I've been driving a borrowed Miata around Los Angeles the last few days. It's a 1993 model, fairly similar to my 1986 MR2 in size and performance and feel. It actually feels just a bit less peppy, with a narrower power band, maybe 3500 to 4500 RPM. It's a convertible, and that finally prompted me to do something I should have done a long time ago: buy Kona a dog seat belt. I did some online research and found very strong warnings against harnesses with load-bearing nylon buckles, which could easily break under stress. At the Centinela Feed pet store, that's all they had, so that's what I got. At least she can't jump out, although I think she's actually a bit nervous of the Miata—certainly she's terrified when I raise the roof with her inside— and she just sits quite still, or lies down so her nose almost interferes with shifting.

I'd like to see the Supreme Court get more liberal. The notion that the justices make rulings out of the pure substance of abstract law is nonsense. Every justice, liberal and conservative, makes decisions out of a combination of legal reasoning, personal beliefs, political motivations, what impact they think their rulings will have on the legal system and thus on society in decades to come, and so forth. There is certainly a need in any society for a trusted, objective arbiter, and we put the institution of the Supreme Court on that pedestal whether it deserves to be there or not because there are few alternatives. I think that mostly it doesn't deserve to be there, but I put it there too. It's certainly a more transparent and institutionally constructive entity than, say, the Guardian Council. But it's still nine people deciding national law by majority vote:

Each year Brennan asked his law clerks to name the most important rule in constitutional law. Brennan gave them the answer after they stumbled around, naming one great case after another. "This," he said, holding up one hand with his fingers spread, "is the most important rule in constitutional law."

So Sotomayor seems like a good enough choice. In some ways it's an uphill appointment for Obama, since she replaces a moderate liberal and at best still doesn't change the overall balance. Mostly I hope she won't be worse than Souter, who was okay. In particular I hope she won't be a stealth anti-abortion vote. I'm happy about the diversity; I'd love to see four or five women on the Court in my lifetime. (And fifty or more female Senators, and a woman president, and heck, some atheists and more gays in office, and everybody to have a pony.) So I'm okay with this pick, but not especially worked up one way or the other.

Well, until now. It turns out she's a Strunk and White zealot:

If you have read Strunk and White, Elements of Style, reread it every two years. If you have never read it, do so now. This book is only 77 pages and it manages, succinctly, precisely and elegantly to convey the essence of good writing. Go back and read a couple of basic grammar books. Most people never go back to basic principles of grammar after their first six years in elementary school. Each time I see a split infinitive, an inconsistent tense structure or the unnecessary use of the passive voice, I blister. These are basic errors that with self-editing, more often than not, are avoidable. —Sotomayor

What's horrible about Strunk and White?

Educated Americans have a tendency to think that (i) intelligence can be directly assessed through the surrogate of compliance with the rules of Standard English grammar, and that (ii) compliance with the rules of Standard English grammar can be checked quickly and easily by glancing in Strunk and White's brainless little pamphlet of 19th-century grammar nonsense. Both propositions are wrong and dangerous, yet tacit acceptance of them is widespread. —Language Log

This style of thinking can lead to nonsense results, such as the ill-considered, often blindly (but poorly) followed rule against using the passive voice, which contradicts common, valid usage in English. Orwell and White both famously advise against the passive; both use it in their own writing as much or more than average. But there are many reasons to use passive constructions and many reasons not to. A more useful rule might be, Don't lie, a special case of which is, don't use language to obfuscate inconvenient facts, a special case of which is, don't use passive verbs to hide agency when that agency is relevant. Example: Don't say "Mistakes were made" as a way to avoid saying, "I made mistakes".

It also leads to pronouncements of fact which are actually based on personal opinion, and often wrong.

"I," said the president, who is inordinately fond of the first-person singular pronoun, "want to disabuse people of this notion that somehow we enjoy meddling in the private sector." —George Will

But Obama turns out empirically to not be inordinately fond of the first-person singular pronoun.

I took the transcript of Obama's first press conference ... a total first-person singular pronoun count of 206 in 7,775 words, or a rate of 2.65%.

For comparison, I took George W. Bush's first two solo press conferences as president ... a total first-person singular pronoun count of 300 in 6,681 words, or a rate of 4.49% ...

For a third data point, I took William J. Clinton's first two solo press conferences as president ... 3.87% for first-person singular pronouns overall ... —Language Log

It seems to me that, while a trial judge's value lies in being an impartial third party who facilitates retail-level access to justice for all parties, an appellate judge is called upon to provide a different flavor of judgment. The parties have had access to the system, and the appellate judge is faced by definition with the tougher, more ambiguous or contradictory problems. The notion of a purely logical judging machine that interprets the law as written is science fiction. A human appellate judge is more likely to be faced with a problem such as, "principle A is sound, and principle B is sound, but in this case they conflict with each other. We all agree we want outcome O, but which balance of principles is more likely to lead to that outcome more often over the next twenty years?" For example, police sometimes exaggerate or even lie to judges in order to obtain search warrants. Exactly how high should the legal bar be all before all fruit of the warrant is thrown out? There will be honest mistakes that look like lies, and lies that look like honest mistakes or that don't get caught. What's the right balance to let honest police do their jobs almost all the time, protect the innocent almost all the time, and deter corrupt police almost all the time? Is such a balance possible?

I worry about the judgment of a Supreme Court Justice who mistakes Strunk and White for good advice. In fact, let's do a litmus test on the current justices, and any of them who also swear by Strunk and White should retire.

