Velocity Over Time

I was never much of a Mac user before OS X came out, mostly because I thought Macs sucked, and until OS X, they never had preemptive multitasking. But I did use System 7 quite a bit in college, because if you wanted to go to the printer and publish a magazine, you had to send them a QuarkXpress file, and the only machines we had access to that had the specs to run Quark were Macs.

The computer labs on-campus were, from what I remember, still mostly running either Win 3.11, Mac OS 8, or BSD, and the biologists seemed to favor Macs, probably because all the genomic sequencing apps only ran on Macs at the time, but since I never did any bench research as an undergrad, I just eyed the real scientists working from a far.

Fast-forward to 2002, when I bought my first notebook, which was an iBook G3 733 MHz. It came with Mac OS X 10.1 Puma and Mac OS 9.2.2 pre-installed, but I never really used Classic mode much. Then I destroyed my iBook by leaving it in the back of a trunk in 90°F weather. I paid Apple $300 to try and fix it, but since they couldn't, they gave me an iBook G4 933 MHz with Tiger pre-installed, and I didn't bother installing Classic mode. I never really missed it.

So when Steve Jobs extirpated Classic mode from Leopard because it would've been a pain in the ass to try and port all that cruft to the x86 architecture, I didn't particularly care.

But what got me thinking about running retro software again was a Friendfeed post by Mona Nomura discussing running Netscape 4 on Windows 95. Any version of Windows excluding XP and possible Win 7 are toxic waste as far as I'm concerned, so I'd never try anything like that, but I wondered if it would be possible to run Netscape 4 on a modern Mac system. Of course, the binaries are only available for OS 9/Classic, so on an Intel Mac I was out of luck.

Or was I?

The only known way to get OS 9 to run in Leopard on an Intel Mac is to use SheepShaver, which is basically a Power PC emulator. I tried to follow the instructions on using SheepShaver to run Mac OS 9 from Dan on uneasysilence. The problem was the ROM file. The link to it is dead. Like Dan says, Apple actually has the necessary ROM available for download, but unfortunately, you have to either already have a running install of OS 9 or Classic to extract it with Tomeviewer, or you had to do things the ridiculously hard way.

---

The hard way involves getting an emulated installation of System 7 up and running. Yes, this does in fact sound like a ridiculous waste of time, but, hey, if you've gotten this far, clearly, you've got plenty of time anyhow.

The way to do this is through Basilisk II, a Motorola 68k emulator. Like SheepShaver, you still need a ROM image to be able to get to work, but the older images are more readily available. This is probably technically copyright infringement, but if you happen to own an old school Mac, then this may actually technically still be fair use. But I am not a lawyer. The 1 MB Mac ROM image will probably sufficient.

If you're a masochist, you can try and compile Basilisk II from source. Macports even has a portfile for Basilisk II, so that may be slightly less insanity-inducing, although when I tried it, the compile crapped out with an error that I had no intention of tracking down. But what I did was just download the universal binary of Nigel Pearson's port of Basilisk II to Mac OS X. Gwenole Beauchesne also has Mac OS X binaries but I haven't tried them.

I was able to pretty much follow the instructions on how to setup Basilisk II and System 7.5.3 on Windows from e-maculation, except, of course, I was doing this on OS X instead of Windows. The System 7 boot disk is key. As far as the disk image goes, what you can do is use Basilisk II to create one with a .dmg extension so that you can later mount it in OS X and transfer files that way until you get System 7.5.3 running, because without System 7.5.3, Basilisk can't mount your home directory in your emulated installation.

Once System 7.5.3 is running, you need to install an old version of Stuffit in your emulated installation, copy over the TomeViewer Stuffit archive to your disk image, then decompress and run TomeViewer on the Mac ROM Update 1.0 file to extract the Mac ROM image, which you can then copy back to your OS X partition

Once you have the ROM, you can follow Dan's instructions exactly and have Mac OS 9.0.4 running. From there, you can install Netscape 4, or even NCSA Mosaic, and surf the web like it's 1999.

