[aggronn]

This is disgusting. I've had to use it extensively for research, and while it 'gets the job done', its such a pain to use. I've asked myself who put this together, and I always assumed it was a team of interns working across a decade for each version.

I can't believe this is the best they could do for $33 million. I just can't. They better have some slick stuff under the hood.

[wpietri]

Good UI comes from collaboration and iteration with users. Apple did something like 100 iterations on the first iPod. Typical consumer electronics companies do 3 to 5.

The government RFP and budgeting process is practically designed to prevent iteration and exploration. I can just hear Senator Blowhard of Cornhusk now: "So you want an unknown amount of money to keep building things until users are happy? That's an obvious boondoggle. Come back with a complete specification and put it out for bids from major contractors. We must be responsible with every penny of taxpayer money. Harrumph!"

So of course the software has terrible usability. Anything built like that will be terrible.

[mbesto]

> Apple did something like 100 iterations on the first iPod.

I don't doubt you, but do you have a reference for this?

[wpietri]

For the Apple figure, I'm pretty sure it was Inside Steve's Brain. This was some years back and I'm away from my books, though. I'm sure it was 100 iterations on some Apple device, but I won't swear it was the iPod. The 3-5 number I got from a consumer electronics designer I took a class with.

[crusso]

Yeah, and for $33M they couldn't build in some iterations and bring in focus groups or hire someone competent to get it at least partially right the first time?

These projects are handed out to reward the connected contractors, contributors and lobbyists that know how to work their target politicians and bureaucrats. There are ZERO incentives to do quality work. No one will lose his job over this software travesty. No one will get demoted. Rather than bemoan the waste of $32.9M on the job, politicians will simply push for higher taxes.

[wpietri]

Putting iterations into a procurement process is very hard. How does one bid against that? How can one control costs? How can we know the cost of the project up front if we don't know exactly what will be done? And if there's a disagreement about whether the project is really done, how do you litigate without a clear spec?

It's not totally impossible to do something agile in a government context, but it requires much more skill and trust from all involved. It's risky. And nobody outside the government is very understanding when things blow up, which happens more when you pick risky choices.

[adrr]

NYTimes paywall cost $40m. Heard bank software refactor can easily exceed 100m and take many years. I've personally seen 25m project fail and the code thrown away. Its not uncommon especially when you have 3 to 4 "support" people per developer especially when dealing with a bunch of stake holders aka executives with fragile egos.