Thread by @rdonoghue

So, after hearing many things in man good places, especially from @ptevis, I finally finished up Skelton & Pais' _Team Topologies_, and it was pretty good.

It has a *little* bit of "This is a really fantastic core couple ideas that got stretched a but to make a book" but at this point I just take that as a given for any business book. (Good news is most start as articles anyway, and you can track them down).

There's a lot of thoughtful exploration of Conway's Law in it, and part of why I like is how strongly aligned with my own philosophy that the most essential part of any system is the handoffs. Gave me some nice language on team architecture too, so that's a win.

One surprise takeaway which made sense in retrospect was the argument *against* collaboration. That sounds weird, but as presented, is a very reasonable case that collaboration is incredibly valuable, but also needs to be *deliberate*.

Which is to say, if you keep two teams who would be better served focusing on their work spending time on collaboration purely because "collaboration is good", you end up wasting a lot of cycles.

Book goes into a lot of thoughtful stuff about *when* collaboration is helpful, and one advantage of framing things in terms of Conway's Law is that you can also flip it and note that if the two *systems* only connect at an API, then that is also the space to limit interaction

(Conway's Law says that designs will mirror the communication patterns of the teams designing, so if you have 4 teams, you'll probably have 4 subsystems. More colloquially, it asserts that HR are our System Archtects)

Anyway, part of why this popped to mind is that I ended up in a discussion of how to re-organize some teams at work last week. Insights from the book were useful, but the curious thing is that this "Maybe don't always be collaborating" was actually a pretty hard sell.

The received wisdom that "Silos are bad" is *really* strong. And it's true! Silos ARE a problem. But the idea that any constraints on interaction between teams will inevitably lead to siloing is...fraught.

Especially when the *other* concerns include "Too Many Meetings" and "Not enough clarity".

If all communication is broadcast, and everyone has to be at all the tables all the time, I'm not sure there's any other outcome than too many meetings.

To come back to @cwodtke's refrain about OKRs - if everything is important, nothing is important. Applies to planning, yes, and it also applies to communication.

Most fields I have sought mastery in have eventually rounded a corner to "do less, more deliberately".

And in every such field, rounding that corner has proven incredibly frustrating to communicate, because it's the hardest thing to argue for.

Because after you've rounded the corner, it *feels* like it's all so simple that people should just be able to jump to it!

But, in practice, they need the time doing more, curiously in order to build a foundation, and you can't skip that.

And, practically, they're going to respond to the suggestion to do less as if it's nonsense. We have such a strong instinct/ethos that the correct response to a problem is to just work harder and do more that any suggestion to the contrary sounds like "shirking".

(Eventually squaring the circle of how the hell to communicate around this sort of gap is usually the NEXT corner to round)