This talk is a must-watch for people in the Event-Driven space. I saw @noctovis.bsky.social giving it live at this year's DDDEU, and I truly believe that she nailed the versioning topic ππ
www.youtube.com/watch?v=6TCb...
This talk is a must-watch for people in the Event-Driven space. I saw @noctovis.bsky.social giving it live at this year's DDDEU, and I truly believe that she nailed the versioning topic ππ
www.youtube.com/watch?v=6TCb...
Read more in my latest article event-driven.io/en/strict_or...
And drop me a feedback or follow up questions π
4/4
Thatβs a good question, and good questions usually require more depth to give a precise answer. Thatβs what weβre here for!
I followed up and discussed how to determine whether we have complete information for our events.
I showed how record revisions from optimistic locking can help with that
3/
...For example, you get an ItemRemovedFromCart event, but the item doesnβt exist in your view of the current state of the cart. Is it an invalid event? Or is there an ItemAddedToCart event that hasnβt come through yet?"
π€
2/
After the last article on Race Conditions in EDA, I got such a question from Ben:
"You described the scenario where you know what events you should receive, just not the order. But what if you donβt know that?
...
1/
Thanks, could you tell me what were your thumb rules between choosing one over another?
Thanks for the extended follow-up. How do you typically integrate two disconnected processes (e.g. between different deployments)? Avoiding async-processes can be a good choice, but sometimes we need to have some orchestration.
Out of curiosity, could you expand on why? Lack of tooling or "other reasons"? π
Out of curiosity, could you expand on why? Lack of tooling or "other reasons"? π
Hence deferring strategy to name "things" might be really desirable, as "naming is framing" - then "models" (of thinking) then are constrainted just by the act of "naming".
I tried to express some thoughts on that topic too talesfrom.dev/blog/the-amb...
I read it as "Models are models but that's ok" :-)
The quote "All models are wrong but some are useful" should not be read as an excuse to stick with your model. It's a call to actively search for more useful models.
#Java and #JVM folks, are you using the Outbox Pattern? If yes, then which tool are you using? It's a genuine question; I'd like to see if my anecdotal observations are real π
Homebrewed? Debezium? Spring Modulith, other? π€
Yeah, time to add a new ADR to apply the NIH principle!
Packages' peer dependency issues are something that makes me hate the world sometimes π«
But I hope you're aware that Temporal (and likes) is actually a message-driven tool?π
My personal take on Temporal is different. Imho most people just need outbox, cron and retry policy when they choose Temporal, and IMHO, there are lighter and less vendor-locking tools in the space to achieve that
Old but seems still relevant udidahan.com/2010/08/31/r.... Crazy how quickly everything changes in software :P
Yup, I like this article and getting back to it from time to time π
Could you expand on your definition of the workflow engine?
So, were you happy with the plot twist in the end? π
Recurring questions during my consultancy are:
- How to deal with race conditions?
- I'm getting events out of order from other services.
I wrote an article today about the technique that should take you far enough explaining how read models can help you in that.
event-driven.io/en/dealing_w...
Recurring questions during my consultancy are:
- How to deal with race conditions?
- I'm getting events out of order from other services.
I wrote an article today about the technique that should take you far enough explaining how read models can help you in that.
event-driven.io/en/dealing_w...
PoP OS, so Ubuntu-based.
It happens everywhere, the mouse scroll "jumps" back and forth sometimes. Increasing speed helped a bit, but it's kinda annoying. I saw others also had that case. It's not a blocker, but really annoying papercut.
Well, I had to change it through terminal, finding which number represents my USB dongle, and then crafting shell script. I'm capable to do it, but I'm incapable to understand how Linux community accepts that as good UX π
Most stuff is fine, but after years I still find surprises like that.
Define "all modern Linux" ;)
This is how it looks on my PoP OS (ubuntu-based).
Nevertheless. Linux is fine, I'm using it for a few years as my main OS, but it surprises me every now and then how advice to "just add startup script" is an generally accepted way to do good UX.
For developers it is, for outside world, probably not so much π For reasons.
If you believe that The Year of Linux on Desktop is close, or even here, then google "how to change mouse scroll speed on Linux".
Victoria Beckham meme - We need monorepo - Be honest - I am honest - You want to just update multiple modules instead of working on their boundaries? - How did you know? - Thank you
Inspired by @ducin.dev