Brian I think people's personal photos *should* be on assets they own/manage/pay for...maybe an asset limit for the "tries.*" type wikis is enough.
09:31:24 From Brian rsnapshot does a good job of using hard links to not replicate backups https://rsnapshot.org/
09:32:57 From Jeff Miller Robert B asks: when is an image copied, versus just a reference? Paul points out that there are cases when you might have abandoned images in the common Assets folder (for images especially) if a wiki is migrated or retired; you can do a link count to see if they're used. The image plugin is currently the only mechanism that makes use of storage in the common folder; an individual wiki only sees its local image links.
09:35:58
09:35:58 From Brian @Eric Dobbs I really enjoyed some of your essays at https://dobbse.net/ Atleast in line with our egos... ;) Speaking as a former horse person too..
09:38:03 From Robert Best You can @mention on here? Had no idea... though I did see if you stick to the same room for a recurring meeting, you can change the chat to be "continuous" so you can chat before during and after meetings and it sticks around in one place https://support.zoom.us/hc/en-us/articles/9862812136205-Creating-and-using-continuous-meeting-chat
09:38:18 From Jeff Miller likewise, looking at the images folders under .wiki
09:42:07
09:42:07 From Jeff Miller e.g.: .wiki/photos.pixiereport.com/assets/plugins/image
09:43:04 From Ward Cunningham https://en.wikipedia.org/wiki/Circle_of_confusion
09:44:51 From Paul Rodwell https://en.wikipedia.org/wiki/Brocken_spectre and https://en.wikipedia.org/wiki/Glory_(optical_phenomenon)
09:46:53 From Brian Google says Hubble's mirrors are hyperbolic.
09:47:14 From Jeff Miller interesting! https://en.wikipedia.org/wiki/Reflecting_telescope for the generic set of designs
09:47:43 From Brian Replying to "Google says Hubble's..." https://science.nasa.gov/mission/hubble/observatory/design/optics/
09:49:56 From Paul Rodwell https://jwst.nasa.gov/content/forScientists/faqScientists.html#howbig
09:51:44 From Brian We are really interested, but will wait for a further explanation.
09:51:53 From Ward Cunningham This pull request is relevant to our earlier discussion of finding and moving assets. https://github.com/fedwiki/wiki-plugin-assets/pull/21
09:52:37 From Jeff Miller Marc says: the physical world and our governance system are nested; but the way we get things done is a network of connections, not a nested recursion.
09:53:28 From Brian I'd like to see that diagram with this insight annotated.
09:53:33 From Jeff Miller When we create institutions with a productive, constructive relationship with the natural and human world, perhaps those connections will keep them more human. We can deliberately create institutions with those connections.
09:55:08 From Jeff Miller Ward mentions "causally efficacious" as a property of enduring life. If you have a thing which can cause things to happen elsewhere (as per Marc's "organizations exist to act"), then DNA is causally efficacious via translation via proteins and ribosomes to have effects in the cell and the world.
09:55:13 From Brian When there are artifacts, like houses, it's a very direct measureable thing...When the artifact is "happiness" it's a lot harder to quantify and relate, though equally important work.
09:55:26 From Jeff Miller How about software? Is software efficacious? Only if people pay attention to the software. A pattern isn't really a pattern unless you can draw a picture of it. (a picture of the pattern's realization in the world?)
09:56:19 From Brian Causal efficacy is a relation between some of the properties of an event and an effect of that event.
09:56:26 From Jeff Miller thus the photographs in Alexander et al's A Pattern Language. "safety is not a pattern, it is a property of patterns, shown via examples of safe stairways and unsafe stairways"
10:01:21
10:01:21 From Jeff Miller oh multiple templates right, yeah (for Word)
10:03:16 From Brian If you can align the values between the Left and the Right, you will have success...when the values are bi-modal, then it sort of stalemates.
10:04:12 From Jeff Miller Marc reflects: the green sets, biophysical living world, truth, physical limits to possibility; the middle, human relationships, like beauty; resolutions among what's good and fair, the right side, and what's possible, the left side. the right, human governance systems, legitimized by aiming at good.
10:05:45 From Jeff Miller Ward reflects: the underlying theory of a lot of Marc's work is that there are enough good people in the world who want to see far, who want to plant the seeds of a truth that will carry things forward and improve the future. Those people should connect and work together; how do we connect them, how do they work together?
