Maybe the only example I’ve seen of technology failing to keep up with proposed changes to the law/policy, usually it’s the other way around:
> New York, she added, also had a distinctly punitive approach when calculating benefits for part-time workers that withholds 25 percent of what they receive in a week for each day they work, no matter how many hours they put in. That means that if an unemployed actress spent just an hour or two a day three days a week delivering groceries for a service like Instacart, she would forfeit 75 percent of her weekly check.
> Legislation to change the rule appeared to have the support of Mr. Cuomo and legislative leaders last year, she said, but no bill was ever signed into law. Mr. Blum said he was told why: The Labor Department’s primitive computers could not be reprogrammed quickly enough to make the adjustment.
So they failed to modernize the technology and then used the outdated system as an excuse not to modernize the law.
Generally, I think it's very much both. People's expectations for what's possible is based on experience. And that experience is driven by by previous expectations.
I've helped some places shift to a much faster delivery cadence. That's really easy to do if you have a clean slate. E.g., when I co-founded something in 2010, we started with a weekly delivery cadence and the turned up up, so that a team of 6 was shipping 2-3 times per day. Our infrastructure and our code was very much tuned for that. It took some work to get to that level, especially in terms of automating absolutely everything, but once we'd done the work it was fine. In fact, it was great!
In contrast, I've tried to get existing companies just up to a weekly cadence. That's an incredible slog because expectations, tools, and processes are all built around much slower cadences. It's not like anybody was bad. It was just that if you were only doing something quarterly or annually, it made no sense to automate it. Or even understand it, really. There was a lot of the business version of, "Honey, where did we put the Christmas lights? Did we replace the wonky tree stand, or did we just talk about it?" It's a nightmare. And the whole time most people are like, "Why are we bothering to go faster? The old approach worked!" And for them, it did work.
For a lot of government stuff, I think the change cadence co-evolves. Law is pretty stable, so people just build with low change frequency in mind. A decade later a change comes along and it's really hard, so the higher-ups learn to minimize change. That makes things even worse; requests are now met with high budgets and long lead times, meaning legislators learn not to propose small improvements. That legal stability biases the tech staff and processes even more to not touching anything.
And of course, so much of this stuff was built long before we figured out how to maintain quality in the face of frequent change. Two decades in, there are a ton of new companies that still can't do CI/CD and automated testing well. It's no shock code from the 1970s hasn't caught up.
> So they failed to modernize the technology and then used the outdated system as an excuse not to modernize the law.
Surely no-one really believes that political will was thwarted here by failed technology? It seems far more likely that someone was looking for an excuse and found one. (At the very least, the fact that the Labor Department still has this primitive technology speaks to the general motivation not to invest money here, if not to any specific sneakiness around this case.)
>the fact that the Labor Department still has this primitive technology //
Sometimes law enshrines the ability for access using particular technologies. That means until the legislation is updated (and after!) government departments might need to support "primitive technology".
Government is often about supporting as close to 100% access as possible. That makes it hard to adopt new tech, and harder to dispose of old tech.
The only need for a fax machine I've had in recent times (in UK) is by proxy, a pharmacy needed a fax from my nephew's doctor in order to issue an emergency inhaler. Of course billions has been spent inter alia on replacing fax, through having integrated, networked medical systems in the UK but the IT companies involved have been "unable" to provide working systems. £20bn in ~2007, £12bn in ~2013 - I don't think we have a working national records system yet?
Companies who can make £100s of millions without delivering a working product wish to continue being paid, the money stops when they deliver the project ...
It seems like you're making the parent point for him/her. Yes, governments should support 100% access. Requiring a fax machine for anything, as in this case, goes against that. (Great that the UK doesn't force this on people though!)
Edit: one line I like to use when asked to fax something is, "Oh, sure thing, I know there's a museum right nearby!"
As I understand the requirement is not that fax is essential to the service, but there is a class of people you may shut out when you turn off the ability to fax.
It's why I'm very skeptical of anything cashless or requiring a smartphone, because there is, for the foreseeable future, going to be a population that can't participate in a cashless, smartphone society.
There is also a decent part that can only participate by smartphone though.