After sending only "Disc 1 of 4" the first time, the Stanford records department apologized and said they would get everything to me right away. True to their word, I got four CDs in the mail yesterday, including the latest pictures from my surgery and followup. I've incorporated them into the main post on the crash. In this post, I'll comment on a few aspects of the imaging process.

All of the image come on CD-ROMs; each CD-ROM has all of the software needed to view the images. The images are all apparently in DICOM format. "DICOM differs from other data formats in that it groups information into data sets. That means that a file of a chest X-Ray image, for example, actually contains the patient ID within the file, so that the image can never be separated from this information by mistake." The software on the CD-ROM is a Windows-only program, of course, and you can view images one CD-ROM at a time. There's no obvious way to combine them, in order to compare series that happen to be on different disks. I guess they don't want to do DVDs.

The data on the disks includes the X-rays, the raw data for the CT scans, a few pages detailing my radiation dosage, and some processed images of CT scans. CT scans are where they take hundreds of x-rays in slices and then the computer assembles a three-dimensional model. The raw data looks pretty terrible but once the computer is done you get some eye-popping results. Here's an example:

With the home version of the software, you can manipulate the original CT data in a 3D model, and play with the contrast, but you can't get anything like the fancy, false-color images that they are producing at the hospital. You get pictures of whatever the original tech thought was interesting when they were looking at the color 3D model in the lab, but if you use the home software to make more, they look like this:

One of the most frustrating aspects of dealing with American emergency medicine is the record-keeping. The hospitals and clinics are clearly going to great lengths to do something about record-keeping: every piece of paper I saw was barcoded; reams of stickers with my name and information were regularly dispensed and attached to pages. When I was at the clinic for my first followup appointment, with x-rays of my clavicle on the computer screen, the doctor was able to bring up images from my hospital stay 2 weeks earlier at a different location.

But when I was originally discharged from the hospital and I asked about records, all they could do was tell me that the records department was closed for the day and give me a form. I send the form in right away, asking for "Medical Records", "Radiology Film/CD", and "Billing Records". Two and a half weeks later, I got a CD in the mail, marked "1 of 4", which had one of my CT scans (I think I had two) and a few of my x-rays, all from the original stay. (Naturally all viewable only in a Windows-only, proprietary program. The images themselves were allegedly in a standard format for the medical field, but four different linux programs were unable to show anything but solid black, and of course we never imaged my brain so that can't be right.) I was planning to take a picture of the screen in the office, and I was able to bring up matching images of my surgically repaired clavicle from two angles compared to how it looked a week ago (my proficiency went unremarked by the doctors), but failed to take a picture of the screen because it had been more than five minutes since I had last recharged my horrible, horrible Gphone, and forgot (again!) to try and save a picture to my thumb drive. I did call the medical records people and explain about the "disk 1 of 4" and just got an apologetic phone message, so perhaps those records will arrive sooner rather than later. I do want to show you, because they are very dramatic.

I stopped by the patient records office at the Stanford Clinic (in a different town from Stanford Hospital, but fortunately on my path between home and work), and they said that there was no way I could walk up to the counter and leave with either billing or medical image records in the same visit, but that I could fill out a form that they would fax to the main Hospital which would mail me my records, and just for billing I could call Billing and ask them. So I did that (650 497-8123), and they said that they only handle physician records, not hospital records. You may recall that American physicians are, presumably for financial and legal and historical reasons, not typically employees of the hospital in which they work, but instead independent contractors or something. So that office promised to send the "physician" portion of my billing record right way, and told me to call hospital billing for the rest. So I did that (1 800 794 8978), and they confirmed that I had three accounts with them, one for each period of contact. I've had four contacts with the Stanford medical system: the initial hospital stay, the followup, surgery, and the second followup, so who knows what relationship the three accounts have to those four contacts. All three accounts were still in processing with insurance, and therefore it was impossible to give me a copy of them. But in 2-3 (more) weeks, Shannon would be sure to keep an eye on the account and then send me a copy. "A copy of exactly what you sent the insurance, or what the negotiated values are, or both?" "What we sent them". So that may be coming. Then I asked if there was any other division of costs that I should investigate, beyond doctor vs hospital. "No." "What about the ambulance?" "Oh, right."

So she transferred me to Stanford Admitting, which transferred me to the "Transfer Center" (that's the center for transferring people between hospitals, not calls between departments), which had absolutely no idea how or who ambulances billed but suggested calling Admitting to talk to a financial counselor, and calling the ambulance directly, and calling my insurance. I happened to know which ambulance company it was from the police report, so I called AMR (888 650 2663) and they gave me the number for billing (800 913 9106), and Billing said they have sent a bill to insurance, but have the capability to send me a copy of the bill without waiting for insurance processing, and in fact told me the total over the phone. It cost $2285 to be treated on the side of the road and then driven to Stanford Hospital. Readers in Seattle are reminded to continue voting for Medic One, publicly funded emergency medical service.

So. Bottom line: one thing that Stanford Hospital is completely incapable of doing is to answer the question, "How much is this service going to cost?" How many other businesses can you think of that can't answer that question? And do you think they would still be unable to answer that question if I had said, "I have no insurance but I do have this briefcase full of cash; how much are we going to need?" And the other thing they can't do is hand you or email you a copy of your medical images or other records. Perhaps I'm in a radically small minority of patients that would actually have wanted these things as I walked out the door. But even if I am, perhaps I shouldn't be.