HealthCare.lol

20 Nov 2013

It should surprise no one that the federal health care program’s problems are much deeper than previously thought. Fresh off the heels of last month’s calamitous rollout, The Washington Post reported yesterday that a huge part of the site’s backend isn’t even finished yet. (Not even if you use the government’s definition of finished–there but you can’t use it.)

Henry Chao, the Obama administration official who oversaw the technical development of the federal health insurance marketplace, said Tuesday that his team has yet to complete 30 to 40 percent of the overall project.

Speaking before a subcommittee of the House Energy and Oversight Committee, Chao said the Centers for Medicare and Medicaid Services is still working on a number of “back office” aspects of the project, including a system to send payments to insurance companies.

Parts of the project that users would see — notably the Web portal, HealthCare.gov — are 100 percent finished, he said. But “the accounting systems, the payment systems, they still need to be” completed, Chao said.

It’s not clear if the the incomplete percentage refers to the backend or the entire project (the story says the former, the headline the latter). What is clear is that Mr. Chao is in hot water. At least the “[p]arts of the project that users would see” are “100 percent finished,” right?

One Hundred Man-Years of Solitude

23 Oct 2013

By all accounts the new HealthCare.gov website, supposedly America’s new frontier for ensuring everyone gets a shot at decent healthcare, sucks. Almost as bad is some of the reporting that’s been done around the failures.

The New York Times recently ran an article on the problems contractors face in fixing the website’s problems. The Times reports:

One specialist said that as many as five million lines of software code may need to be rewritten before the Web site runs properly.

[…]

According to one specialist, the Web site contains about 500 million lines of software code. By comparison, a large bank’s computer system is typically about one-fifth that size.

In response, David Auerbach at Slate rightly takes issue with these claims but completely misses the point. Auerbach—a former software engineer at Microsoft and Google, according to his personal website—sees it as a semantic issue: not all lines of code are treated equal. That’s, of course, true. All lines of code are not created equal, and that’s why no engineer/text editor/product manager/designer worth their salt even bothers talking about this pointless metric. The implication here is that the HealthCare.gov code is not as complicated as the Times suggests. That’s also undoubtedly true, but not because one woman’s 50 million lines of code is another woman’s 50 lines of code.

The elephant in the room here is that there is absolutely no way that the engineers who developed HealthCare.gov wrote 50 million lines of code, much less 500 million.1 We all understand that there are many moving pieces there, but that’s the case with any sufficiently complicated software. It seems to me that this rewriting-50-million-lines-of-code claim is a smokescreen for the real scandal here: the federal government’s horribly inept contractor-development practices. Hiring one contractor to do the database, one do to the design, another to do the login system, another for income verification, etc, is the road to doom. In other words, the point isn’t that HealthCare.gov is complex. The point is that it is a deeply mismanaged project. Now everyone from government leaders to the contractors working on the project are beside themselves promising quick fixes. It’s not going to happen. As Auerbach’s Slate colleague Matt Yglesias points out, Brooks’s mythical man-month is as elusive as ever: you can’t add manpower to a project to get it done faster. No, the contractors who made this mess are stuck with themselves and with each other until it’s fixed. And you can bet that will not be anytime before the end of the year.

  1. By comparison, the most recent major version of the Linux kernel has fewer than 17 million lines. Even huge operating systems, like the one that runs Macs, have many fewer than 500 million. Mac OS X 10.4 (froom 2006) had 86 million lines of code. That’s surely increased by now but is definitely nowhere near 500 million. Even the Curiosity rover, a massive feat of human engineering, only carried 3.8 million lines of code with it to Mars.

    The only other explanation for the exaggerated number I’ve seen was suggested by TPM’s Josh Marshall, who suggests that the number might include dependencies on other code (e.g., open source software and other code the developers relied on but didn’t themselves write). But those numbers cannot be included in a proper counting of the code, and even if they could—as Marshall himself then acknowledges–they are still unlikely to reach anywhere near 500 million.

Linkmania

30 May 2013

Recently I was reading a blog post by Jeff Moser about the future of programming and how programmers think about the languages they use. At just over 2,000 words, it’s not a very long post. But it included some 77 links, for a word-to-link ratio of 27 to 1—one link for every 27 words. That’s more than one link for every two sentences. (Note: it’s actually a three-part blog post, but since the second two parts suffer from the same high word-link ratio, I’m ignoring them.)

