1619: Watson Medical Algorithm

This forum is for the individual discussion thread that goes with each new comic.

Moderators: Moderators General, Prelates, Magistrates

niauropsaka
Posts: 85
Joined: Mon Apr 22, 2013 8:50 pm UTC

1619: Watson Medical Algorithm

Postby niauropsaka » Mon Dec 21, 2015 7:05 am UTC

Image
Title text: "Due to a minor glitch, 'discharge patient' does not cause the algorithm to exit, but instead leads back to 'hunt down and capture patient'."

Well, I laughed enough to feel it deep in my lungs. Thanks for the cardiovascular workout, Randall.

User avatar
Eternal Density
Posts: 5547
Joined: Thu Oct 02, 2008 12:37 am UTC
Contact:

Re: 1619: Watson Medical Algorithm

Postby Eternal Density » Mon Dec 21, 2015 7:23 am UTC

This chart seems to be incorrect. I found this to be rather banal, rather than amusing.
Play the game of Time! castle.chirpingmustard.com Hotdog Vending Supplier But what is this?
In the Marvel vs. DC film-making war, we're all winners.

User avatar
rhomboidal
Posts: 788
Joined: Wed Jun 15, 2011 5:25 pm UTC
Contact:

Re: 1619: Watson Medical Algorithm

Postby rhomboidal » Mon Dec 21, 2015 7:35 am UTC

For "Comfort Patient," Watson selects a random knock-knock joke.

If the patient fails to respond, surgical extraction of laughter is licensed.

User avatar
StClair
Posts: 404
Joined: Fri Feb 29, 2008 8:07 am UTC

Re: 1619: Watson Medical Algorithm

Postby StClair » Mon Dec 21, 2015 7:42 am UTC

"Surgically adjust patient to match" (standard height/weight chart) - ah, Procrustes' sign.

Ordinalade
Posts: 4
Joined: Thu Nov 19, 2015 10:20 pm UTC

Re: 1619: Watson Medical Algorithm

Postby Ordinalade » Mon Dec 21, 2015 7:53 am UTC

Man, Watson's algorithm gives away the secret for Laser Eye!

User avatar
da Doctah
Posts: 893
Joined: Fri Feb 03, 2012 6:27 am UTC

Re: 1619: Watson Medical Algorithm

Postby da Doctah » Mon Dec 21, 2015 7:58 am UTC

I see one glaring omission. Nowhere does the chart mention checking to see if the patient has health insurance. (Yes: proceed; No: shove him out into the street.)

I tried following my own present condition through the flow chart, and got stuck after "remove and inspect skeleton". There are branches for too few bones and for too many bones, but no branch for exactly the right number.

User avatar
keithl
Posts: 637
Joined: Mon Aug 01, 2011 3:46 pm UTC

Re: 1619: Watson Medical Algorithm

Postby keithl » Mon Dec 21, 2015 8:09 am UTC

It's uncanny - this is precisely the procedure described in the upcoming 20th edition of "Harrison's Principles of Internal Medicine", the standard reference for clinical practice. I just downloaded the galleys from McGraw Hill's experimental new document management system, also from IBM research.

User avatar
Neil_Boekend
Posts: 3220
Joined: Fri Mar 01, 2013 6:35 am UTC
Location: Yes.

Re: 1619: Watson Medical Algorithm

Postby Neil_Boekend » Mon Dec 21, 2015 9:08 am UTC

StClair wrote:"Surgically adjust patient to match" (standard height/weight chart) - ah, Procrustes' sign.

One size fits all, including knife to cut off the parts that won't fit.
Mikeski wrote:A "What If" update is never late. Nor is it early. It is posted precisely when it should be.

patzer's signature wrote:
flicky1991 wrote:I'm being quoted too much!

he/him/his

User avatar
Reecer6
Posts: 100
Joined: Fri Nov 30, 2012 12:59 am UTC

Re: 1619: Watson Medical Algorithm

Postby Reecer6 » Mon Dec 21, 2015 9:25 am UTC

Of course, the medical industry ignores the existence of those with Vitamin D that's just okay, like maybe a 6/10. We truly live in a dystopia.