10:07:20 From Jeff Miller Brian reflects: in looking for how to find new directions for myself, my boss suggested that I talk with senior folks. I went to them and asked them what they find valuable. I've learned a lot of different insights about different divisions of the organization, and learned more about myself. a "genius", in the Roman sense, of a common spirit and habit of thought and interest (to Ward's reflection on who shows up for federated wiki, and stays around)
10:09:27 From Jeff Miller a Twitter question which sparked a lot of senior people and experienced people to respond in the same sense of the space of possibility and reality in information technology: https://twitter.com/ccieby30/status/1712801666020392998 including "we were always the adults in the room"
10:16:17
10:16:17 From Jeff Miller Ward, we lost you at "there was a time when corporations trained their own"
10:16:39 From Paul Rodwell https://www.recurse.com/
10:17:22 From Jeff Miller mob -> ensemble hacker school -> recurse center (hacker vs. cracker)
10:18:29 From Jeff Miller XP vs. agile agile as diluted brand
10:18:37 From Brian Agile became a business model rather than a philosophy.
10:18:42 From Jeff Miller "Certified Scrum Master" as eh.
10:19:53 From Brian For me, XP was a collection of best practices that seem to work well together. If they worked well, do them a lot, and adjust as necessary.
10:20:54 From Brian I think it was revolutionary to have an iterative contract vs a fix-cost/time waterfall kind of thing that was often used before, but I was just entering the workforce then, so not really sure.
10:21:21 From Jeff Miller Kent coining "Extreme Programming" as deliberately provocative keeping the mindless consultants out of the game by giving them a brand they couldn't sell as words.
10:21:51 From Brian Also, including the customer/user as part of the software team was a significant aspect X-Games was just exploding
10:22:14 From Jeff Miller That was the anchor of the feedback loop to stablize the direction of progress.
10:22:15 From Paul Rodwell https://books.google.co.uk/books/about/Extreme_Programming_Explained.html?id=G8EL4H4vf7UC
10:22:38 From Brian XP leaned into the "coding while bungee jumping" joke that was going around too.
10:22:59 From Jeff Miller arena rock sellouts :)
10:23:34 From Brian Institutions change slowly and there was a need for SCRUM or something so that there could be a contract with someone to point fingers at...but it's gone well beyond it's use, imo.
10:24:31 From Jeff Miller "lightweight methodologies" as an umbrella term
10:24:41 From Paul Rodwell Kent’s book was 2000, Snowbird 2001
10:24:45 From Jeff Miller vs. the heavier methodologies like RUP
10:25:01 From Brian XP had a user...but the scrummaster was much more formalized in SCRUM than it was in XP. If you ever see SAFe, run away...lol.
10:25:42 From Paul Rodwell https://www.theatlantic.com/technology/archive/2017/12/agile-manifesto-a-history/547715/
10:25:47 From Brian I don't think that SAFe can ever succeed by any definition that you'd find acceptable.
10:25:52 From Jeff Miller "Scrum Master" -- articulating how the processes were refined (Ron Jeffries in C3 project) Ron J. likes Scrum as a form of work, and as a name for what he does.
10:27:02 From Jeff Miller "scientific management" by number-izing the process in ways that are not useful a sprint goal for Scrum is a focus point and "don't interrupt the sprint" as a way of protecting the process
10:28:17 From Jeff Miller common across IBM: agile in any flavor was better than what was practiced before (Ward relates IBM's study of team practices)
10:29:06 From Brian It's values and power/control. Bosses want the control and SCRUM/KPIs/etc prioritize their power over the worker power...
10:29:07 From Jeff Miller corporate interests as a source of corruption of the values of XP
10:30:16 From Jeff Miller My complaint toward management was often "you don't understand opportunity cost!" when process was added for fear's sake, rather than capability to recover as ability to reduce damage.
10:31:47 From Brian https://en.wikipedia.org/wiki/Wardley_map
10:32:01 From Jeff Miller Ward reflects: I studied Wardley at New Relic because I thought it was important, but I realized that I didn't want to be good at navigating that entire terrain; much of the map is about product management in a corporate institutional context.
10:36:14
10:36:14 From Brian @Marc Pierson reminds me a little of http://paniolo.xyz:8085/view/welcome-visitors/view/social-information I wrote about sender-receiver having three channels in their communication, the text/words, the context/setting, and the "subtext" the part not spoken. And how that differs when the sender and receiver are part of the same group or are in different groups.
