LWN: Comments on "Fighting human trafficking with self-contained applications"

Cryptocurrency and Human Trafficking

linuxrocks123 — Fri, 19 Sep 2025 09:45:44 +0000

Anyone interested in the intersection of technology and human trafficking may want to read "Number Go Up: Inside Crypto's Wild Rise and Staggering Fall". It was an interesting -- and at times very depressing -- read.

https://en.wikipedia.org/wiki/Number_Go_Up

Groupthink

marcH — Thu, 18 Sep 2025 16:38:18 +0000

> The mythical “invisible hand of the market” favours whatever people think will make more money.
> Turns out letting users pick what they prefer is very low on the list of things people think will make money

I agree but just to be clear, the (naive?) "invisible hand" theory does not. That theory says customer pressure is stronger and forces producers _not_ to do all the things you listed. In reality, it's as you wrote just one pressure among others that works sometimes and not that much. Yet it's revered like a religion by some people[*], probably because we love simple answers to complex problems.

FWIW, international trade stopped being entirely based on that theory.

Groupthink

Wol — Wed, 17 Sep 2025 18:31:19 +0000

Whoops - I'm describing predictive cruise control.

But unfortunately, my car provides no way to have one without the other :-(

Cheers,
Wol

Groupthink

pizza — Wed, 17 Sep 2025 14:54:12 +0000

> How on earth is he supposed to do that if his car accelerates unexpectedly without any driver input whatsoever? And he has no way to over-ride other than to turn the whole damn thing off? Or equally the car brakes unexpectedly on a motorway without any driver input whatsoever?

You reveal a fundamental misunderstanding of what "adaptive cruise control" actually refers to:

https://mycardoeswhat.org/deeper-learning/adaptive-cruise...

> Classic feature creep where nobody actually does a risk assessment on each individual creep until we have a major accident ...

On the contrary, the regulatory bodies have done _exhaustive_ assessments (including multiple rounds of public comments) to establish the operational requirements for these systems, and after being widely deployed for the better part of a decade data unequivocably shows them to be objectively safer (ie lower accident rate per distance driven).

...I'll leave it at that.

End here, please

jzb — Wed, 17 Sep 2025 14:44:54 +0000

And... once again... we've wandered far, far off-topic for the article and LWN. Let's end the thread here.

Groupthink

Wol — Wed, 17 Sep 2025 14:39:49 +0000

> adaptive cruise control,

I damn well hope not. The driver is legally obliged to be in full control of the vehicle at all times. How on earth is he supposed to do that if his car accelerates unexpectedly without any driver input whatsoever? And he has no way to over-ride other than to turn the whole damn thing off?

Or equally the car brakes unexpectedly on a motorway without any driver input whatsoever?

(Both usually caused because the car has an internal database and also reads external signs. Throw roadworks into the mix and I've had the car run the pretty much the entire gamut of every legal value between 20mph and 70mph in the distance of maybe 100 yards ... talk about getting confused. Oh, and this is in an area where the City Speed Limit is 50mph.)

My car does both. Fortunately, it's usually on familiar roads and I'm expecting it, but adaptive cruise control is a complete liability. My wife hated the basic cruise control on our previous car. I sure as hell hope she never accidentally activates it on my new car because she'll just lose control ...

Classic feature creep where nobody actually does a risk assessment on each individual creep until we have a major accident ...

Cheers,
Wol

Groupthink

pizza — Wed, 17 Sep 2025 13:55:30 +0000

> Actually cars are another product where corporate greed leads to over weight and over featured products

A considerable amount of that "overweight and overfeatured" is driven by government regulation.

For example, any passenger vehicle sold today has to have a pretty long list of safety [1] and emissions/efficiency [2] features that considerably drive up the minimum production cost of any vehicle. (And that's before manufacturers try to game the metrics..)

Additionally, thanks to the joys of competition folks expect an ever-greater degree of baseline sophistication and functionality, and there is an ever-wider gap between what-they-say-they-want vs what-they-actually-purchase.