Unrelatedly, I've had my eyes surgically removed the last 5 times I've been to the hospital and I have no idea what to say to make Watson stop doing that but I can't read this chart because my eyes have been somehow surgically removed 5 times.

Hiferator
Posts: 81
Joined: Fri Feb 15, 2013 8:23 am UTC

Re: 1619: Watson Medical Algorithm

Postby Hiferator » Mon Dec 21, 2015 9:33 am UTC

Wait, we remove extra limbs if there's FEWER than 100?

LockeZ
Posts: 51
Joined: Mon Jan 19, 2009 8:30 am UTC

Re: 1619: Watson Medical Algorithm

Postby LockeZ » Mon Dec 21, 2015 10:18 am UTC

Eh, I've had worse doctors.

da Doctah wrote:I tried following my own present condition through the flow chart, and got stuck after "remove and inspect skeleton". There are branches for too few bones and for too many bones, but no branch for exactly the right number.

Kind of an edge case, probably not worth worrying about. Just knock one of your teeth out and try again.

User avatar
Envelope Generator
Posts: 582
Joined: Sat Mar 03, 2012 8:07 am UTC
Location: pareidolia

Re: 1619: Watson Medical Algorithm

Postby Envelope Generator » Mon Dec 21, 2015 11:25 am UTC

da Doctah wrote:I see one glaring omission. Nowhere does the chart mention checking to see if the patient has health insurance. (Yes: proceed; No: shove him out into the street.)

I tried following my own present condition through the flow chart, and got stuck after "remove and inspect skeleton". There are branches for too few bones and for too many bones, but no branch for exactly the right number.


The only solution seems to be to hold your breath for a couple of minutes before starting the flowchart.
I'm going to step off the LEM now... here we are, Pismo Beach and all the clams we can eat

eSOANEM wrote:If Fonzie's on the order of 100 zeptokelvin, I think he has bigger problems than difracting through doors.

User avatar
orthogon
Posts: 2941
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1619: Watson Medical Algorithm

Postby orthogon » Mon Dec 21, 2015 11:39 am UTC

LockeZ wrote:Eh, I've had worse doctors.

da Doctah wrote:I tried following my own present condition through the flow chart, and got stuck after "remove and inspect skeleton". There are branches for too few bones and for too many bones, but no branch for exactly the right number.

Kind of an edge case, probably not worth worrying about. Just knock one of your teeth out and try again.

Teeth are obviously not bones. But is a boner a bone?
xtifr wrote:... and orthogon merely sounds undecided.

User avatar
Echo244
Posts: 511
Joined: Wed May 20, 2015 9:49 am UTC
Location: Ping! Ping! Ping! Ping!

Re: 1619: Watson Medical Algorithm

Postby Echo244 » Mon Dec 21, 2015 11:43 am UTC

No.

Anyway. I've seen worse decision trees. It's entertaining, though, and captures the spirit of craziness that you get when you start looking to optimise processes.

I love that the surgical adjustment of patient height and weight occurs before any decision can be made. We will be standardised!
Unstoppable force of nature. That means she/her/hers.
Has committed an act of treason.

User avatar
orthogon
Posts: 2941
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1619: Watson Medical Algorithm

Postby orthogon » Mon Dec 21, 2015 11:57 am UTC

I went to the 2014 BCS/IET Turing Lecture which was given by Bernard S. Meyerson of IBM. It's an interesting lecture, and you can watch it online. Although it fell somewhat outside the scope of the talk, he spoke briefly about Watson. Although it seems to be quite clever, his explanation of how it works wasn't particularly enlightening. In particular he put a lot of emphasis on the way that it learns, rather than being pre-programmed with all the relevant rules up front; but he presented this as though it was a new thing, whereas as far as I understand it, that's been the default approach in AI since at least the '80s. In a way that's part of the joke of this comic: the idea that a cutting-edge AI would be based on old-school expert-system-style flowcharts.
xtifr wrote:... and orthogon merely sounds undecided.