OK, I know Input Managers are notorious for causing system instability, but I'm crossing my fingers with SIMBL, the Smart Input Manager Bundle Loader, which hopefully really will minimize the possible collisions.

The reason why I installed SIMBL was because of GreaseKit, which gives Safari (and any other WebKit-based app) some of what Greasemonkey gives Firefox: user scripting.

The reason why I wanted to install a user script was Helvetireader, which is a modification of Google Reader's look and feel.

Disclaimer: I do not own an iPhone, although I do own an iPod Touch.

I had to buy a new phone in July because I had lost my old one (a Motorola SLVR.) I could've decided to wait a week for the roll-out of the iPhone 3G. I even thought about waiting for the Blackberry Storm to come out, until I realized that it wouldn't be for several months. But not having a cel phone is kind of difficult when you don't have a land line, so I ended up with a Palm Centro, which is, eh, OK, but sometimes I do wish I had waited for the iPhone 3G.

But my friend B and my cousin J have been pretty happy with their Blackberries, and I almost ended up with a Pearl instead. I actually forget why I didn't get one. I hope it was a good reason.

Nonetheless, knowing the buzz surrounding RIM's products, I'm kind of shocked that the newly released Storm is getting panned. Oh, sure, it's obvious that they bit off the iPhone with regards to the touch screen. Ironically, that was actually one of the things that made me hesitate about getting an iPhone. Even now, I'm not very good with the virtual keyboard on my iPod Touch, and I keep wishing I had physical keys to touch so I wouldn't have to be looking at the screen all the time. But knowing how popular their other phones are, I'm still kind of surprised.

Of course, this could all be early-adopter remorse. I've heard the iPhone 3Gs were notorious for having all sorts of irksome bugs, and maybe RIM has since released updated firmware since. But the keyboard interface actually sounds kind of atrocious. One of the things that I liked about moving from a numeric keypad to a QWERTY keyboard is that you didn't have to deal with the idiosyncrasies of predictive text. The Sure Type keyboard sounds like a step backward. And if you actually have to pound your touch screen in order for a key press to register, that would suck.

Back in May, John Gruber predicted that RIM was screwed. It's perhaps a little premature to say that he's right, but clearly the winds are starting to shift, if only for the fact that the on-going financial crisis is going to be one hell of a wind storm. Apparently it's a pain-in-the-ass to develop for RIM. How can they compete for developers against Apple's and Google's readily available SDKs? If the Storm really does tank, if Apple soon comes out with a more featureful iPhone, or if Google starts making more deals and getting more Android phones cranked out, the landscape is sure to change.

It's odd, but I'm starting to look at the disaster that has befallen Wall Street like something akin to a person having a stroke or a heart attack. Sure, the symptoms are most dramatic during the acute phase, and it's the acute phase that's probably going to kill you, but it's important to recognize that what a stroke or heart attack means is that this person has been sick for quite a while. Coronary artery disease and cerebrovascular disease don't just develop over the course of a week. It takes years of unhealthy living to get you to the point where you are vulnerable to having a heart attack or stroke.

And while we've got acute interventions like angioplasty, stenting, and emergent bypass, all that's going to really do is keep you from dying. It's not going to actually cure the root cause of your problem. Once you've had a heart attack or stroke, there's no going back to "normal." You had your chance at normal over the course of probably the last 25 years, but you never exercised and never fixed your diet. You can exercise and eat healthy now, after you've had your heart attack or stroke, but even that isn't going to actually reverse any of the damage that has already occurred.

So that's kind of where we're probably at with this Wall Street crisis. Years of deregulation, corruption, and sheer incomptence have rendered our economy vulnerable to disaster, and now the chickens have come home to roost. This $700 billion blank check is just like stenting or bypass. It's going to keep the economy alive. But it's not going to fix any of the damage that has already been done.

In other words, our economy is still going to collapse, unless we start restructuring it so that things are actually sustainable.

And just like how lots of cardiac patients generally fail to exercise or change their diet, I fear that most Americans are not going to change their spending habits or the way they manage risk.

