yvonnezlam

yvonnezlam, 5 days ago to random

Greetings new followers! I believe many of you got here because @kevlin (thank you!) mentioned a tweet of mine in a talk. My stance on many of the questions/discussions taking place in my mentions might be addressed by the original thread, so here it is:

Or even "we don't have guests, and the guest room bathroom is mysteriously out of toilet paper. Can whoever is using that bathroom please make sure it's restocked?" The thing with tech debt is that in order to have a useful discussion, you need to be able to talk about
@yvonnezlam - As anyone who has ever argued about housework knows, housework arguments are disturbingly specific. It's always "Who minds the most?" and "l don't use that," and "Who's going to do the work?" and (rarely) "Who's going to manage getting this done?" @yvonnezlam - Bringing some of that uncomfortable specificity to discussions of tech debt could be really useful. We need to be able to talk about who minds and why and where the work should live. @yvonnezlam - We also need to talk about how what we do affects other teams/people. Team A can decide not to worry about concurrency right now, but they might be sticking Team B with that work down the road. @yvonnezlam - Right now, with the financialization metaphor, no one cares, because we know "down the road" might never happen. But: that means we don't think "Hm. We're not going to develop in-house expertise on concurrency in a hurry. Maybe we should hire/send people to training/support ... @yvonnezlam - ... people who are interested in this kind of thing doing some study?" or any number of other things to spread the load when the time comes. We don't think about what work Team B might need to put down in order to pick up the problem that everyone knows has been brewing forever.
@yvonnezlam - We don't think "Team A took on this debt because they didn't know how to do the right thing and they had to do something. Not their fault, but...that wasn't good." @yvonnezlam - Whereas if we thought of it as more like housework, we could think "ok, the kids made this mess, now someone has to clean it up, can we get some of the kids involved since they have context for what they did, and it'd help everyone to get context for what to do next time?" @yvonnezlam -The debt conversation assumes a kind of statelessness, because that is the magic of financialization and financial metaphors. Anyone who has ever dealt with tech debt in an org knows that it's not stateless. @yvonnezlam - Mar 29, 2021 Housework is stateful. You don't get a clean slate. You often end up moving stuff from Place A to Place B so that you can clean and tidy Place A properly. And people have state (feelings) about it. O boy do they have feelings about it. @yvonnezlam - Mar 29, 2021 We picked a metaphor for "tech work that is about cleaning up old stuff instead of making new stuff" that sounds like something that would make sense to The Business, but (a) it's not clear it does, and (b) it confuses us thereby falsely constraining our options. /fin

hazelweakly, 6 days ago to random

Tell me you're doing the GitOps thing without telling me you're doing the GitOps thing

diana, 8 days ago to random

@yvonnezlam Made famous in the NewCrafts closing keynote … was a perfect moment set up by other talks here.

grimalkina, 13 days ago to random

"When people are treated unfairly, for example, when they are not allowed to have input into decisions that will affect them, or when they are not given good explanations of why certain decisions were made, the symbolic message may be that the organization does not think highly enough of them (to provide input or to be given good explanations)."

• Brockner & Sherman, 2019
  https://www.sciencedirect.com/science/article/abs/pii/S0191308520300058

grimalkina, 21 days ago to random

Have you ever thought about the process by which people answer self-report measures about their own experience? Because there's a really interesting literature on this. Provided that people HAVE answers to the questions being asked, and feel comfortable accurately reporting them (not a small concern obviously), it is an incredibly depthful and layered process.

In fact, better to think of self-report measures in psychology as being CONSTRUCTIVE dialogue with participants, not passive sampling

grimalkina, 28 days ago to random

Listen it's been a really hard week (four months), if I share a new working paper are you folks gonna be nice to me

gvwilson, 1 month ago to random

When I tried to teach Docker to data scientists with biology backgrounds I had to back up every few minutes to explain underlying concepts. What does it mean to mount a filesystem? What does it mean to background a process? What's a port, what'a s symlink, why should I ever need to "sudo", and what actually happens when I install a Python package? 1/n

gvwilson, 1 month ago to random

"An Empirical Comparison of Packaging Difficulty Across Major Language Ecosystems" is a paper I would read. What does it take to build and distribute something that e.g. fetches a handful of web pages, counts words, and creates a plot in Java, Python, R, C#, Lua, Rust, Go, etc.? And how long does it take a newcomer familiar with the language who's never built a distributable package before to get it working? I realize it's not as hot right now as AI, but if what you want is readers… #nwit

gvwilson, 1 month ago to random

Two papers at #ICSE2024 have "adoption" in the title, two have "transfer", and six have "collaboration". None of them are about getting practitioners to use what researchers have discovered or about getting researchers to study problems that practitioners actually care about.

gvwilson, 1 month ago (edited 1 month ago) to random

Starting to wonder if AI is like spreadsheets: for every programmer pointing out flaws and deficiencies, a double dozen people are using 'em to do something they find useful. 1/4

Added: please see https://mastodon.social/@gvwilson/112265751571981599 for clarification.

grimalkina, 1 month ago to random

You know it's bad when you have to pull up your self regulation playlist to look at reviewer comments

yvonnezlam, 1 month ago to random

Re the libxz hack, I wonder what would have happened if the person who experienced the symptom had not been someone who could also debug the problem in source.

yvonnezlam, 1 month ago to random

Somehow this reminds me that I've never once taken the State of the Devops survey, because I have never felt that it is addressed to me:

https://mastodon.social/@anthrocypher@hachyderm.io/112162777960394133

yvonnezlam, 2 months ago to random

This part of Marco's thread rings very true to me. Previous job was a low trust, high complexity, highly transactional environment, and it squashed me flat. As someone who works on cross-functional, platform/devtools/SRE things, I now dread hearing the words "spotify model" because I associate it with not owning my time and not being supported.

https://mastodon.social/@polotek@social.polotek.net/112107016033538099

yvonnezlam, 2 months ago to random

I would love to participate in a zine-based tech conference along these lines: https://diymethods.net/

danderson, 2 months ago to random

Tepid take: I wish binding sockets to specific interfaces or IPs was harder and took more ceremony, including having to type the equivalent of "I know this is a bad idea generally but I am in one of the very few cases where it's necessary".

Restricting address binding costs users so much pain and fucking around, and causes no end of support demands on maintainers, both for "please let me make this mistake" and for "I have made this mistake and broke it, fix it".

yvonnezlam, 2 months ago to random

I cried reading this remembrance of Aaron Bushnell. I hope his cats are being loved and cared for.

https://crimethinc.com/2024/02/29/memories-of-aaron-bushnell-as-recounted-by-his-friends

hazelweakly, 3 months ago to random

Infrastructure as code is great and all, but why is it that we've completely failed as an industry when it comes to best practices as code?

eaton, 3 months ago to random

there is no good storage structure for sufficiently complex content, everything is a compromise and you just have to invest in understanding the choices you’re making

yvonnezlam, 3 months ago to random

This, in conjunction with a conversation I was having with a former coworker elsewhere, crystallizes something for me: at least in ops/SRE-land, where the work is highly interrupt-driven, we often don't do a good job of talking about how easy or hard a tool is to pick up or put down.

https://mastodon.social/@grimalkina/111853781539013123