User avatar
Angua
Don't call her Delphine.
Posts: 5774
Joined: Tue Sep 16, 2008 12:42 pm UTC
Location: UK/[St. Kitts and] Nevis Occasionally, I migrate to the US for a bit

Re: 1619: Watson Medical Algorithm

Postby Angua » Mon Dec 21, 2015 11:58 am UTC

orthogon wrote:
LockeZ wrote:Eh, I've had worse doctors.

da Doctah wrote:I tried following my own present condition through the flow chart, and got stuck after "remove and inspect skeleton". There are branches for too few bones and for too many bones, but no branch for exactly the right number.

Kind of an edge case, probably not worth worrying about. Just knock one of your teeth out and try again.

Teeth are obviously not bones. But is a boner a bone?

Not in humans. However, you can still fracture it.
Crabtree's bludgeon: “no set of mutually inconsistent observations can exist for which some human intellect cannot conceive a coherent explanation, however complicated”
GNU Terry Pratchett

User avatar
cellocgw
Posts: 1917
Joined: Sat Jun 21, 2008 7:40 pm UTC

Re: 1619: Watson Medical Algorithm

Postby cellocgw » Mon Dec 21, 2015 12:44 pm UTC

Eternal Density wrote:This chart seems to be incorrect. I found this to be rather banal, rather than amusing.


You must be one of those pre-med weenies who spends the whole weekend studying in the library and never has a lick of fun in college. :twisted:

As a wise friend of mine once said, "Computers do exactly what you tell them to do." So blame the Medical Cognoscenti Starchamber Inquisition Conspiracy for the chart, not the poor computer.
https://app.box.com/witthoftresume
Former OTTer
Vote cellocgw for President 2020. #ScienceintheWhiteHouse http://cellocgw.wordpress.com
"The Planck length is 3.81779e-33 picas." -- keithl
" Earth weighs almost exactly π milliJupiters" -- what-if #146, note 7

User avatar
Neil_Boekend
Posts: 3220
Joined: Fri Mar 01, 2013 6:35 am UTC
Location: Yes.

Re: 1619: Watson Medical Algorithm

Postby Neil_Boekend » Mon Dec 21, 2015 12:55 pm UTC

cellocgw wrote:As a wise friend of mine once said, "Computers do exactly what you tell them to do.".
Only if you use "the cooperation between you and the programmers that made the software on the computer". If it did exactly what I told it I wouldn't have half the trouble I have with computers. Half the time it's a piece of software that makes incorrect decisions, because the programmer made incorrect assumptions of my needs or didn't consider my use case, a piece of software that is making a billion calculations that it doesn't need or in some rare cases an actual bug.

In most cases the computer is NOT doing what I told it to do. It is doing what the programmer told it to do. In most of those cases that works fine, but sometimes it doesn't.
Mikeski wrote:A "What If" update is never late. Nor is it early. It is posted precisely when it should be.

patzer's signature wrote:
flicky1991 wrote:I'm being quoted too much!

he/him/his

User avatar
orthogon
Posts: 2941
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1619: Watson Medical Algorithm

Postby orthogon » Mon Dec 21, 2015 1:30 pm UTC

Neil_Boekend wrote:If it did exactly what I told it I wouldn't have half the trouble I have with computers.

This. The most infuriating issues are the ones where you find yourself looking for the "--actually-do-it" command-line switch or the "Ignore my requests" checkbox (selected by default).
xtifr wrote:... and orthogon merely sounds undecided.

speising
Posts: 2267
Joined: Mon Sep 03, 2012 4:54 pm UTC
Location: wien

Re: 1619: Watson Medical Algorithm

Postby speising » Mon Dec 21, 2015 1:57 pm UTC

That's what DWIM (Do What I Mean) systems are for.

User avatar
orthogon
Posts: 2941
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1619: Watson Medical Algorithm

Postby orthogon » Mon Dec 21, 2015 2:51 pm UTC

speising wrote:That's what DWIM (Do What I Mean) systems are for.

