Fork me on GitHub

☀️ 😎 💻


So… y’all had a heatwave recently I heard? It was 38C in the house here this afternoon/evening 😞


(and we’re promised much the same all through the weekend 😓 )


Yes. So f*ck home office – I'm back at the air-conditioned company 😁


Summers in france without aircon are miserable for me


But can you drink 🍻 in the office? 😅


Friday afternoon 🍻 ? For sure. :rolling_on_the_floor_laughing:


Enjoying these thick stone/brick walls for once. I usually curse against them when any kind of work has to be done.


We ll reach 30c today... In southern Sweden


Good Morning!


Rainy Oslo today, still warm though. Oh, and good morning!


good morning


Having a meeting with some computational linguists this morning who are making technical decisions without really having much… technical insight at all… this has already resulted in a couple of zoom meetings where the most well-intentioned inquiries like “so why did you choose to do it this way rather than [this other way]?” are shot down in a defensive zoom rage (mostly by one guy who has a stick up his ass). Now I’ve read their most recent proposal and it just raises so many more questions. I’m honestly a bit afraid at what their reactions are going to be when I ask them about it. For the record, I am going to inherit their design as I am fully responsible for step 2 of this project.


I have become very disillusioned by academic collaborations like this. Most seem to treat it as a game of getting funding for their own personal projects.


I suggest to keep asking and explain your point. I'm far from academia but the funding game can ruin a lot of good projects I suspect.


Also "Let's not jump into suggesting solution before we understand the problem" is a good guideline. However, I can see that this is hard to execute in a situation like what you described.


If they are convinced their (technical) decisions are fine, the ask them for the design documents that lead to the decision. They are scientists after all, aren't they?


> If they are convinced their (technical) decisions are fine, the ask them for the design documents that lead to the decision. They are scientists after all, aren’t they? That is exactly the approach I’ve been taking and that is what leads to this one guy spitting rage on zoom 😛


If asking “why did you choose…” results in them feeling attacked, maybe you can challenge their ideas with listing the hard results those decisions will cause? “I see that you plan on doing x. If we do this, it will probably lead to y, and this might cause problem z in the end. How should we approach here?” There are some people with a really analytical communication style, which understand this way of talking about solutions better.


yeah, I will try that.


Is there a chance of having the discussion via a written document, having a written down proposal? Eg wiki page or google docs and comments? It could let people to hold their tempers, react less emotionally, and develop their thoughts more.


I think I will write down my comments and have them reviewed by the more friendly members of the group instead, before publishing them.


Came across and some related nubank repos yesterday, and couldn’t stop thinking about @ordnungswidrig


hehe, I was not involved in the naming 🙂


But in the "ordnungsamt" would fine "ordnungswidrig_es" behvaiour.


suddenly microservices makes a bit more sense


Almost as much sense as a well structured monolith 🙂


Someone retweeted the other day, was it @ghadi? that microservices and monorepo doesn’t really make sense? Yes it was:


I think there is multiple dimensions where you can go from fine grained ("microservice") coarse grained ("monolith"): revisions control, release cycles, domain model and so on. I'm pretty sure there are good architectures where one dimension might be fine grained and the other is monolithic.


Or you can consider "monoliticality" as an extreme form of arbitrary coupling.


I think there should be a balance. I've been in situations where multiple teams shared one service, and where one team was running about 100 services. My preference would be 1-10 services for a team. Preferably that team is also the only team that has to know the internals en manage those services.


it's really tricky, imo it depends on a lot of things, from scale of dev teams to domain


True, but having many services as a team gave a lot of operational complexity. Which would have been less of a problem if there was something like an SRE team to make sure the infra was managed. But working with multiple teams on one service is something you wish to avoid in my experience.


Are you people saying that the answer is “it depends”?

🎯 12

that depends 😉


We also have one app which is pretty much a monolith but depending on environment variables it does different things, so we can deploy it for multiple purposes


It nearly is… 😂


Morning, friends! I’m a bit pissed at the local roe deers who have eaten all our strawberry plants and all potato plants as well. Probably no home grown strawberries or potatoes this year for the Strömberg family. Also a bit worried that the roe deers feel quite sick now, because potato plants are toxic, right?


Only the leaves are toxic to deer, it seems, but they do eat potatoes regularly (and sometimes they do eat the toxic leaves).


I’m sure the deer think you were very kind to put out such lovely food for them! 🦌


Yeah, we feed them with tulips and all kinds of flowers. Regarding the potatoes, it’s the leaves they have eaten. Quite a lot of it.

😞 3