Accountability Sinks

It's interesting we always talked about the Holocaust and the Nuremberg trials when talking about accountability, as if similar atrocities aren't currently happening. It's because breaking an accountability sink of people who are long dead doesn't have any impact other than the explanation itself. Breaking an accountability sink of currently living people and currently active wars is much more dangerous.

This is the same reason that computers suck.

Every program you ever run will precisely follow the same set of rules, because it is those rules.

There's a missing piece that no one has really managed to implement on computers: backstory. The reason why a program's rules are written is much more important than the rules themselves, yet we haven't found any way to write the reason why.

The most important feature of backstory is that it's dynamic. The meaning of a story can be completely changed by simply replacing its backstory. Whether it's a computer program or a societal organization, a decided system must be ignorant to its backstory. There is no place in a decided system to implement context. It turns out that this is a core feature of computable systems: they are context-free.

---

I've been working on a way to change this, but it's such an abstract idea, it's been hard to actually find (and choose) where to get started.

The discussion near the end about how leadership taking responsibility can beneficially relieve accountability reminded me of the story of the Naval Tactical Data System (NTDS) [0].

[1]:

> When NTDS was eventually acclaimed not only a success, but also one of the most successful projects in the Navy; it amazed people. Especially because it had stayed within budget and schedule. A number of studies were commissioned to analyze the NTDS project to find why it had been so successful in spite of the odds against it. Sometimes it seems there was as much money spent on studying NTDS than was spent on NTDS development.

[2]:

> ...the Office of the Chief of Naval Operations authorized development of the Naval tactical Data System in April 1956, and assigned the Bureau of Ships as lead developing agency. The Bureau, in turn, assigned Commander Irvin McNally as NTDS project “coordinator” with Cdr. Edward Svendsen as his assistant. Over a period of two years the coordinating office would evolve to one of the Navy’s first true project offices having complete technical, management, and funds control over all life cycle aspects of the Naval Tactical Data System including research and development, production procurement, shipboard installation, lifetime maintenance and system improvement.

[1]:

The Freedom to Fail: McNally and Svendsen had an agreement with their seniors in the Bureau of Ships and in OPNAV that, if they wanted them to do in five years what normally took 14, they would have to forego the time consuming rounds of formal project reviews and just let them keep on working. This was reasonable because the two commanders were the ones who had defined the the new system and they knew better than any senior reviewing official whether they were on the right track or not. It was agreed, when the project officers needed help, they would ask for it, otherwise the seniors would stand clear and settle for informal progress briefings.

The key take-away is that the NTDS was set up as a siloed project office with Commanders McNally and Svendsen having responsibility for the ultimate success of the project, but other than that being completely unaccountable. There were many other things the NTDS project did well, but I believe that fundamental aspect of its organization was the critical necessary condition for its success. Lack of accountability can be bad, in other circumstances it can be useful, but diffusion of responsibility is always the enemy.

How many trillions of dollars are wasted on projects that go overbudget, get delayed and/or ultimately fail, and to what extent could that pernicious trend be remedied if such projects were led from inception to completion by one or two people with responsibility for its ultimate success who shield the project from accountability?

[0]: https://ethw.org/First-Hand:No_Damned_Computer_is_Going_to_T...

[1]: https://ethw.org/First-Hand:Legacy_of_NTDS_-_Chapter_9_of_th...

[2]: https://ethw.org/First-Hand:Building_the_U.S._Navy%27s_First...

Some of the seminal works on accountability as applied to systems and particularly the business world, are the works from Gerald M. Weinberg.

"Are Your Lights On?" - https://www.goodreads.com/book/show/1044831.Are_Your_Lights_...

"The Secrets of Consulting" - https://www.goodreads.com/book/show/566213.The_Secrets_of_Co...

"More Secrets of Consulting" - https://www.goodreads.com/book/show/714345.More_Secrets_of_C...

> Somehow, the airline has constructed a state of affairs where it can speak to you with the anonymous voice of an amorphous corporation, but you have to talk back to it as if it were a person like yourself.

Welcome to scale. Every business that wants to grow faces this, and those that grow exponentially face this way before they could ever have established a company culture of treating people like humans, which only comes with years of face-to-face interactions, sometimes that don't go so well. Customers sometimes disappointed and you have to make it up to them; when they do, they feel valued. But in today's economy means you can endlessly screw customers, and as long as your business/your userbase/the sector keeps appearing to grow before your exit, giving a shit is an active impediment to that sweet sweet millionaire payoff in the end.

i came to search for a word: advocate. void.

> Concentrating the Jews was thus crucial to the success of the genocide

Oh interesting. Maybe I should be moving out to the country and not among other trans women then.

That is some Tetsuo level bullshit.

Did you already engulfed precious North American spirit in the process, thus, killing it by suffocating it in nonsensical debate?

Let us know when that happens so we can wake up Akira.

> The unsettling thing about this conversation is that you progressively realise that the human being you are speaking to is only allowed to follow a set of processes and rules that pass on decisions made at a higher level of the corporate hierarchy. It’s often a frustrating experience; you want to get angry, but you can’t really blame the person you’re talking to. Somehow, the airline has constructed a state of affairs where it can speak to you with the anonymous voice of an amorphous corporation, but you have to talk back to it as if it were a person like yourself.

Welcome to modern day customer support. Phone or email agents have zero agency and their jobs are more often than not outsourced to some ultra low wage country... the only ones with actual authority tend to be C-level executive assistants and social media teams because a bad experience gone viral can actually threaten a massive financial impact.

[dead]

[flagged]

[flagged]

I am deeply suspicious of "blameless" post mortems. I agree that we should work in ways that minimize fear. We should, to some degree, celebrate the learning we glean from our failures.

But I keep seeing "blameless" being construed as lying about why something happened. It's construed in such as way that anyone can hide from their misdeeds. People screw up, and we need to hold them accountable, and THEY need to hold THEMSELVES accountable. Not necessarily with "punishment" (what does that even mean in a professional context) but perhaps atonement and retraining.

Ah, I love when I'm a software engineer sitting for coffee in the morning, and I open up my tech newspaper to read some extremely overly verbose way of explaining to me like I was just born that yelling at floor staff doesn't change anything (this is actually not a product of modern society, you could yell at a soldier fighting against you and that also won't change anything). Had to stop after that second massive quote. Seriously, what? I thought this was going to be about managing the 1000 compliance settings in Azure and how that sucks.

I'd say he loves the sound of his own voice, but everything worthwhile here is in a blockquote. Oh well, even a poor collator has value as such.

The reader can feel a glimpse of the author's ego the moment he explains his skills as a Google Site Reliabiliy Engineer and his glorious work on improving gmail post-mortems right after the section where a hospital team saves various people in a mass casuality situation by empowering nurses to perform formally doctor-only tasks.

Only with a healthy dose of cynicism I can understand where he's going. While the topic of accountability sinks is quite interesting, I'm searching for the author's reflection of their own accountability.

They worked at google, made a boatload of money for the advertising company and himself, and now philosophically lectures others how to detect and/or design accountability sinks.