That's a stage removed from what we're talking about. To rephrase Neil's point: computers only really do what you tell them at the level of machine-code instructions. Compilers are generally pretty good, so the same sort-of applies at the level of source code. But even at those levels, they only do what you tell them if "doing what you tell them" is defined as the the normal behaviour plus all the corner cases, exceptions and errors that might occur, but then that's almost a tautology. When you say "ADD R1, R2, R3", you're not telling it to add register 1 and register 2 and put the result in register 3; you're telling it to try to do that, but wrap around modulo 232 if the result doesn't fit, and set some flags, and so on. All you can really say for sure is that "ADD R1, R2, R3" will do what "ADD R1, R2, R3" does, which is circular. And if you select "copy to clipboard" from a menu, there's absolutely no guarantee that it will copy anything anywhere. It's the same basic problem, but much more involved and with orders of magnitude more reasons for it not to do the thing that was implied by the text of the menu option.
xtifr wrote:... and orthogon merely sounds undecided.

User avatar
DennyMo
Posts: 46
Joined: Fri Apr 23, 2010 10:50 am UTC

Re: 1619: Watson Medical Algorithm

Postby DennyMo » Mon Dec 21, 2015 3:22 pm UTC

Go home, Watson; you're drunk.

rmsgrey
Posts: 3412
Joined: Wed Nov 16, 2011 6:35 pm UTC

Re: 1619: Watson Medical Algorithm

Postby rmsgrey » Mon Dec 21, 2015 4:26 pm UTC

Hiferator wrote:Wait, we remove extra limbs if there's FEWER than 100?


A) If there's more than that, it would take too long
B) 1002?
C) Yeah, sometimes branches get switched.

User avatar
Flumble
Yes Man
Posts: 2024
Joined: Sun Aug 05, 2012 9:35 pm UTC

Re: 1619: Watson Medical Algorithm

Postby Flumble » Mon Dec 21, 2015 4:40 pm UTC

rmsgrey wrote:
Hiferator wrote:Wait, we remove extra limbs if there's FEWER than 100?


A) If there's more than that, it would take too long
B) 1002?
C) Yeah, sometimes branches get switched.

It's probably C: it's just part of Watson's genetic learning algorithm.

And with genetic I mean using actual genes from previous patients, chopped up at certain locations and then translated to machine code.

ps.02
Posts: 378
Joined: Fri Apr 05, 2013 8:02 pm UTC

Re: 1619: Watson Medical Algorithm

Postby ps.02 » Mon Dec 21, 2015 7:19 pm UTC

Very Kafkaesque. I like how recording the patient's name is lower priority than getting a blood sample. Or ... well, I assumed it was a blood sample ... but considering the rest of the chart, it seems more likely that the procedure is to extract all the blood. This would, after all, make several of the subsquent steps easier.

Also, yay for non sequiturs like the little nod to autoconf.

xtifr
Posts: 327
Joined: Wed Oct 01, 2008 6:38 pm UTC

Re: 1619: Watson Medical Algorithm

Postby xtifr » Tue Dec 22, 2015 2:55 am UTC

Neil_Boekend wrote:
cellocgw wrote:As a wise friend of mine once said, "Computers do exactly what you tell them to do.".
Only if you use "the cooperation between you and the programmers that made the software on the computer". If it did exactly what I told it I wouldn't have half the trouble I have with computers.

The problem is that you're when you run someone else's software, you're asking someone else to tell your computer what to do. So you shouldn't be surprised when the results don't match what you want.

What you're really saying is that you don't actually know how to tell the computer how to do something, and the people who try to fill that role in your life are unreliable. I suspect most of us would agree with that second part. :)

Of course, the claim "computers do what you tell them" depends on having a working computer. If you have a memory chip that's glitching, and occasionally dropping bits, it may end up doing something entirely different from what it was told to do, no matter who actually did the telling.
"[T]he author has followed the usual practice of contemporary books on graph theory, namely to use words that are similar but not identical to the terms used in other books on graph theory."
-- Donald Knuth, The Art of Computer Programming, Vol I, 3rd ed.

User avatar
Neil_Boekend
Posts: 3220
Joined: Fri Mar 01, 2013 6:35 am UTC
Location: Yes.

Re: 1619: Watson Medical Algorithm

Postby Neil_Boekend » Tue Dec 22, 2015 8:10 am UTC

xtifr wrote:
Neil_Boekend wrote:
cellocgw wrote:As a wise friend of mine once said, "Computers do exactly what you tell them to do.".
Only if you use "the cooperation between you and the programmers that made the software on the computer". If it did exactly what I told it I wouldn't have half the trouble I have with computers.