10:40:21
10:40:21 From Jeff Miller https://www.linkedin.com/in/timtischler/ looks like this person work in understanding how to connect larger modules and subsystems; Eric Dobbs and Tim Tischler were in the same class;
10:41:35 From Jeff Miller the low-pass filter from The Mythical Man-Month (a person common to Eric and Tim's experience)
10:42:52 From Jeff Miller I froze for a moment, blip on my connection
10:45:06 From Jeff Miller Thompson Morrison observed: there's a fork on the trail. Some folks went to California to look for gold; other folks went to the Willamette Valley to find farmland and build; it's also a reflection on the difference between California (SFBay) infotech culture and Portland infotech culture. Tektronix - radar engineering after WW2
10:46:07 From Jeff Miller (vs. Westinghouse, a different branch of radar engineering after WW2, from Jeff's in-laws)
10:47:33 From Brian A casual loop diagram. :)
10:48:04 From Jeff Miller Eric's story: Tim went around the database team's customers and showed the database team's systemic problems to the other teams; the other teams made changes to un-load the DB, and that let the DB team have the ability to maneuver and improve its resilience.
10:49:47 From Jeff Miller Eric's story about Tim: what do the most useful developers do when they're operating on an incident; when the senior engineer highlighted a thing that made a difference, WHAT LED UP TO THAT MOMENT? (Full Story interaction watching from the UI)
10:51:54 From Jeff Miller or "how do you know this query?" and he got the background from the most experienced engineers that made Tim more deeply useful for using New Relic to figure out its own system behavior Eric reflects on Tim's work at Wayfair, doing domain-driven system mapping, event mapping to figure out where the points for leverage and changing were.
10:55:47
10:55:47 From Jeff Miller > I just saw a posting for a leadership role in technology that was about leading “a transformation” at a company that I knew people leading a transformation a decade ago. At this point, haven’t we learned that transformation is a ongoing process that never ends? That’s the nature of large enterprises. What are the advantages to acting as if it’s a discrete event? Or, more importantly, what does it miss? (3 weeks ago)
10:57:17 From Brian Problems are opportunities in disquise.
10:58:26 From Jeff Miller tools which are not very discoverable versus "what is a question I could ask, that would be great if I could understand how to use it"
11:00:22 From Brian KPIs
11:01:55 From Paul Rodwell from tremendous hope, to disappointment and abuse.
11:02:51 From Jeff Miller (discussion about tools like Full Story which allow improvement of user experience, or understanding of high-performing troubleshooters; but may threaten abuse) "metastable failures" If you get into a metastable failure state, it's difficult and tedious to get out of it.
11:04:16 From Jeff Miller a Humpty Dumpty event
11:04:22 From Brian A system experiences a metastable failure when it is in a vulnerable state and a trigger causes a temporary overload that sets off a sustaining effect — a work amplification due to a common-case optimization — that tips the system into a metastable failure state. Reminds me of async state-machines. *EVERY* state needs to have a path back to "normal" to have a stable one.
11:05:43 From Jeff Miller https://en.wikipedia.org/wiki/Catastrophe_theory (unfortunately none of the pictures are ones I remember)
11:07:01 From Paul Rodwell better - https://www.bbc.com/culture/article/20180525-every-story-in-the-world-has-one-of-these-six-basic-plots
11:07:11 From Eric Dobbs https://www.usenix.org/publications/loginonline/metastable-failures-wild And I think this is the paper from 2021 is the origin of the term in tech: https://sigops.org/s/conferences/hotos/2021/papers/hotos21-s11-bronson.pdf
11:09:32 From Jeff Miller 1. Rags to riches – a steady rise from bad to good fortune 2. Riches to rags – a fall from good to bad, a tragedy 3. Icarus – a rise then a fall in fortune 4. Oedipus – a fall, a rise then a fall again 5. Cinderella – rise, fall, rise 6. Man in a hole – fall, rise
11:09:37 From Brian I designed an analog circuit that would have run fine, *if* I could have charged the capacitors up initially...it was a good learning experience.
11:09:43 From Jeff Miller from the BBC paper, that includes diagrams!
11:11:47 From Jeff Miller https://en.wikipedia.org/wiki/Metastability_(electronics)
11:11:56 From Brian CDMA -- ethernet...
11:12:23 From Jeff Miller with randomly jiggered exponential backoff
11:13:31 From Jeff Miller "agile has missed, in the world of DevOps, being an enabling and empowering force in this world" Eric reflects: the ability for tests to help us refactor code depends on the limits of the combinatorics of code that we don't throw away. Jeff: or forced dependency updates.
11:14:39 From Brian Or the additional options. Having 1 way vs 2 vs 10 apis for the same thing is very different complexity impositions on rest of program. There is cost when you aren't backward compatible too. It's a difficult balance.
