|
canis minor posted:So, I'm integrating new payment gateway today. While their documentation makes sense, the example code is...
|
# ? Nov 16, 2016 19:55 |
|
|
# ? Jun 7, 2024 18:01 |
|
Ralith posted:Is it just me or is every last one of those variable names using a prefix meant to designate "null-terminated string" in a language that doesn't expose the notion of null terminators? Love that szPassword, too. I'm sure it's nicely encrypted, though.
|
# ? Nov 16, 2016 20:42 |
|
Cuntpunch posted:
One knock on LINQ prior to Visual Studio 2015(?) was that you couldn't debug lambda expressions.
|
# ? Nov 16, 2016 21:03 |
|
Ralith posted:Is it just me or is every last one of those variable names using a prefix meant to designate "null-terminated string" in a language that doesn't expose the notion of null terminators? yup, that's correct! We had a look and it appears to be a software package that "payment merchants" can just buy and resell
|
# ? Nov 16, 2016 21:05 |
|
wide stance posted:One knock on LINQ prior to Visual Studio 2015(?) was that you couldn't debug lambda expressions. Oh yeah, finding that in old legacy code might make some semblance of sense. That comment is one month old on a UWP project so we are all working with vs2015
|
# ? Nov 16, 2016 23:52 |
|
Spotted in logs "No need to log this"
|
# ? Nov 17, 2016 02:00 |
|
Ghost of Reagan Past posted:Spotted in logs I wonder how often "this should never happen" or similar phrases show up in the production logs of your average Big Software Company.
|
# ? Nov 17, 2016 02:32 |
|
boo_radley posted:Love that szPassword, too. I'm sure it's nicely encrypted, though. Symmetrically encrypted with the Zodiac cipher
|
# ? Nov 17, 2016 02:33 |
|
TooMuchAbstraction posted:I wonder how often "this should never happen" or similar phrases show up in the production logs of your average Big Software Company. well code:
|
# ? Nov 17, 2016 03:35 |
|
Ralith posted:Is it just me or is every last one of those variable names using a prefix meant to designate "null-terminated string" in a language that doesn't expose the notion of null terminators? Way back in the day I remember seeing tutorials that used an "lpstr" prefix for string variables in ASP code, written by someone who worked for Microsoft.
|
# ? Nov 17, 2016 11:59 |
|
The future is now https://www.youtube.com/watch?v=F0r9D6HZEnU
|
# ? Nov 17, 2016 12:03 |
|
mmkay posted:The future is now Truly a great time to be alive
|
# ? Nov 17, 2016 15:20 |
|
mmkay posted:The future is now Something like a decade ago I watched a talk given by a programmer who got carpal tunnel syndrome but still needed to be able to code. During what little "real typing" time he had each day, he hacked one of the more modder-friendly voice recognition tools to add a bunch of special "words" that corresponded to actions and symbols that are common in programming but not in normal text. Like, "zzp" mean { and "pzz" meant }, and he had words for un/indenting, creating a ( ) and moving to the interior of the parens, etc. He had some crazy setup to get the software to work, like, he did his programming in Vim running in a Windows VM in his Linux workstation, because the voice recognition software was Windows-only. But he demoed doing some nontrivial programming and it actually worked pretty well.
|
# ? Nov 17, 2016 17:03 |
|
Speaking of funny comments, found this at the start of one of our tree searching algorithm. I've guessing the dev got really bored...code:
code:
Gavinvin fucked around with this message at 17:38 on Nov 17, 2016 |
# ? Nov 17, 2016 17:35 |
|
Ghost of Reagan Past posted:well I wonder if that's related to trying to land Neil and Buzz on a huge boulder. Gavinvin posted:Speaking of funny comments, found this at the start of one of our tree searching algorithm. I've guessing the dev got really bored... The good thing is that it's obviously a joke and you're not obliged to read it to understand the code. wide stance fucked around with this message at 18:07 on Nov 17, 2016 |
# ? Nov 17, 2016 18:02 |
|
TooMuchAbstraction posted:Something like a decade ago I watched a talk given by a programmer who got carpal tunnel syndrome but still needed to be able to code. During what little "real typing" time he had each day, he hacked one of the more modder-friendly voice recognition tools to add a bunch of special "words" that corresponded to actions and symbols that are common in programming but not in normal text. Like, "zzp" mean { and "pzz" meant }, and he had words for un/indenting, creating a ( ) and moving to the interior of the parens, etc. This guy? https://www.youtube.com/watch?v=8SkdfdXWYaI
|
# ? Nov 17, 2016 18:58 |
|
Yeah, that's the guy! Clearly I misremembered some of the details (including how long ago it was ) but it's still really impressive.