The problem is that you're when you run someone else's software, you're asking someone else to tell your computer what to do. So you shouldn't be surprised when the results don't match what you want.

What you're really saying is that you don't actually know how to tell the computer how to do something, and the people who try to fill that role in your life are unreliable. I suspect most of us would agree with that second part. :)

Of course, the claim "computers do what you tell them" depends on having a working computer. If you have a memory chip that's glitching, and occasionally dropping bits, it may end up doing something entirely different from what it was told to do, no matter who actually did the telling.

I am not an experienced programmer and even if I was I wouldn't be able to program an OS on my own. I choose to use general software although it makes my computer not always do what I tell it to. It is a rather simple choice, really.
Computers haven't done what a single person told it for years. The software has gotten too complex. All the commands get filtered and changed by choices that the programmers of the OS and the other software made. This has gotten so complex that IMHO you can't say that the computer does exactly what I tell it to.
Note that I don't think this is inherently bad. Computers can do so much more than before the layers of complexity were piled on, despite me disagreeing with many of the choices (M$ and their freaaking autoplay, hiding of file extensions and whatnot (yes I know how to turn those two "features" off)).
Mikeski wrote:A "What If" update is never late. Nor is it early. It is posted precisely when it should be.

patzer's signature wrote:
flicky1991 wrote:I'm being quoted too much!

he/him/his

User avatar
Pfhorrest
Posts: 4807
Joined: Fri Oct 30, 2009 6:11 am UTC
Contact:

Re: 1619: Watson Medical Algorithm

Postby Pfhorrest » Tue Dec 22, 2015 9:04 am UTC

It's true that your computer does a complicated function of what you and other people have told it to do, and not strictly exactly what you told it to do. But you really wouldn't want it to do exactly what you told it to do, and only what you told it to do; a large function of delegating a task rather than doing it yourself is trusting that the thing or person you delegate it to knows how to do the task better than you, and will do something that you will like more than that what you would have done yourself, or what you would have told it/them to do if you had to spell it out for them.

Imagine you hire, say, a plumber, assuming you're not a plumber yourself. You tell them some things about what you don't like about your plumbing and the kind of thing you would prefer. This is really terribly incomplete information, not at all even a full description of what you would like your plumbing to do, much less instructions on how to make it do that. You're relying on someone else (or many someones else, over possibly decades) having told the plumber both some assumptions he should make in various circumstances about what plumbing behavior will best please customers, and how to transform plumbing that differs from that state into that state; and how much input to seek from the customers to better tailor their behavior to the circumstances, too. Sometimes, a plumber will and should ignore what you tell them to do, because they know better and they really shouldn't do that because it will result in something you're really unhappy with even though you technically asked for it. Of course, sometimes a plumber might also ignore what you tell them to do when it's a simple matter of your preference with no non-obvious problems that must be avoided for your own good. The circumstances in which plumbers listen to you or solicit your input may differ and be superior or inferior in quality, depending on their training, but we certainly don't want plumbers to all be trained to do all and only exactly what we tell them to, because half the reason we pay them is to know better than us (because they were trained by someone, or many someones, who cumulatively know better than most of us individually).

Same is true for computers. The circumstances in which computers listen to you or solicit your input may differ and be superior or inferior in quality, depending on their programming, but we certainly don't want computers to all be programmed to do all and only exactly what we tell them to, because half the reason we use them is to know better than us (because they were programmed by someone, or many someones, who cumulatively know better than most of us individually).
Forrest Cameranesi, Geek of All Trades
"I am Sam. Sam I am. I do not like trolls, flames, or spam."
The Codex Quaerendae (my philosophy) - The Chronicles of Quelouva (my fiction)

User avatar
Neil_Boekend
Posts: 3220
Joined: Fri Mar 01, 2013 6:35 am UTC
Location: Yes.

Re: 1619: Watson Medical Algorithm

Postby Neil_Boekend » Tue Dec 22, 2015 9:33 am UTC

Exactly.
Mikeski wrote:A "What If" update is never late. Nor is it early. It is posted precisely when it should be.

patzer's signature wrote:
flicky1991 wrote:I'm being quoted too much!

he/him/his

Tova
Posts: 100
Joined: Wed Mar 24, 2010 3:44 am UTC

Re: 1619: Watson Medical Algorithm

Postby Tova » Tue Dec 22, 2015 10:09 am UTC

Eternal Density wrote:This chart seems to be incorrect. I found this to be rather banal, rather than amusing.