A blog post critiques a throw-away line from "The Dark Knight":

It's interesting how the general public has such a different view of schizophrenia than what a health care professional is taught. I wonder if it's simply because psychiatry is such a relatively new field, and neuropsychobiology is practically brand new. Maybe schizophrenic just meant something entirely different than what it does now.

While people tend to remember the so-called "positive" symptoms of schizophrenia better—auditory hallucinations, delusions, so-called formal thought disordered—there are so-called "negative" symptoms as well that are basically lacunae of certain characteristics—apathy, non-emotionality, inability to relate normally to other people. These are evidence of executive dysfunction, a hallmark of many mental illnesses. The "negative" symptoms tend to make it less likely that a schizophrenic will actually turn violent and assault you.

If you're worried about being assaulted by a person with a mental illness, I'd watch out for little old ladies with Alzheimer's dementia. They can really throw a mean sucker punch, and woe to you if you didn't expect it. Another mental illness that makes a person more likely to assault you is mania. When lay people think of "crazy", I really think manics fit the stereotype better than schizophrenics do. About the only thing good about mania is that the prognosis is far better than in schizophrenia, dementia, or the personality disorders. Other potentially assaultive conditions include anti-social personality disorder (which is basically the updated name for what used to be known as a sociopath) and borderline personality disorder.

In an episode of synchronicity, I just finished Philip K Dick's novel Martian Time Slip, which, among other things, features an autistic child and another character who suffered some sort of psychotic break. Dick uses the term "schizophrenia" quite loosely, and I'm thinking he's confusing it with psychosis. It's perfectly possible to have a psychotic episode without being schizophrenic. It's actually more likely to happen in patients with major depression, since major depression is a lot more prevalent than schizophrenia. It's also possible to be schizophrenic without being psychotic, particularly if you're on some sort of neuroleptic medication. We can actually treat "positive" symptoms fairly well. Unfortunately, we haven't really figured out the "negative" symptoms. There is also a big, easy-to-tell difference between schizophrenic apathy and a cluster B personality disorder such as avoidant/dependent personalities who also have a hard time dealing with their environment and with other people. The former simply don't care, in the sense that they don't perceive the lack. The latter are acutely aware of their problem.

The House of God is a satirical novel written in the mid 1970's by a physician who goes by the pseudonym Samuel Shem. The book is about the experience of an intern physician trying to survive the rigors of the residency program associated with the mythical Best Medical School (a thinly veiled reference to Harvard.) The Fat Man is one of the senior residents in this residency program, and he came up with a set of rules that I find terrifyingly useful.

1. Gomers don't die
Now, what exactly is a gomer? Supposedly, the word was devised as an initialism that stands for "get out of my emergency room", which is what the ER physician will say when they find one of these folks hanging out in triage. It is actually a somewhat difficult word to define, although I think we eventually all get a sense of what a gomer is. At the VA, these are the guys who have (as we like to put in chart lore) multiple medical problems, which generally include COPD and PTSD. And despite having multiple types of cancer (lung cancer, prostate cancer, colon cancer, etc.) and despite having experience multiple cardiac arrests, they still manage to survive on.

Unfortunately, those episodes of cardiac arrest aren't inconsequential, and sometimes they leave their indelible hypoxic/ischemic mark on the brain. Hence, the O-sign, and the Q-sign.

At a county hospital, these kinds of gomers don't really show up much. For one thing, if you're forced to subsist on Medicaid, or if you're truly indigent, you probably won't survive to live this long anyway. Instead, what we have are "frequent fliers" who are generally younger than their VA counterparts, and who often have rather exotic chronic medical conditions owing to some heroic surgical measure (often performed when the patient was a pediatric patient) that did not really take into account the patient's social situation. Lots of kidney and liver transplant victims recipients. A lot of their problems are due to (1) the lack of available funding or (2) simple non-adherence to treatment regimens. Although I have to say, I've met a few frequent fliers who actually like coming back repeatedly to the hospital. (I suddenly think of a poor woman with lupus nephritis who was given a renal transplant that eventually failed because she didn't always take her meds. Her technique for getting into the hospital was to pop her colostomy bag so that stool would leak over her chronic panniculitis and lower extremity ulcers, immediately generating a fever. And then there was that creepy lady in the ER who shows up every week begging that a male resident perform her pelvic exam.)