The answer is to accommodate both, not to shut out one or the other. That's what 100% access is.
Only want to participate by smartphone. I guess smartphone payers can handle cash, but people that can have credit cards can't pay via a phone?
> Sometimes law enshrines the ability for access using particular technologies.
> Government is often about supporting as close to 100% access as possible.
Isn't that a good thing? You probably don't need the latest whizbang tech stack and UI framework that needs a screaming CPU, tons of memory, and a 4G connection to implement a usable government service website.
You do need a good design process at both the UX and data processing level, focused on clarity and simplicity, with lots of unit and integration tests.
Yes, I'd agree slow progress with near universal access is good; I was being descriptive. But it still needs to utilise the best of established technology.
I'd be inclined to invoke Hanlon's razor here.
And I'd be inclined to counter with Grey's Law: "Any sufficiently advanced incompetence is indistinguishable from malice."
http://wikidumper.blogspot.com/2007/07/greys-law.html
> Surely no-one really believes that political will was thwarted here by failed technology?
Have no clue about the political will here, but the outdated failed technology story sounds very believable to me!
To be sure, I don't doubt that the technology failed. I meant only that it was surely not a case of political will to action being stymied by failing technology, but rather something between political inertia and will to inaction meeting happily with the failing technology which was, after all, its bequest.
I think people are overestimating the competence of government IT programs.
It's possible to quickly make the technical changes, but when politicians asked the small group of people with direct knowledge of the issue, this is the answer they got.
The small group of people are not plugged into SV mentality quick iterative work. They include (for example) a Perl programmer who has had the same job for 25 years and isn't interested in learning new skills.
This is not a very charitable view.
Keep in mind that at least in the Anglosphere, a point often raised about the government is that they are incompetent, that they need to be starved out, so usually
- they don't have the resources or backing to hire people who are up to date at competitive compensation levels
- the people left are those who have been able to survive such onslaught from their bosses, so they are presumably very jaded, cynical, and careful about new proposals
Our politicians do not empower the bureaucracy to be nimble, there is no reward for doing so. So I wouldn't expect it to be.
The incentives for government consultants are also generally not aligned to deliver quickly or competently.
There seems no way to win with government, unfortunately. If you de-fund it, you get cost cutting, reduced workforce, reduced capability, corruption, and disfunction. If you fully fund it, you get inflated costs, a bloated workforce, inefficiency, waste, grift, and more disfunction. No matter how you fund it, you get disfunction at the end of the day.
Maybe more funding to increasing salaries for government employees would be a good start instead of more funding in order to hire more people/bloat the system. I think it could over time make government agency work more attractive and prestigious and in turn attract the best talent.
You classify it as strategic readiness and have something like a Core of Engineers do the work.
Really the US military forces need an additional branch, the Civilian Core of Engineers. They would be held to high ethical standards (on the job), and all work they do for the government would become public domain. I imagine this could be combined with a New New Deal where public-works projects are generally funded, as well as a job guarantee / internship place to build a resume and gain social connections through.
As a minor nitpick, the word is "Corps".
I actually did mean Core as inspired by the Corps*. The distinction being that this group of individuals would be more like a not-for-profit company than the military, and thus, they aren't a Corps.
"In many armies, a corps is a battlefield formation composed of two or more divisions, and typically commanded by a lieutenant general." https://en.wikipedia.org/wiki/Corps
High ethical standards and the US Military do not go hand-in-hand.
This seems rather myopic.
Singapore is, by a lot of accounts, a well-functioning government with competent bureaucrats.
It's also a dictatorship.
The way to win is to refactor and reorganize properly. Unfortunately that is a difficult task in both management and politics as every special interest involved with the power to change is selfish towards their own goals. How democratic vs dictatorial only changes how the interests are distributed.
Maybe government shouldn’t exist in steady state?
It’s possible that people don’t have the ability to be consistent and you have to tack between the two strategies.
Look into the Judicial Council of California’s case management boondoggle. A few billion was spent with nothing to show for it. I’m pretty sure every “there isn’t enough funding” argument can be countered with real examples of waste, fraud, and abuse.