11:15:28 From Jeff Miller Ward says: the class of software for which tests work well is known, and there's respect now even for senior engineers to be writing tests.
11:16:58 From Brian IGT (vegas slot machines. 1/3 was QA, 1/3 was legal and math, the remaining 1/3 of staff was all the engineering and manufacturing) Because then you don't have Elon Musks...
11:17:52 From Jeff Miller Eric says: a response to a production failure is sometimes done by investing in QA, but the business pressures increase the combinatoric explosion in production, from greater scale and varying paces, and it's not possible to keep up. (set your customers on fire...) [sell the company and take the profits]
11:18:28 From Brian A friend of mine has a consulting firm and it's grown and now has an HR and travel department and he was lamenting about how much overhead there is compared to the days when it was 5 engineers with an admin...
11:20:18 From Jeff Miller Marc asks: is it possible to not undertake the things at that beyond-the-manageable limit of complexity?
11:20:27 From Brian Ironically, the algorithms * containers with OOP initially created exponential growth until generics were added.
11:21:05 From Jeff Miller Eric says: introducing general managers of a division is a way where a large company can be run closer to that part of the business; but it was typical to unwind the decision. Ward observes: the reality on the ground didn't permit the general managers to exercise their authority effectively.
11:21:54 From Brian I dislike the frequent use of the "enshitification" term, but it sure seems like it accurately describes the business models of a lot of places right now.
11:22:09 From Jeff Miller that's somewhat specific to squeezing the customers rentierism as compared to the problems of the essential complexity of large dynamical systems, which can frustrate people trying to manage them for any purpose, not just to squeeze the business model.
11:23:02 From Brian YAGNI is a pretty powerful concept. I try to keep in in mind before I put fingers to a keyboard. One of my soapboxes is that heros is an anti-pattern that we should be paying a lot more attention to. All businesses become banks, e.g. airlines, amazon, etc.
11:25:00 From Brian A number of years ago I discovered that McDonalds was a real-estate company and that marked a change in the way that I viewed corporations.
11:25:09 From Jeff Miller nod
11:26:39 From Jeff Miller yes, a system (org, nation, etc) needing heroes is in a bad state, and if it keeps needing heroes, it's stuck in a bad state
11:28:08 From Jeff Miller "are there jobs which we should not do?" (a different angle on YAGNI)
11:28:09 From Brian A similar rule of thumb from Kent B is that you can never should work a second week of overtime. That is an indicator broken process. Combinatoric on way down too, with strategy.
11:29:00 From Jeff Miller steering away from the combinatoric mess by not committing to the coupling
11:29:22 From Brian Chuck Moore was an extremist in minimalizing complexity.
11:29:25 From Jeff Miller steering away from the combinatoric mess (Eric) by decommissioning services to reduce the operational complexity
11:30:12 From Brian It's specialized, not generic. 3 Act play?
11:30:52 From Jeff Miller yes, Chuck Moore was a visionary in the microscopic sense, to perfectly fit the system to the case
11:31:23 From Brian Spiking! Do it all the time, most of the time actually.
11:31:45 From Jeff Miller Ward talks about an exploratory approach, doing ten percent of the work to have it tell you what the rest of the work should be. (Jeff: I am not good at this approach, as I was learning yesterday trying to bring up an in-memory database API; I ended up with much more printf-/console.log debugging)
11:34:11 From Jeff Miller "here is a flat list of tree-shaped names, what can I do to better understand how I've used a capability that FedWiki has had for a number of years?" a famous-ish game which I actually won: a copy of Mr. Penumbra's 24-Hour Bookstore, by playing chicken with prime numbers.
11:35:48 From Jeff Miller the fifth root of 371298 ? hmm, the fifth root of 248832 is 12.
11:37:16 From Jeff Miller oh there we go the fifth root of 371293 is 13, the successor of 12.
11:37:32 From Brian JS is so *something*...
11:38:53 From Jeff Miller symbols versus strings yes
11:39:07 From Brian Ran across https://en.wikipedia.org/wiki/Ishikawa_diagram again, I think there might be some uses with fedwiki, either for naviation, or other tree/hierarchy structures that want to be visualized.
11:39:19 From Paul Rodwell https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Symbol
11:40:51 From Brian Do it with incorrect comments. Tell it is a binary search or something.
11:41:52 From Jeff Miller http://clhs.lisp.se/Body/f_gensym.htm "gensym" in Common Lisp
11:43:39 From Jeff Miller Lisp: you can do anything with it; but you have to do it all yourself. refactoring as finger exercises
11:44:37 From Eric Dobbs Essential or accidental complexity
11:44:38 From Jeff Miller (Ward uses the metaphor of finger exercises on a piano in order to work with small programs effectively)
11:44:43 From Brian I think the formulas need to be in a reduced form...
