Hacker News new | ask | show | jobs
by lucb1e 1913 days ago
Proprietary route:

input: 5+ years and 100 million dollars

output: crappy system

maintenance: still necessary, hire anyone but probably the expensive guys from before

The route that these parents demonstrated:

input: a few months and presumably a whole lot less money

output: something that seems to be well-liked

maintenance: still necessary, hire anyone

Regardless of "but who's going to maintain it", the benefits should be clear here. People don't work on open source code because they get paid to and shrug when they do something useless or even detrimental just because the boss says so. They work on it for a passion. Now if you hire a company to write the open source code, you kind of lose that benefit, but if the development is out in the open, the public can at least keep track of it and say "but this doesn't make sense" or "let's get a working system before we spend another 4 years over-engineering and bloating this". This open model is how the corona tracker was developed in the Netherlands and it worked super well. The question is now whether the government will dare to do it again with the next IT project.

2 comments

> maintenance: still necessary, hire anyone but probably the expensive guys from before

Don't forget: sometimes the copyright is still owned by the contractor that developed it, at which point the options are only "hire the expensive guys from before." Want to make a change and the vendor can't/won't? Oops, guess you're starting over from scratch! Or you don't make the change you wanted to and live with it as-is.

> Regardless of "but who's going to maintain it", the benefits should be clear here.

Are they clear?

Or is this the "Chrysler Comprehensive Compensation System" all over again? aka the gigantic disaster that somehow spawned the Extreme Programming "experts" and implemented the easy 80% while missing the really hard 80% (yes, that totals to 160% intentionally).

It's really easy to produce something that majority like but doesn't get even basic use cases. For example, let's start with some simple stuff:

- Does it meet GDPR guidelines?

- Does it meet accessibility guidelines?

- Does it meet security guidelines for protection of personal information of minors? (Apparently the original government software completely blew this off. As always.)

As I have pointed out previously on HN, that adds a bunch of cost to government software that MUST be paid if the software is part of government functionality.

I can whip out a single page app in a hurry. Ask me to comply with those issues and I'm going to have to spend a lot more time on things.

Don't get me wrong. Big IT projects like these always become boondoggles. However, everybody always simply gives the open source project the benefit of the doubt at being "better" when it probably just blows off a lot of functionality.

Government CRUD applications have to be able to handle the majority while still allowing the 0.1% to be handled.

> Are they clear?

Yes they are. The benefits are clear, because apparently a whole bunch of users got together a spent a bunch of effort building something that solve real problems that they themselves were having.

> everybody always simply gives the open source project the benefit of the doubt at being "better" when it probably just blows off a lot of functionality.

If people are using it, and spent a bunch of time and effort to solve their problems, then almost by definition, it is solving a problem that they were having.

So yes, we can just assume that it is better in some ways. If a bunch of people are using it, then almost by definition, it is providing value to some people.