It’s important to recognize that the way we consume information online–hypertext–is different from the way we consume it offline–static ink. In ink, you’re forced to explain yourself or assume your readers know what you’re talking about. On the web, we have the benefit of hypertext. Obviously, hypertext has been there from the beginning and has done as much as anything to catalyze the web’s exponential expansion. But there is great cost in saturating thoughts you’re trying to convey with links to other sources. Many times, it is necessary to link–for example, it would be unfair for me to write this criticism without linking to the criticized post. So it’s clear that I’m not advocating a linkless Web. But consider this sampling of the links in just the first third of Moser’s post:

  • “programming languages class” -> the author’s programming class from Purdue
  • “Scheme” -> Wikipedia article for “Scheme”
  • “latest version of C#” -> a .doc of the C# language spec
  • “lambda functions” -> Wikipedia article for “Lambda calculus”
  • “Plato expands on this” -> Amazon page for Notes on the Synthesis of Form by Christopher Alexander, with title ‘The quote came from the front of this book.’
  • “Phaedrus” -> Wikipedia article for Phaedrus
  • “fanboy” -> Wikipedia article for “Fanboy”
  • “think different” -> YouTube link to classic Apple campaign

Moser even includes a link under the text “Intergalactic Network” to a page called “Internet History”!

My point here is not to single out Moser, to say that links are bad, or to say that citations are unnecessary. But I do think that most of Moser’s links are bad, and are symptomatic of a larger problem that bloggers have, which is that they assume their readers are as invested in their subjects as they are. This is certainly a false assumption. If your aim is to contribute something new, then you need to eschew linkmania and explain what you mean. If you’re not willing to explain what Scheme is, or what a fanboy is, or where and when your programming language class took place, then it’s either (a) not relevant or (b) not necessary. In other words, I don’t care which programming language class you took and I don’t need to be told what Scheme is, because I’m reading an article about the future of programming languages.

From a pure reading perspective, it’s incredibly distracting to have to stumble over a link every sentence or two, because links tease: click me for more information. Sometimes, it’s obvious where the link goes–e.g., many book titles are linked to their Amazon page. Most of the time, though, the reader will have to roll over the link to see where it goes, requiring him to: (1) break his concentration to roll over the link; (2) try to decide whether to follow the link now, ignore it, or save it for later; (3) remember where he was in the text and return to it.

But of course, many links are useful, informative, necessary—so which ones? I think a decent heuristic is: does the link sufficiently enhance the experience of reading the main material, and will more than half of your readers click on it?

Who wants a built-in TV on an airplane?

13 May 2013

The New York Times recently ran a story about changes in the airline industry in recent years. A few highlights:

But most have begun putting Wi-Fi and individual televisions aboard their planes, installing more comfortable seats for business passengers and investing in mobile technology that gives passengers more control over their travel plans.

[…]

Delta also spent $140 million over two years to develop a new Web site, which it unveiled last year. The site allows the airline to sell more services to passengers, including upgrades to premium seats, or booking hotel rooms or cars. At a conference last year, one executive compared the company’s aims to something Amazon has on its Web site — the ability to bundle offers or suggest products based on a passenger’s history and preferences.

I recently flew from Seattle to Chicago on United on a Boeing 737-800, which has 126 standard coach seats. Every one of those seats was equipped with a display hooked up to DirecTV. Only a couple people, from what I saw, took advantage of the programming and watched something during the 5-hour flight. What most people did was ignore their display without turning it off, leaving it to loop endlessly over the 5-minute DirecTV marketing pitch.

I have no figures, but how much might it cost an airline to install and maintain these individual displays? Add on the costs to get the content, plus the cheap headphones most airlines give passengers these days, and you’re wasting a bunch of money on something that barely anyone wants. I, like everyone else, travel with my phone, which has all my music on it, and my headphones. And, like many people, I travel with my iPad. It seems to me that having good Wi-Fi on the plane is orders of magnitude better than having a personal display, because I can simply load Netflix on my iPad. (Even assuming a lack of Netflix-capable WiFi, I’d rather watch a downloaded movie on my retina-display iPad than the low-quality airline display.)

Everything the Times article cites as progress—WiFi, site redesigns, better mobile apps—are good things. Except the seat displays. As tech has become more personalized, especially with the popularity of tablets, we’re getting more accustomed to carrying our entertainment around with us. In that world, the airlines would be better off ditching the seat displays, the awful headphones, and the content deals. Instead, they should invest in power outlets and Wi-Fi.

The BRAIN Initiative

02 Apr 2013

Today, President Obama unveiled Brain Research Through Advancing Innovative Neurotechnologies:

Today at the White House, President Obama unveiled the “BRAIN” Initiative—a bold new research effort to revolutionize our understanding of the human mind and uncover new ways to treat, prevent, and cure brain disorders like Alzheimer’s, schizophrenia, autism, epilepsy, and traumatic brain injury.

A few things:

  1. This is fantastic.
  2. Contrived acronyms like this and countless others in D.C.—PATRIOT, etc.—are nauseating and senseless.
  3. If the name were BRAIN Research Through Advancing Innovative Neurotechnologies, it would almost be redeemable, because then it would be a recursive acronym. Since recursion—self-reference—is so incredibly central to the idea of consciousness, I think it might provide some value and undo the stupendous ridiculousness of the acronym.