Same. It's the first time I gave up reading an xkcd comic before getting to the end.

xtifr
Posts: 327
Joined: Wed Oct 01, 2008 6:38 pm UTC

Re: 1619: Watson Medical Algorithm

Postby xtifr » Tue Dec 22, 2015 11:58 am UTC

Neil_Boekend wrote:I am not an experienced programmer and even if I was I wouldn't be able to program an OS on my own.

And I'm not saying you should. But if you don't, you don't get to complain about how the computer doesn't do what you told it to—because you didn't tell it what to do. You delegated that task to someone else. You can complain that it doesn't do what you want, and you can complain that the people who wrote the programs you use are idiots. But the computer is not at fault. It does what it's told. Every single freaking time. Whether that's what you wanted it to do or not.

The quote you originally responded to was incomplete. The full quote is "the computer does exactly what you tell it to do—not what you want it to do!" It was intended as a reminder to programmers, not a promise to hopeful users. :)
"[T]he author has followed the usual practice of contemporary books on graph theory, namely to use words that are similar but not identical to the terms used in other books on graph theory."
-- Donald Knuth, The Art of Computer Programming, Vol I, 3rd ed.

User avatar
Neil_Boekend
Posts: 3220
Joined: Fri Mar 01, 2013 6:35 am UTC
Location: Yes.

Re: 1619: Watson Medical Algorithm

Postby Neil_Boekend » Tue Dec 22, 2015 12:41 pm UTC

xtifr wrote:
Neil_Boekend wrote:I am not an experienced programmer and even if I was I wouldn't be able to program an OS on my own.

And I'm not saying you should. But if you don't, you don't get to complain about how the computer doesn't do what you told it to—because you didn't tell it what to do. You delegated that task to someone else. You can complain that it doesn't do what you want, and you can complain that the people who wrote the programs you use are idiots. But the computer is not at fault. It does what it's told. Every single freaking time. Whether that's what you wanted it to do or not.

The quote you originally responded to was incomplete. The full quote is "the computer does exactly what you tell it to do—not what you want it to do!" It was intended as a reminder to programmers, not a promise to hopeful users. :)

I know the quote. I just disagree.
And I am not complaining. I think it is a reasonable trade. The current complexity of software gives rise to awesome features and I wouldn't want to go back to a commodore 64. See the following quote:
Neil_Boekend wrote:<snip>Note that I don't think this is inherently bad. Computers can do so much more than before the layers of complexity were piled on,<snip>
Mikeski wrote:A "What If" update is never late. Nor is it early. It is posted precisely when it should be.

patzer's signature wrote:
flicky1991 wrote:I'm being quoted too much!

he/him/his

User avatar
pkcommando
Posts: 535
Joined: Fri Feb 10, 2012 3:22 pm UTC
Location: Allston, MA

Re: 1619: Watson Medical Algorithm

Postby pkcommando » Tue Dec 22, 2015 1:25 pm UTC

In the Big Hero 6 sequel, Watson reprograms Baymax with this algorithm, hilarity ensues.
"The Universe is for raptors now!" say Raptors, as they take over all of Universe.

User avatar
orthogon
Posts: 2941
Joined: Thu May 17, 2012 7:52 am UTC
Location: The Airy 1830 ellipsoid

Re: 1619: Watson Medical Algorithm

Postby orthogon » Tue Dec 22, 2015 1:42 pm UTC

Neil_Boekend wrote:And I am not complaining. I think it is a reasonable trade. The current complexity of software gives rise to awesome features and I wouldn't want to go back to a commodore 64.

I'm not so sure. What I would like to go back to is 1980s levels of responsiveness. Electronic stuff used to respond in the time it took a few TTL gates to flip; these days I simply cannot comprehend what the machine is doing during the several billion cycles that elapse between pressing, say, a key on a ticket machine touchscreen, and it moving onto the next step.
xtifr wrote:... and orthogon merely sounds undecided.