11:44:45 From Marc Pierson Battery almost dead. Will go and save a little .
11:44:53 From Brian Later Marc.
11:45:04 From Paul Rodwell vs complexity simply for the sake of it
11:45:04 From Jeff Miller Essence and Accident in [...]
11:45:44 From Brian I think LISPers added macros that weren't portable...
11:46:56 From Jeff Miller Eric reflects on what Lisp hackers learned in a direct way, which later on became points of reference to Eric's understanding; for Eric, Perl was his reference point for a pointer and for the thing it pointed to. A Perl reference as analogous to a pointer.
11:48:40 From Jeff Miller https://en.wikipedia.org/wiki/XOR_linked_list a destructive traversal
11:50:25 From Jeff Miller this reminds me of the drum memory hacks in the day
11:51:27 From Brian I'd guess that in most cases, the savings being beneficial are rather limited and probably shouldn't be using a doubly link list...especially with today's caches.
11:52:39 From Eric Dobbs This example, XOR linked list, delights me as much as the first time I learned the terms “row major” vs “column major” for traversing nested arrays. https://en.wikipedia.org/wiki/Row-_and_column-major_order
11:53:11 From Jeff Miller Paul refers to weak sets (can be GC'd and not referred to again)
11:53:22 From Brian There are several lists like https://graphics.stanford.edu/~seander/bithacks.html Yeah, matrix transpose is surprisingly expensive in HPC.
11:54:55 From Jeff Miller https://en.wikipedia.org/wiki/Skip_list
11:55:24 From Brian Sort of a binary linked list structure? An index of sorts.
11:57:09 From Brian Those are cases where really nailing down the properties needed, can lead to large payoffs with the constraint. Journal the bloomfilter…
11:58:14 From Jeff Miller checkpointing of Bloom filters as a way for New Relic to be able to have a high rate of data collection and be stable to restarts (via the checkpointing)
12:02:32
12:02:32 From Jeff Miller Part of why I think I've gravitated toward tool development is that the complexity is smaller and more testable. I tend to think in terms of finite state machines. oh hang on, I've worked next to the inventor of skip lists and I only knew him for FindBugs! https://en.wikipedia.org/wiki/William_Pugh_(computer_scientist) small world!
12:06:20
12:06:20 From Jeff Miller my old Algorithms professor went into health care statistics https://en.wikipedia.org/wiki/Udi_Manber Ward describes New Relic's use of high insert rate and then ElasticSearch for retrieval
12:09:11
12:09:11 From Brian It was hard to write performant JVM code at that time period...at least, that was my impression.
12:09:46 From Jeff Miller New Relic started by monitoring Ruby on Rails applications and opened up a division to support monitoring Java.
12:10:28 From Brian Intercept the JIT and replace it with traced/instrumented version?
12:11:28 From Jeff Miller aspect-oriented programming glued in using annotations
12:12:09 From Brian TheIncrediblyLongMethodNamesThatRequireIntellisenseIsWhatGetsToMe()
12:12:36 From Paul Rodwell https://www.youtube.com/watch?v=8Ab3ArE8W3s
12:13:15 From Jeff Miller Episode 2 of the podcast named ... (back to the annotations: people think of Spring and maybe maybe JBoss, and have mostly forgotten AOP and AspectJ)
12:13:49 From Paul Rodwell https://www.causalislands.com/podcast
12:14:50 From Brian I went to langnext (https://www.youtube.com/@25msr/search?query=langnext) one of the years and met one of the JVM architects and really got a better appreciation for everything the JVM did and how well it did it.
12:15:41 From Jeff Miller Crista Lopes, The Future of Conferences
12:16:11 From Brian Flying should be primarily vacation and entertainment, not business travel...but don't see that happening anytime soon.
12:16:34 From Jeff Miller having satellite conference centers and linking them to make a telepresence based grand conference, a Strange Loop ending keynote
12:18:12 From Brian There are a lot of communication lessons to have been learned and I hope we do better for the next time...
12:19:33 From Brian As always, make the desired thing easy to do and the less desired thing harder to do, and people will naturally fall into doing the desired thing. Within the tolerance of the plastic printer, probably good enough.
12:20:46 From Brian Hang it on the wall. Mother's day gift....
12:23:51
12:23:51 From Brian I've been to NM, but not Arizona.