Hacker News new | ask | show | jobs
by jacobyoder 893 days ago
Absolutist takes like this never achieve much.

Yes, it could be better. But an mvp-type launch that does some things well, without trying to achieve everything on day one, would definitely not be a "complete waste". If successful, it's laying the base for the next iterations of functionality.

For a lot of folks, they have one or two W2 forms to enter, and that's it. It's really not that hard. If even a million or so folks get benefit from it to start with, that's great, and we should then build on that.
2 comments
nerdponx 893 days ago
Let's also remember the context of this project. It's not just an overly-cautious MVP rollout. This project is something that has tremendous resistance against it, for stupid reasons, but the resistance exists nonetheless. Much like ACA benefits, people will only realize it's actually a good idea after they've gotten used to it. Starting small, with a minimal project that is most likely to succeed, is the best way to ensure that the project can continue at larger scale in the future.
link
GreedClarifies 893 days ago
The IRS should play try to create an actual MVP. Create an API which allow for efile (which I assume already exists given turbotax) and create an API which allows the download whatever information the IRS has on the tax payer.

Then allow companies or open source software to create whatever value add they want on top of these APIs.

That sounds like a minimalist launch.

link
dimal 893 days ago
That does not sound like an MVP. Then it’s up to random chance that they create the right API format and then someone makes an app to use said API and doesn’t also use it to steal data from gullible people. You’ve multiplied the number of ways this can go wrong and reduced the chance that anyone would use it.
link
jacobyoder 893 days ago
You've increased the places where things can go wrong (in third party client apps) while not having any control over their quality. We'd hear "oh, IRS opened things up!", then deal with a dozen half-baked clients, and conclude "IRS sucks at something so simple!".

Agreed - public API first is the wrong place to start.

link