December 1, 2013
"In order to boast how far the federal site has come, HHS laid out how bad it used to be."
Wise-cracks The Atlantic's headline writer as Garance Franke-Ruta displays charts from the HHS report.
Subscribe to:
Post Comments (Atom)
Posts
20 comments:
My take on some folks at the Atlantic is that being a right-thinking feminist or modern feminist comes first.
If so, Obama and his liberal/progressive coalitions are natural allies.
Hence, the constant cheerleading and reluctance to point out just how much of a mess Obamacare is.
At least Ezra Klein, uber-progressive health-care wonk, knew his job was to try and be objective as possible on the rollout to save his career and some of his progressive principles.
As I see it, the law is still the problem. Most supporters are still shuffling around, hoping it'll all work out.
Didn't Domino's Pizza do this?
Self-Flagellation as Corporate Strategy: Domino's "We Suck!" Campaign Piles on the Sales
New headline:
The Biggest News in the Healthcare.gov Report: Just How Bad It Was
Now with 100% less snark.
They are defining improvement with metrics like 'throughput' and 'stability', though there's no mention of the site's ability to allow people to buy insurance. Is there anyone verifying the ability of the site to accomplish it's primary purpose? Anyone?
Didn't think so. Besides, Obama had moved on...
They have now done the impossible--a relatively simple technological task that tens of thousands of businesses do every day.
It was inevitable.
Which is why the focus on the web site by the critics has been so foolish. It was fixable.
The passed a golden opportunity to use the web failure to discuss the more important issues, and to propose alternatives.
The Stupid Party.
rehajm said...
They are defining improvement with metrics like 'throughput' and 'stability', though there's no mention of the site's ability to allow people to buy insurance. Is there anyone verifying the ability of the site to accomplish it's primary purpose? Anyone?
That will come. For now all they have to do is show that the web site is not crashing. There is still plenty of time to sign up, and they can always change the law by presidential proclamation to make the requirements be consistent with the actual capacity of the web site.
Again, the issue is not the mechanics of the web site.
Obamacare: We Suck Less!
Error messages are designed to blend in with the background so as not to alarm the user: success!!!
So does the website crash only when used by more than 10 people? 20 people? 40?
And why the disclaimer on the website that your data is not secure? So how do they justify violating HIPAA regs?
Hope for change in 2014. Lots of democrats up for reelection. Most of them voted for Obamacare.
"Didn't Domino's Pizza do this?"
Sure, the pizza's bad. But you get two!
"they can always change the law by presidential proclamation to make the requirements be consistent with the actual capacity of the web site."
Change the law by presidential proclamation??
Lefties just yearn for a dictatorship or fascism, don't they?.
Democracy and separation of powers be damned.
The key fact in the article is that "per page system timeouts or failures" is still running at .75% (down from 6%). So if it takes (optimistically) 10 pages to go through the registration process, that's a 10% freeze out, right there. So for 50,000 concurrent users, you're looking at 5000 or more stalled applications.
It's far from clear from the reports thus far that the website is anything near acceptable levels of reliability.
Looking over the graphs provided, it looks like they've increased the capacity of the site by about 5x and decreased the crashes by about the same factor. At 95% uptime, that's still 3 or 4 minutes out of every hour that the website is down.
It's hard to say what the dividing line should be between "terrible experience, but doable" and "impossible nightmare," but it's not clear to me that they've passed that threshold yet. Something like 90% of the targeted population hasn't even seen the site yet. Are they going to give partial credit for what is still likely to be a very bad experience?
Almost five PERCENT downtime?
Over half a percent blocking-error rate per page load?
Oh no no no. This isn't fixed. This isn't good. This isn't even at "levels vaguely acceptable for beta if we're not being very strict". Two orders of magnitude better than this (99.9% uptime and > 1 in 10,000 fatal errors per page load) would be what I'd be looking for, to call a site of this type and criticality "sort of maybe ready to launch".
This is horrible performance. There's no other word for it.
As I suggested last month, I'm sure they can make the site usable for as long as it takes to get through a press conference. The problem is that the site still doesn't do what it was supposed to do in the first place and I seriously doubt it's maintainable at any level. Also, the guys fixing site shouldn't be the same guys defining metrics and measuring how close the site is to meeting them. BTW, it really shouldn't take longer to put up a working website than it took FDR to defeat the Nazis.
I give this warning freely to you all. Do not use the governments website. If you do, you might as well just throw your wallet with all your personal id, and pin's out into the street for any random jackass to find. I'll never use this site. Never.
95% uptime is atrocious. You're dead if you're in the private sector with 95% uptime.
Post a Comment