Yeah, this slash and burn approach is legitimately good in many circumstances. You cut through cruft, break down old unneeded processes, and things that are needed but were undocumented become quickly documented and maybe refurbished. Things that break are quickly noticed and reinstated.

It really doesn't work for huge complex systems where the breakages aren't felt fully for many years, and building the systems back up take decades, and things break so far down the line that it's impossible to even tell what needs to be fixed.

I have to hope that the intentions are actually good, because the appearances from the outside look a lot like the start of a dictatorship.

When you’re refactoring a system (which is really what he’s claiming to want to do) you don’t just delete the repository and start with main(). You figure out where the edges are, how they fit together, where are the Chesterton Fences and how can we protect them until they’re understood. A/B unit testing.

This takes a non-zero amount of time and it takes careful consideration by subject experts.

I can’t really speak to your hopefulness of good intentions - as a parent that relies on government agencies to help my special needs child with the tools she needs, I see no hope here.

What does success look like? How long should the chaos and pain last before you’re beyond hope? Will there be remuneration for people that are materially hurt by this scorched earth policy? How will you be hurt, personally?

I generally agree with you, but I have worked with many balls of mud that actually can not be understood even with experts and research without an unreasonable amount of time. Things like a program that runs in a database that was bought from a vendor that has since gone out of business, and the program ostensibly formats some rows and puts them into another table, but we don't know and can't know if anything else is actually using it because observability is not a part of this product, and we can't reach anybody who actually was involved in making it or setting it up. In these cases, a scream test is easier, faster, and more effective than "proper" research. And nearly the entire system is built on these inscrutable processes that nobody understands.

Sometimes it involves people, too. I did consulting work at a company that maintained an entire department of dozens of people who did nothing but mechanically opened spreadsheets that were dropped into a share, copied a specific set of rows from them into another spreadsheet, and then copied that spreadsheet into another share. It turned out the entire process was redundant, because what it was built to temporarily fix had already been permanently fixed elsewhere, and they still kept growing this useless department for years. Happily, the majority of these people were reassigned rather than fired.

For a collection agency, breaking these things is temporary pain. For a government, it can be deadly for constituents.

> I can’t really speak to your hopefulness of good intentions - as a parent that relies on government agencies to help my special needs child with the tools she needs, I see no hope here.

I am really sorry to hear that. It's not a good position to be in. I have a trans child myself, and I am also worried about being able to care for them in an environment that is increasingly hostile. I've spent weeks of this past month depressed and panicked over the current "governance". I'm just trying to cope, and sometimes that involves a bit of denial.

> What does success look like?

Don't get me wrong here. No success is possible. If their intentions are good, they're still destroying things that will kill people and take decades to repair. There will likely be no remuneration. I'm just hoping for the outcome that is least bad. I have zero hope of anything good resulting.

Thank you for the thoughtful response- and best of luck to you and your child.