MuEta
Posts: 15
Joined: Tue Dec 22, 2015 2:44 pm UTC

Re: 1619: Watson Medical Algorithm

Postby MuEta » Tue Dec 22, 2015 2:47 pm UTC

Does anyone know how I could get a larger version of this comic? My dad (a physician) wants to post a poster of it in one of his exam rooms. He also discussed highlighting certain steps, to scare patients. If you have any advice for what to highlight, feel free to provide suggestions.

User avatar
Flumble
Yes Man
Posts: 2024
Joined: Sun Aug 05, 2012 9:35 pm UTC

Re: 1619: Watson Medical Algorithm

Postby Flumble » Tue Dec 22, 2015 3:02 pm UTC

orthogon wrote:
Neil_Boekend wrote:And I am not complaining. I think it is a reasonable trade. The current complexity of software gives rise to awesome features and I wouldn't want to go back to a commodore 64.

I'm not so sure. What I would like to go back to is 1980s levels of responsiveness. Electronic stuff used to respond in the time it took a few TTL gates to flip; these days I simply cannot comprehend what the machine is doing during the several billion cycles that elapse between pressing, say, a key on a ticket machine touchscreen, and it moving onto the next step.

Me too. I can't fathom how people are fine with an application (or worse, the OS) hanging for a second when they press a button.
Even my phone's default apps (most notably the app store c.q. updater) have blocking views that first have to load <something big> before anything is rendered and touch events are handled (and yes, they are all queued, so don't touch anything before it's loaded).

Quey
Posts: 28
Joined: Wed Sep 24, 2014 12:05 am UTC

Re: 1619: Watson Medical Algorithm

Postby Quey » Tue Dec 22, 2015 3:35 pm UTC

Flumble wrote:
orthogon wrote:
Neil_Boekend wrote:And I am not complaining. I think it is a reasonable trade. The current complexity of software gives rise to awesome features and I wouldn't want to go back to a commodore 64.

I'm not so sure. What I would like to go back to is 1980s levels of responsiveness. Electronic stuff used to respond in the time it took a few TTL gates to flip; these days I simply cannot comprehend what the machine is doing during the several billion cycles that elapse between pressing, say, a key on a ticket machine touchscreen, and it moving onto the next step.

Me too. I can't fathom how people are fine with an application (or worse, the OS) hanging for a second when they press a button.
Even my phone's default apps (most notably the app store c.q. updater) have blocking views that first have to load <something big> before anything is rendered and touch events are handled (and yes, they are all queued, so don't touch anything before it's loaded).


A lot of that is probably network access, which can take a lot longer than internal gate flipping. This is especially true for smartphones, which are expected to deliver a high standard of experience with a tenth of the hardware, in broad terms. And while there is an advantage to the client-server model (insert cloud jargon here), that delay is one of the consequences.

What is totally inexcusable is an unnecessary blocking UI animation. I turn off things like the weird genie animation on OS X where I can. Some of them have training purposes, but after that it's a second delay or longer for some of the most common interactions with the OS. So much wasted time. I'm also very critical of Elite: Dangerous, since 1000 years in the future your holographic displays take an appreciably long time to open up, during which time they're all fuzzy like some sort of degaussed CRT. Connecting to a space station's network is also slow (which is due to the limitations of the real world) but while that's loading, it sounds like an 80s hard drive, which I imagine is a callback to the experience of playing the original Elite.

xtifr
Posts: 327
Joined: Wed Oct 01, 2008 6:38 pm UTC

Re: 1619: Watson Medical Algorithm

Postby xtifr » Tue Dec 22, 2015 8:20 pm UTC

Neil_Boekend wrote:
xtifr wrote:
Neil_Boekend wrote:I am not an experienced programmer and even if I was I wouldn't be able to program an OS on my own.

And I'm not saying you should. But if you don't, you don't get to complain about how the computer doesn't do what you told it to—because you didn't tell it what to do. You delegated that task to someone else. You can complain that it doesn't do what you want, and you can complain that the people who wrote the programs you use are idiots. But the computer is not at fault. It does what it's told. Every single freaking time. Whether that's what you wanted it to do or not.