It's possible for both things to be true.
Money doesn't magically fix management issues caused by a lack of money; people who are not used to spending money may not know how to spend it well once they get it. Lottery winners generally fizzle out their winnings after a few years.
This is also a problem if you spend billions on consultants rather than shoring up the civil service; consultants aren't incentivized properly to deliver a good system on time.
Or underestimating the bureaucracy of the system. I did some work for a startup that interfaced with the government, and basically the government insists on doing everything in triplicate, getting approval from multiple layers of hierarchy, and does things on a yearly cadence.
I think the average HN-er doesn't realize just how slow/inefficient government really is. But when you read these stories it's apparent that the problem exists at an institutional level, and isn't really as simple as people think.
Most large organizations are like this. At the megacorp I used to work for the running joke was that any change, no matter how small, to the legacy systems started at 1 year & 1 million dollars.
At work, our customers are the Oligarchist Cell Phone Companies (about three at this time). As I was joking with a higher up about adapting agile programming, it doesn't matter if we're on a two-week sprint when our customer is on a two-year sprint. And yes, it took over two years to get them to adapt a change.
There's nothing about perl that makes it hard to change a mathematical formula for benefits.
Not are any "new skills" needed. Just funding to do the work.
Is it actually Perl though? I have my money on COBOL which is an entirely different story.
Perl? If only. We're talking COBOL here.
Avoid ageist comments, please.
Well you pretty much have to be 50 to have had the same programming job for 25 years, it's not so much the age as the tenure. But I will take out the age.
I would still say it's pretty shitty to assume that "SV culture" is the only thing that innovates, or innovates quickly, which is what you're implying.
There are major impediments that are non-technical for those of us in government positions, just take a look at the bidding process for projects that are pretty much required for something the scale of "overhauling" the system backing state or federal programs. Is it necessary? Probably not. Does it happen anyway because it's required? Yes. This isn't the fault of some tenured programmer, hell, they're probably not even consulted as a knowledge worker as part of the bid process. This isn't a failure of culture, this is the failure of a byzantine beauracracy built around protecting itself instead of the people it's meant to serve. You can have all the innovation that comes out of the NIH, the NSF, DARPA and the NSA; but the amount of time even practical papers and systems that are produced by these institutions are put into practice by the government thay funded them is exceedingly slim due to the institutional rust that's in place. Many of these "SV culture" companies are actively building things that were first concepts in a government lab, but by all means, blame the underpaid, overworked technical staff that actually designed them.
That's the culture. The SV innovation is that startups have gestalt views of things not a "management made it like this so I am unable to operate". The bureaucracy and the machinery that supports it is the government culture. People who don't understand this, often don't understand the natural offshoots that come from it: blameless post-mortems, prioritizing falsification, unlimited paid vacation, etc. These all work because the people participating have the culture.
The SV startup culture is a full-spectrum ownership thing. If your executive team isn't aiming at the right direction, you tell them that, and if you lose faith in the future as a result you go elsewhere. The executive team always relinquishes its power over you in order to maximize your effectiveness because that's the only thing that matters.
Government culture is different. It optimizes in a different constraint space. There has to be visibly equitable hiring, the needs of political expediency require things like "heads to roll", regression is more dangerous than lack of progression, etc. Because of this, government organizations operate differently. They're not inefficient. They're just optimizing to different constraints and they're incredible at doing so. I would hypothesize that you cannot retain SV startup culture at the government for any appreciable amount of time because of the constraint space - much of which is the result of natural law.
The age probably gives a misleading impression anyway, giving the idea that dynamic, change-the-world things are not done by 50 year olds.
But a widely cited statistic (don't know if it's true of course) is that the average age of a founder of a successful startup is 45.
https://hbr.org/2018/07/research-the-average-age-of-a-succes...
Even as an old, I'm not seeing that as ageist. The problem is not somebody with 25 years of experience. It's that they have settled into something comfortable and haven't kept up with 20 years worth of change.
What's the betting that the alteration needed is to change 0.25 to 0.025, say, and then record hours worked instead of days ... ?
But, hey they might end up paying someone more than a pittance, and that would never do, eh.