2. Gomers go to ground
In this era with so many neuroleptic agents, this actually isn't as much of an issue (although it's still true.) Most patients can be kept in their beds without even so much as putting on a posey vest, much less four point restraints. All you need to do is order a little "Vitamin H" (haloperidol) or even some "Vitamin R" (risperidone) which is available as an orally disintegrating tablet, and your patient will remain blissfully staring at the ceiling for almost an entire nursing shift. That said, it is well to bear in mind that you can indeed raise the bed to different levels, and if you forget to reset it after performing whatever procedure you had to perform, you may well be guilty of malpractice. I am still amused by Shem's discursus into the different levels you can adjust the bed: the orthopedic height (where if your patient falls, they'll break a hip), the neurosurgical height (where if your patient falls, they'll have an epidural or subdural hematoma), and the pathologic height (where if your patient falls, they're going to get transferred to the morgue.)

3. At a cardiac arrest, the first procedure is to take your own pulse
While TV depicts every "code blue" as a traumatic, high-stress situations, in reality, they rarely are. Some care units will call a code blue if your patient sneezes funny or actually does start breathing at 20 breaths per minute. A lot of times, they get called because someone got a little too aggressive with the opiates or benzodiazepines, and the respiratory tech had to manually bag your patient for a little while until the narc wore off, or until someone administered the antidote. But many times, it's because the patient never really had a chance in the first place, you knew they were going to code when you admitted them but forgot to or were too chickenshit to ask the family to change the code status, and now all you're really doing is breaking their ribs, and pissing away thousands of dollars of life-saving pharmaceutical agents before you package them up for the transfer on the metal gurney to the basement. Whatever the case, remember what the Hitchhiker's Guide to the Galaxy advises, and "Don't Panic!" There is no situation that yelling will not fail to make worse.

4. The patient is the one with the disease
There are actually a lot of ways to interpret this one. The obvious one is that you should never so strongly self-identify with the patient that you get emotionally wrapped up in their suffering. As my ethics instructor was wont to remind us, there is a great difference between empathy, and sympathy. And once you lose your ability to rationally assess your patient, the probability that you'll do more harm than good increases.

The other thing is that remember, you're not the one who's sick. While being on-call for 30 hours without sleep actually mimicks being sick fairly well (even to the point that you'll probably develop a slight leukocytosis with neutrophil predominance), your chances of ending up in the morgue are a probably still a lot less than your patient's chances. (Unless you drive home delirious from sleep-deprivation and get hit by a Mack truck, which is certainly possible, but I digress.)

The more subtle interpretation is the whole "treat the patient, not the numbers" mantra. Just because the monitor says ventricular fibrillation does not make it so.

5. Placement comes first
Thankfully, most training hospitals now come well-equipped with discharge planners who have the infinite patience to deal with the circuitous madness known as health insurance coverage. If there's funding, there's always somewhere you can send them. But it is a little disturbing to have to write the order: "discharge to street."

6. There is no body cavity that cannot be reached with a 14-gauge needle and a good strong arm.
Disturbingly, this is probably now very less true because of the awful obesity epidemic going on in this country. There are places in certain patients where even an angiocath cannot go, no matter how many people you have holding up the pannus. And while we now have interventional radiologists who can wield said 14-gauge needles with impunity because at least they don't have to go in blind, most IR suite tables can't handle people who are over 350 lbs, 450 lbs max.

7. Age + BUN = Lasix dose.
I used to think this was a pure joke until I realized that the correct way to dose any drug is to determine their kidney function through their glomerular filtration rate (GFR). Since direct measurement of the GFR is tedious, it's way more expedient to estimate it. A popular way to estimate it is by using the Cockcroft-Gault equation:

estimate creatinine clearance = (140-Age) x mass x 0.85 if female/72 x serum creatinine.