The quote you originally responded to was incomplete. The full quote is "the computer does exactly what you tell it to do—not what you want it to do!" It was intended as a reminder to programmers, not a promise to hopeful users. :)

I know the quote. I just disagree.

I'm not sure I understand which part you disagree with.

If you click a button marked "Save", you're not telling the computer to save your file. You're telling it to run a set of instructions attached to that button, probably by some third party (unless you are a programmer). If you know what you're doing (and this doesn't require programming skills on most internationalized software), you can change the button so it says "Burp" instead of "Save", but that won't change what you're telling the computer to do when you click on it. Just what it looks like.

I can't disagree with your other point about performance, though. That's one of the things I like about LInux. Even without programming it, I can choose between several different user interfaces, ranging from big-and-slow ones that try to do everything, to tiny, minimalist ones where I need to know what I'm doing, and manually reload its hand-edited configuration files whenever I want to make a change, but boy they're zippy without all that bloat.... No, its still not something I'd recommend for the average user, but it definitely works for me. ;)
"[T]he author has followed the usual practice of contemporary books on graph theory, namely to use words that are similar but not identical to the terms used in other books on graph theory."
-- Donald Knuth, The Art of Computer Programming, Vol I, 3rd ed.

*Kat*
Posts: 75
Joined: Sun Dec 26, 2010 1:36 am UTC

Re: 1619: Watson Medical Algorithm

Postby *Kat* » Tue Dec 22, 2015 8:54 pm UTC

Three thoughts:
1) Somebody is in a mood.
2) I hope Randall's cold hasn't turned into pneumonia.
3) Sounds more or less like my friend's recent experience with the medical "professionals" she encountered while caring for her terminally ill partner. (Perhaps the "best" of the bad experiences came the day he died when she called the hospice center to let them know... and was made to answer a multitude of questions before being told to, "Have a nice day.")

User avatar
Flumble
Yes Man
Posts: 2024
Joined: Sun Aug 05, 2012 9:35 pm UTC

Re: 1619: Watson Medical Algorithm

Postby Flumble » Wed Dec 23, 2015 12:17 am UTC

Quey wrote:
Flumble wrote:Me too. I can't fathom how people are fine with an application (or worse, the OS) hanging for a second when they press a button.
Even my phone's default apps (most notably the app store c.q. updater) have blocking views that first have to load <something big> before anything is rendered and touch events are handled (and yes, they are all queued, so don't touch anything before it's loaded).

A lot of that is probably network access, which can take a lot longer than internal gate flipping. This is especially true for smartphones, which are expected to deliver a high standard of experience with a tenth of the hardware, in broad terms. And while there is an advantage to the client-server model (insert cloud jargon here), that delay is one of the consequences.

Indeed, in the case of the main screen of the app store the <something big> is fetching lists of popular apps and their thumbnails –though I think that should be done asynchronously too, even if it means half the content is missing.
However in the case of the update screen of the store the <something big> is compiling a list of installed apps (which is about 5+20 bloatware) with "check for updates" as a button, not a procedure that's started when going to the view.
Luckily, since the last system update I have yet to see a view (which is not a splash screen) that takes the UI thread hostage for more than a second. So I'll quit my ranting for now. :mrgreen:


xtifr wrote:If you click a button marked "Save", you're not telling the computer to save your file. You're telling it to run a set of instructions attached to that button, probably by some third party (unless you are a programmer).

I guess there's a difference in perspective: I press a button labeled "save" with the intent to have the computer save the current data (depending on context either a designated/previous save location or the program asking me where to store it). If the instructions attached to that button include "open the web browser and post a knock-knock joke on facebook", the computer isn't doing what I want it to do. Note that "the computer" here is more than the hardware and a set of instructions: it's also the conglomeration of layers upon layers of abstraction and human brainfarts errors thoughts resulting in the good-for-nothing* software I'm using.

*if the button were labeled "post knock-knock joke on facebook" it would be a great program. Maybe an editor that lacks a save function isn't perfect, but at least it keeps your social media life fresh at the click of a button. :)


Return to “Individual XKCD Comic Threads”

Who is online

Users browsing this forum: cellocgw and 28 guests