4 ms·

I enjoy reading things like this. It formalizes a lot of intuition and provides a precise vocabulary. The issue I have is when your superiors (tech lead) seems

by ttymck 5y ago

I enjoy reading things like this. It formalizes a lot of intuition and provides a precise vocabulary. The issue I have is when your superiors (tech lead) seems to lack this intuition, let alone awareness of these formalized concepts. How do you work through those scenarios? Just build the "bad way" at their discretion? You can't exactly tell them to "just read this article" or recite a bunch of snippets at them. You could, but I don't think anyone would find that compelling. I'd be happy to build "the bad way" if the decision makers bore responsibility for the consequences, but I'm not going to last long on-call for a rube Goldberg system I had no say in, so I tend to move on.

th3iedkid 5y ago
Bad leaders are a good excuse to jump, unless you just want to bid time.Nothing wrong with either, given you know why are you doing what you are.I suppose the org doesn’t take feedback seriously either, in which case, it only becomes simpler
ris 5y ago
Pragma over "formalism". Every time. I leave the discussion when an architecture astronaut starts talking about Domain Driven Design, Data Mesh Architecture or any other pointy headed nonsense. I will listen if they can convince me a particular approach is better using real, concrete arguments. This sort of cargo cultism is how we ended up with incoherent "microservice" architectures polluting the industry.
- ttymck 5y ago
  Fair enough. I recognize the formalism approach is tantamount to cargo culting. I guess what you are saying is: broach the ideas in the context of the problem. "The foobar intermediary service will help us transition safer and faster than rewriting at once and doing a big switch". Instead of "we should use transitional architecture to avoid a Big Bang". If you follow me. But I have no real evidence that "safer and faster" is empirically true. I can appeal to authority in Fowler's website, but that's what we are trying to avoid. If my lead can't see the intuition in the idea, or refuses to seek out case studies, etc, you can't do much right?
  - ris 5y ago
    This might be overly glib, but I suspect that anyone who has time to write up enough information about a project to be useful in a "case study" isn't close enough to the code to know what's really going on. Not to even mention the political aspect of architects documenting how well their ideas actually worked out.

by ttymck 5y ago

th3iedkid 5y ago

Bad leaders are a good excuse to jump, unless you just want to bid time.Nothing wrong with either, given you know why are you doing what you are.I suppose the org doesn’t take feedback seriously either, in which case, it only becomes simpler

ris 5y ago

Pragma over "formalism". Every time. I leave the discussion when an architecture astronaut starts talking about Domain Driven Design, Data Mesh Architecture or any other pointy headed nonsense. I will listen if they can convince me a particular approach is better using real, concrete arguments. This sort of cargo cultism is how we ended up with incoherent "microservice" architectures polluting the industry.

ttymck 5y ago
Fair enough. I recognize the formalism approach is tantamount to cargo culting. I guess what you are saying is: broach the ideas in the context of the problem. "The foobar intermediary service will help us transition safer and faster than rewriting at once and doing a big switch". Instead of "we should use transitional architecture to avoid a Big Bang". If you follow me. But I have no real evidence that "safer and faster" is empirically true. I can appeal to authority in Fowler's website, but that's what we are trying to avoid. If my lead can't see the intuition in the idea, or refuses to seek out case studies, etc, you can't do much right?
- ris 5y ago
  This might be overly glib, but I suspect that anyone who has time to write up enough information about a project to be useful in a "case study" isn't close enough to the code to know what's really going on. Not to even mention the political aspect of architects documenting how well their ideas actually worked out.