An even more complicated formula is the MDRD formula

estimated GFR = 186 x serum creatinine^-1.154 x Age^-0.203 x 1.21 if black x 0.742 if female

hilariously, BUN does factor in a more elaborate version of the MDRD formula

estimated GFR = 170 x serum creatinine^-0.999 x Age^-0.176 x 0.762 if female x 1.180 if black x BUN^-0.170 x albumin^0.318

8. They can always hurt you more.
Essentially a more succinct version of Murphy's Law. Just when you think it can't get any worse, it almost certainly will. My senior resident when I was an intern put a positive spin on it, though: They can always hurt you more, but they can't stop the clock. There's something reassuring about knowing that you do get to sign-out when 1 pm rolls around post-call, that a rotation won't last more than 28 days, and that an internal medicine residency program lasts only 3 years. Unfortunately, the rest of your career doesn't have quite as well-defined end-points, and the only real way out is retirement or death.

9. The only good admission is a dead admission.
I never bought this one, because even if they come in dead, you still have to write an history-and-physical and a discharge summary. About the only good thing is that (1) the admission will still count towards your admission cap and your census cap and (2) at least you know where the patient is going to be discharged to.
10. If you don't take a temperature, you can't find a fever.
This is basically a core internal medicine superstition. Do not order tests on things that you don't need to know, because you will find something abnormal, and you'll end up going down some god-awful circuitous route that mandates invasive testing and probably surgery that will probably lead to your patient's death, when all they came in with was a simple case of pneumonia, and all they really needed was an overnight stay for some IV fluids and antibiotics, and they would've lived if you didn't fuck with them.

11. Show me a medical student who only triples my work, and I will kiss his feet.
It really all depends on the medical school and the medical student. The medical school I went to basically threw us to the wolves and lots of us practiced medicine without a license. The medical school associated with my residency program barely let their med students do anything. The fact of the matter is that most medical students don't know enough to understand the previous rule, and this will almost certainly increase your work load by an order of magnitude, mostly to prevent the proceduralists and the surgeons from trying to stick things into/take things out of your patient. What is particularly grueling is working with the so-called sub-intern, who is supposed to function like an intern, except that he/she can't actually write orders. I would like to know whose bright idea this was. It was probably Willian Henry Osler's, while he was high on amphetamines.

12. If the radiology resident and the medical student both see a lesion on the chest x-ray, there can be no lesion there.
Don't get me wrong. Without the radiology department, I would've killed far many more people than was necessary, but sometimes it all seems so futile when they dictate "significant of this finding is unclear, recommend clinical correlation.

13. The delivery of good medical care is to do as much nothing as possible.
The fact of the matter is that the human body is a whole hell of a lot better at fixing itself than any of our poisonous medications and crude surgical tools, and a sure sign that you're doing something wrong is that the number of consults and invasive procedures you're ordering are rapidly increasing, even as your patient's lab values and vital signs all drift toward abnormality.

I don't know why only this particular invocation worked for me, but I finally figured out how to share my desktop screen over the Internet

My desktop machine at home is a Mac Mini 1.2 GHz PPC. I'm trying to connect with a MacBook 2.4 GHz Core 2 Duo.

MacBook% ssh -f -C -L 5901:localhost:5900 username@macmini.dynamic-dns-service.org 'sleep 5'; open vnc://localhost:5901

I found this on Extra Pepperoni.

5901 can be any port you want on your client machine. 5900 is where Screen Sharing/VNC lives. Everything else is pretty self-explanatory, I think.

Of course, viewing a 1680 x 1050 screen on an 1280 x 800 screen is probably not recommended.

 

Interesting that KDE appears to be going through what GNOME went through back in the 1.4 > 2.0 transition. There's all sorts of bitching and moaning going on about the recent KDE 4.1 release. Some writers have even employed the hyperbolic description of a civil war.

I still remember those days when criticisms like these would immediately get blasted by "write code or STFU!" Which is probably appropriate. KDE is not some Microsoftian abomination with a ton of awful legacy crap attached to it that you're plunking down several hundred dollars to license. It's free. No one is forcing you to use it. Keep using KDE 3 if you like.

