Error Categories/Types

[Elliot Temple]

Writing about error correction. Trying to organize errors into categories/types. Suggestions? Or just give a bunch of error examples.

Elliot Temple
www.elliottemple.com

[Alisa Zinov'yevna Rosenbaum]

On Mon, Jan 13, 2020 at 04:09:15PM -0800, Elliot Temple wrote:

> Writing about error correction. Trying to organize errors into categories/types. Suggestions? Or just give a bunch of error examples.

In the Necessary & Sufficient video series ( https://www.toc.tv/TV/video.php?open=excerpt&id=62 ), Eliyahu Goldratt talked about categorizing mistakes made in *execution* (as opposed to planning). Here's my partial transcript of the online excerpt from toc.tv:

> If we bear in mind that what we want is to induce people to continuously improve, what we have to ask ourselves is: what should we concentrate on measuring when we are talking about execution? Should we concentrate on measuring what is done properly? Or should we concentrate on measuring what is *not* done properly?

> [... he decides to focus on what's *not* done properly ...]

> I have to find a way to classify, to categorize the things that are not done properly in such a way that I will not get more than 5 categories. The categories that I'm getting, less than 5, should cover everything that is not done properly and should have zero overlap. That's what I'm looking for.
>
> Now, this is not a triviality. And if you think more and more and more, you'll find out that you really have to dive deep enough to the generic way to get the answer. And then of course, once you see the answer, it's so obvious.
>
> So, let's give you the answer. What I want to emphasize: it took me years to reach it, simply because it's so simple.
>
> There are only two categories. The first one is:
>
> - Things that should have been done and were not.
>
> The other category is:
>
> - Things that should not have been done but nevertheless were.
>
> How do you like that? Only two categories, covering everything. Zero overlap. Exactly what we want.

[Andy Dufresne]

On Jan 13, 2020, at 5:09 PM, Elliot Temple <cu...@curi.us> wrote:

> Writing about error correction. Trying to organize errors into
> categories/types. Suggestions? Or just give a bunch of error examples.

In one sense I think all errors are process errors: the process you were
following didn’t prevent the error.

And then there’s the 2-way categorization of errors as errors of
commission (things you did wrong) and errors of omission (things you
didn’t do but should have).

There’s also the 2-way categorization of errors as process knowledge
errors (didn’t know of a process to avoid the error) vs. process
execution errors (you knew a process to avoid the error but failed to
execute it).

Beyond that I think we start to get into categorizations that are more
useful in some domains or purposes than others.

Like on FI posts some of the main error categories I notice are:
Typos
Reading Comprehension
Quoting & attribution
Grammar
Logic errors
Justificationism
Induction
Evasion
Hostility

Likewise, there are tons of lists of common programming error categories
like:
https://techtiptrick.com/10-most-common-programming-and-coding-mistakes/

I don’t know of a categorization that would be more detailed than the
2-way examples and broadly applicable to ex: both coding and writing FI
posts.

--Andy

[Anne B]

On Jan 14, 2020, at 8:52 AM, Andy Dufresne <foxhunt1...@gmail.com> wrote:

> On Jan 13, 2020, at 5:09 PM, Elliot Temple <cu...@curi.us> wrote:
>
>> Writing about error correction. Trying to organize errors into categories/types. Suggestions? Or just give a bunch of error examples.
>
> In one sense I think all errors are process errors: the process you were following didn’t prevent the error.
>
> And then there’s the 2-way categorization of errors as errors of commission (things you did wrong) and errors of omission (things you didn’t do but should have).
>
> There’s also the 2-way categorization of errors as process knowledge errors (didn’t know of a process to avoid the error) vs. process execution errors (you knew a process to avoid the error but failed to execute it).
>
> Beyond that I think we start to get into categorizations that are more useful in some domains or purposes than others.
>
> Like on FI posts some of the main error categories I notice are:
> Typos
> Reading Comprehension
> Quoting & attribution
> Grammar
> Logic errors
> Justificationism
> Induction
> Evasion
> Hostility

Overreaching

[GISTE]

On Jan 13, 2020, at 6:09 PM, Elliot Temple <cu...@curi.us> wrote:

> Writing about error correction. Trying to organize errors into
> categories/types. Suggestions? Or just give a bunch of error examples.

(Not meant to be complete, just initial thoughts)

- Social-reality-focussed: Judging in terms of the opinions of other
people and their social status instead of judging things in terms
facts/logic/reality.

- Overreaching.

- Hostility to facts, logic, reality. Evasion. Dishonesty.

- Superstition/faith

- Reading/writing/math/logic errors.

- Everything else.


Some of these contribute to other ones. So it would be better expressed
as a mindmap.

——

this is what I read right before thinking about putting
social-reality-focussed at the top of this list.

https://curi.us/2277-static-memes-are-about-social-dynamics#15188

> When I talk about problems like overreaching or lack of paths forward,
> those don't work with most people because they are reality/facts/etc
> oriented. They seem fundamental to me from my perspective, but they
> aren't designed to have the right social meanings to work for
> socially-focused people. Overreaching is not the fundamental problem
> of an overreacher. Living in social reality instead of actual reality
> is their fundamental problem.

——

This was my first draft. I didn’t send it because I hadn’t finished
it.

brainstorming…

There are systemic errors and non-systemic errors. Systemic errors make
it harder to fix non-systemic errors.

Overreaching is a systemic error.

Discussing in bad faith is a systemic error.

Hostility is a systemic error.

[not done]

-- GISTE