[1] Half a dozen or so airbags, lane departure detection, adaptive cruise control, forward collision warnings and automatic braking, stability control, camera systems (which effectively mandates having a infotainment screen), etc. Granted some of these aren't required _now_ but will have to be present within the next few years.
[2] More stringent requirements require more advanced (and expensive, and less reliable) designs.

Groupthink

nim-nim — Wed, 17 Sep 2025 13:34:10 +0000

Actually cars are another product where corporate greed leads to over weight and over featured products and a slowly shrinking market. Seems earning a little more short term at the cost of major problems long term is not a software exclusive.

Groupthink

alfille — Tue, 16 Sep 2025 17:12:46 +0000

First, the goal of the program is wonderful. It sounds like a very constructive response to a terrible personal experience.

But the discussion about UI complexity is interesting. I think the push complexity is more fundamental than corporate greed. It's human nature -- easy of use and learning vs customization, options and, frankly, personal greed. We think we want it all.

Not just commercial software, open source tools are often extremely complex too. Look at open source text editors or the command line switches for wget, sed or awk. People use only subset of features, but not always the same subset. And if you are going to invest the time to use a tool, why not the most powerful one. For the developers, adding features is a balance -- enjoying the challenge, pleasing your users, vs simplicity and maintainability. Also it's hard as an expert to see how a new user sees the interface.

Actually the same is true for physical devices -- washing machines, cars, etc. It's hard to invest in something that doesn't have more features, just in case.

Mass-market HCI is hard, actually

notriddle — Tue, 16 Sep 2025 15:27:48 +0000

I think this is the biggest reason why buggy and slow software is so common.

The HCI/UX problems, however, seem like a more fundamental tension, not a market failure. Consider the description that the article gave the interface:

> it's just a single window with some text and four buttons.

I've worked on software in a similar situation, where I got feedback from non-technical users and they really liked it because of how self-explanatory it was and how frustrated they were with all the competitors and their complex, clunky interfaces. At least, that's how it was for our first customer.

Then we tried to onboard customer #2, and they liked what they saw, but absolutely needed a feature we didn't already have, and, more importantly, they needed to enforce that their employees were actually using it, so we needed to add a check. Customer #1 could not accept this, so the whole feature needed to be configurable.

Ten customers later, and you wind up in one of two places:

1. Hell of combinatorics: your program is 50% conditional on its settings, and your test suite is either woefully incomplete or enormous. LWN has an article on configuring Linux kernel builds that illustrates this problem.

2. An aggressively limited scope that leaves the product as a compromise that leaves everyone equally annoyed.

Thank-you for covering important topics like this

karath — Tue, 16 Sep 2025 13:53:41 +0000

Among all the high-profile and emotive topics (which are important), it's great to see coverage of the smaller tools that really might change some lives for the better.

Groupthink

nim-nim — Tue, 16 Sep 2025 07:58:48 +0000

The mythical “invisible hand of the market” favours whatever people think will make more money.

Turns out letting users pick what they prefer is very low on the list of things people think will make money (perception counts more than measurements, no human actor is perfectly objective, culture weights more than hard facts). Cornering markets, stuffing your product with useless nice to have features, shortening its life, adding money-earning trackware, lobbying (corrupting) market institutions (ie cheating) are all perceived to have a better effort to profits ratio.

Groupthink

marcH — Mon, 15 Sep 2025 21:42:34 +0000

> where the application puts the explanation of what it does right next to the individual buttons that do those things. Several users have told her that they wished other software was written like this, to her bafflement.

This is so typical. In theory, the mythical "invisible hand of the market" provides choice and let users pick what they prefer. In reality, oligopolies and vendor lockdowns leave no room for alternatives or experimentations and they only come from unexpected people and places.

Not just with computers.

prior art

shironeko — Mon, 15 Sep 2025 20:41:27 +0000

This is so cool. I wonder how much free software is used in forensics, seem it would have some overlap with archivists. Another thing I wonder with those paid solutions is how much of the price is liability transfer (real or perceived) or regulatory capture where some software/format is designated as the standard.