List of Eponymous Laws

04 Mar 2013

I was recently reminded of Brooks’s Law—that “adding manpower to a late software project makes it later”—and it led me to Wikipedia’s full list of eponymous laws.

Besides the classic Asimov robotics laws and Clarke’s laws, a few of my favorites:

Benford’s Law: In any collection of statistics, a given statistic has roughly a 30% chance of starting with the digit 1.

Betteridge’s Law: Any headline that ends in a question mark can be answered with a “no”.

Conway’s Law: Organizations which design systems are constrained to produce designs which are copies of the communication structures of the organization.

Gall’s Law: A complex system that works is invariably found to have evolved from a simple system that worked.

Hofstadter’s Law: It always takes longer than you expect, even when you take into account Hofstadter’s Law.

Kerckhoff’s Principle: A cryptosystem should be secure even if everything about the system, except the key, is public knowledge.

Linus’s Law: Given enough eyeballs, all bugs are shallow.

Mooers’s Law: An information retrieval system will tend not to be used whenever it is more painful and troublesome for a customer to have information than for him not to have it.

Niven’s Law of Time Travel: If the universe of discourse permits the possibility of time travel and of changing the past, then no time machine will be invented in that universe.

Peter Principle: Employees tend to rise to their level of incompetence.

Now I’m off to inspect the even-longer Wikipedia list of paradoxes.

Peter Staley to Pat Buchanan: “Use a condom”

26 Feb 2013

Over the weekend I finally got around to watching How to Survive a Plague (available on Netflix), a wonderful documentary about the AIDS activism group ACT UP (the AIDS Coalition to Unleash Power). The film makes heavy use of archival footage from the ’80s and ’90s. One particular moment that I found illuminating was a clip from Crossfire featuring an exchange between Peter Staley, a prominent ACT UP activist, and Pat Buchanan. The first half of the segment is available on YouTube, but I haven’t been able to find the full clip. However, you can see it in the film at the 30-minute mark. Here’s a transcript of the segment:

: Peter Staley, you have the AIDS virus, and I am sorry. But, don’t you think that the Federal Drug Administration [sic] has a responsibility not to let people such as you have quacks that could cause even more harm than you already have?

: The problem is, is that the FDA is using the same process to test a nasal spray as it is to test AIDS drugs. And it’s a 7-10 year process.

: You have the FDA giving you a drug. So far you’ve got AZT. Why would—

: Which I can’t take because it’s far too toxic, and over half the people that have HIV can’t take it.

: OK, but the FDA says there is nothing else that is worth anything.

: Mr. Staley, this is gonna astonish you, but I agree with you a hundred percent. I think if someone’s got AIDS and someone wants to take a drug, it’s their life, and if it gives ’em hope, you oughta be able to take it. What I wanna ask you is whether you know of anything that you think might be some kind of miraculous cure that you think they’re sittin’ on at FDA.

: There are over one hundred and forty drugs out there that the FDA has identified as possibilities, and are in some stage of being looked at right now.

: Why are they holding back—

: Among that one hundred and forty, there’s gotta be one or a combination thereof that can slow down this virus or halt it in its tracks.

: You’re just simply carrying the virus, is that correct?

: I have a few very minor symptoms, and my immune system is virtually shot.

: What would you like to take?

: I would like to be able to take dextran sulfate, legally. I’m taking it on the underground right now.

: Well why not, Mr. Braden?

: Because for—I don’t know anything about dextran sulfate, and neither do you–

: Well I’ll tell you this: it’s an over-the-counter drug in Japan, and has been for twenty years.

: But—

: Over the counter.

: Over the counter in Japan?

: Yes.

: But if the FDA says—

: Mr. Staley–

: I’m only asking that they be released after there’s a minimal amount of efficacy, not a one hundred percent test.

: Well, a final question to you. Let me get something here. You’ve got the pink triangle on your shirt.

: Yes.

: I gather that means you’re a homosexual.

: Yes.

: Lookin’ in the camera, what would you tell some kid—say you had a younger brother, twenty-one years old, who also might have homosexual tendencies—What would you tell him if you wanted him to live a long life?

: Use a condom. And also to use a lubricant by the way that has a medicine that can—

: But aren’t you, this is Russian roulette.

: It is not Russian roulette. It is Russian roulette to not give people this information when human nature dictates that they’re gonna go out there and they’re gonna have sex.

: You mean celibacy is impossible?

: It’s just not gonna work. People aren’t gonna do it and lots of people are gonna die. Now would you rather have a lot of people cheating under celibacy with thousands of people dying or would you rather save those lives and let them have sex?

: Well I think that uh, well. Thank you very much Peter Staley, thanks for being in our studio, Mr. Braden and I’ll be back in a minute.

Some things never change.

Hello, World

17 Feb 2013

Hello, world!

Hopefully more interesting things are to come.