In an open source environment, the idea that the users don't matter as much as the developers really comes to fore. This is sort-of true even in proprietary environments. It's more important to keep the developers happy, because the people who write the actual code are the people who will keep the platform alive. It's the apps that matter. If developers don't write apps for the platform, the platform will die. Notice that no one writes GNOME 1.4/GTK 1.2 apps anymore. Someday, KDE 4.x-only apps are likely to appear.

---

When I started using Linux exclusively in 1999, KDE had the larger installed base. It seemed more featureful and more stable than GNOME at the time. But it had the bugaboo of a non-GPL'ed license attached to it, and so I stuck with GNOME, which was quite an awful piece of crap at the time. (Although it was still less frustrating than Windows 98!) Trolltech eventually saw the light and released Qt as GPL, allowing KDE to shed the non-GPL stigma. Even when GNOME 2.0 came out, KDE still seemed to have the lead. They also had a much better file manager (Konqueror vs Nautilus, which always, always crashed for me) But inertia is strong, and I stayed faithful to GNOME. For one thing, there were a lot of neat GTK apps, especially on the mail client and web browser front. And, as a testament to the awesomeness of Open Source, there was nothing to stop me from running KDE apps on GNOME.

Eventually, though, I was seduced by the candyliciousness of Mac OS X. But don't get me wrong. Remember that Apple is a hardware company, and I bought a Mac because I like the hardware. Most of the apps I use are still Open Source (Safari/Webkit, Camino, Firefox, Seashore, Transmission, Abiword, Smultron, Aquamacs, just to name the ones sitting in my dock) But the GNOME vs KDE saga receded to the background. I still keep X.org sitting around my hard drive, although I don't launch it very often.

Still it'll be interesting to see if GNOME will ever be considered the more user-friendly, more stable desktop environment. I'm sure that they got a boost when Sun decided to adopt it as their CDE replacement (JDS, the Java Desktop System, is basically a rebranded and customized version of GNOME.) And GNOME 2.22 sounds like it's pretty solid.

While there have been talks about forking KDE, it's probably not going to happen if the main proponents are non-coding users.

Date	Event	Magnitude	Damage	Deaths
1 Oct 1987	Whittier Narrows	5.9	$380 million	8	My first major earthquake; 5.6 magnitude aftershock on 4 Oct 1987
29 Jun 1991	Sierra Madre	5.6	$40 million	2	epicenter 12 miles from my parents' house. I almost forgot about this one because it didn't cause a lot of damage and didn't have a lot of aftershocks.
17 Jan 1994	Northridge	6.7	$20 billion	60	7,000 injuries, destroyed Golden State Fwy/Antelope Valley Fwy interchange (I-5/CA-14) and a large segment of the Santa Monica Fwy (I-10)

Now don't get me wrong. I'm a big fan of social networks and microblogging/nanoblogging, but I'm not really ready to buy into the hype about the obliteration of traditional media.

The Chengdu/Sichuan quake

in May was ballyhooed as a triumph of Twitter over traditional media, and in fact there was a significant time delay before stories from traditional channels finally hit. But this is more a commentary on the limitations of the press in a repressive Communist regime that has a stranglehold on information, not to mention a fact that the epicenter was in a relatively remote rural region almost 100 miles away from Chengdu.

The case is significantly different when you're dealing with a major metropolitan area in an industrialized country under a regime that allows relative freedom of the press. The local news happened to be on-the-air when the Chino Hills quake hit. They got in touch with Cal Tech seismologists relatively rapidly. In contrast, the first reports on Twitter localized it wrong (San Diego is over 100 miles away from the epicenter) and most of the chatter was simply reiterating what the local news was reporting.

---

Not to say that Twitter didn't have any value. Certainly, since the cel phone network got overwhelmed by panicked callers and since land line service was disrupted by the quake, the only avenues of communication remaining were SMS and the Internet (thank you, ARPAnet!) But it's really more of an adjunct than a replacement to traditional media, and there's a lot of noise and misinformation.