Sometimes the best thing you can do is stop someone from writing code.
As for your other point, that research is similar to the research on working from home. It measures a controlled thing, but certainly nothing I would use to extrapolate that software engineering teams work less efficiently when people can interrupt each other with questions easily.
If you have any studies that you think are good feel free to link them, I'm open to being wrong.
It's unfortunate that companies and individuals, like yourself, can't respect the fact that others _are_ more effective remote and when they're not getting tapped on the shoulder by random colleagues.
As your GP comment pointed out remote vs in-office productivity likely breaks even. Companies should treat their employees like adults and let them work where they feel most productive and not cater to the preferences of upper-management or a desire to return to the status-quo.
The entire team at an in-person org pays a global commuting tax. Employees are not an idealized frictionless surface; if they have to commute to work, that comes out of their productive capacity in some way.
What you are describing here is a team that is not managed well. If I had to guess, that team would exhibit features such as:
- critical information only shared in meetings and not recorded in writing (or recorded in a place the team does not use as habit)
- that same critical information re-discovered in subsequent meetings because it was not written down
- team cadence extremely sensitive to PTO because key knowledge is only accessible by speaking to the relevant person
- team cadence extremely sensitive to staff onboarding/replacement because both of these require hands-on in-person time from existing staff
- etc.
It's worth noting that all of these inefficiencies are experienced as "progress" by the team: "we made a key decision" (that we had decided and not written down a month prior) feels like something positive. "I spent most of today helping X get ramped up" also feels like an accomplishment rather than the more correct "our documentation and process is so opaque that we lost $X of engineering time today when an engineer was required to help another get onboarded."
Essentially, teams that need to be in-person to function typically underutilize the old technology (writing) and because so many teams work that way, most people do not notice how much time they waste.
This is not every team that leans heavily on in-person interaction; obviously there are outliers that are capable of using the technology of the written word while also augmenting with high-bandwidth interactions.
This research is categorically different from the research on working from home in that:
- it measures a facet of the workplace that is commonplace and standard
- it has been studied for several decades, and the results have held up. (IIRC the first such studies were done in the 1950s or 1960s.)
IMHO those arguing that people have changed enough that now they can better recover from interruptions than their predecessors could 20 or 30 years ago have the burden of proof, not the other way around.
> I want to be able to quickly resolve any issues in person
I think this was more true for me when I was earlier in my career and remote tech really sucked. But now I see it as generally negative to have to sync with someone else to resolve an issue they (or I) am working on. In my experience, in-person bias is a symptom of a team that does not communicate well in writing (emails, Slacks, Jira, etc.).
> where someone can tap me on the shoulder when they need help leads to overall higher productivity
There is literally decades of published research that refutes this for knowledge work. IIRC the first studies were done in the 1950s or thereabouts, and I am not aware of any big studies that have come since that concluded differently. (IIRC Microsoft's campus was originally designed to give every engineer a solo office because Gates read _Peopleware.)
I will admit that working in interrupt-driven style does produce the feeling of getting more work done, whether or not that is actually true.