|
# ? Nov 17, 2016 19:27 |
|
Wow that's really cool, that's kind of what I was hoping for someone would do when I was watching the other video.
|
# ? Nov 17, 2016 20:01 |
|
TooMuchAbstraction posted:Something like a decade ago I watched a talk given by a programmer who got carpal tunnel syndrome but still needed to be able to code. During what little "real typing" time he had each day, he hacked one of the more modder-friendly voice recognition tools to add a bunch of special "words" that corresponded to actions and symbols that are common in programming but not in normal text. Like, "zzp" mean { and "pzz" meant }, and he had words for un/indenting, creating a ( ) and moving to the interior of the parens, etc. https://www.youtube.com/watch?v=lQ91SVKryYU
|
# ? Nov 17, 2016 22:01 |
|
All this voice recognition stuff ... Bullshit. What we need is the ability to just think and the computer to just do: https://en.wikipedia.org/wiki/Brain%E2%80%93computer_interface . With a bit more research we can overcome whatever barriers exist right now in BCI to be commercialized. That's where the future is.
|
# ? Nov 18, 2016 02:06 |
|
As someone who does research highly related to brain computer interfaces, I really don't see anything being viable for general use in any time soon unfortunately. Invasive technologies (require surgery to implant into the brain) like ECoG are showing some limited promise in very specific applications to my knowledge, but are obviously not viable for the general healthy public and require specific tuning for each individual. The only non-invasive technology (no surgery) that is even remotely viable is EEG. The rest of the technologies that currently exist either require a massive device and powerful electric/magnetic shielding (MEG,fMRI) or give almost no information (FNIR). EEG has a fairly terrible signal to noise ratio, poor spatial localization due to the scalp's effects on the signal as an insulator, is drowned out by any muscle movement including blinks, and can't even detect signal from more than a centimeter or so below the surface of the skull. Furthermore, the EEGs that actually give worthwhile signals cost tens of thousands of dollars and require set ups that are far from mobile. I'd say we still don't have a particularly great idea what level of information can even be decoded from an EEG signal (my area of research). Simple devices like the P300 speller (https://www.youtube.com/watch?v=wKDimrzvwYA) that take several seconds and full concentration + visual attention to print out a single letter are still basically state of the art. Here's a recent publication where they achieved a 120 bits per minute information rate with a p300 device: http://iopscience.iop.org/article/10.1088/1741-2560/13/2/026024/meta Hopefully some breakthrough physics will allow engineers to develop some technology that can actually extract reasonable amounts of information from the brain and be mobile. I don't have any idea what that would be since electrical information is limited by the distance square law and the insulation of the scalp, and magnetic information is at the femtotesla (10^-15) level, which literally anything that produces a electricity/ a magnetic current will drown out. As it stands, I don't see BCIs being used outside of specific medical and research settings for a very long time. I hope I'm wrong though!
|
# ? Nov 18, 2016 03:08 |
|
Dairy Power posted:Hopefully some breakthrough physics will allow engineers to develop some technology that can actually extract reasonable amounts of information from the brain and be mobile. I don't have any idea what that would be since electrical information is limited by the distance square law and the insulation of the scalp, and magnetic information is at the femtotesla (10^-15) level, which literally anything that produces a electricity/ a magnetic current will drown out. As it stands, I don't see BCIs being used outside of specific medical and research settings for a very long time.
|
# ? Nov 18, 2016 03:30 |
|
Ralith posted:I wonder if you could do some sort of filtering on the magnetic signals. Surely muscle contractions produce very different looking impulses than neuron activations? Well, one of the biggest issue with the magnetic signal is that it's extraordinarily hard to measure something at the femtotesla level. It involves devices that require a super conductor to measure a macroscopic quantum phenomena (https://en.wikipedia.org/wiki/SQUID). To even have these things functions at the femtotesla level, the sensors have to be cooled by liquid helium -- liquid nitrogen isn't cold enough. So, toting around liquid helium and having the requisite insulation is already a huge barrier. Worse yet, the presence of any electrical device (or anything else that produces a magnetic field) can completely destroy the tuning on these things. One of the grad students I worked with accidentally walked into the MEG room with her cellphone in her pocket, which led to the medical imagining researcher we worked with having to manually recalibrate all 306 sensors. He wasn't too pleased with that situation! But, you can filter out the signals from muscle contractions. I'm not sure about real time methods, but people will frequently use independent components analysis to filter out artifacts that are created by eye blinks after the fact.
|
# ? Nov 18, 2016 03:48 |
|
Dairy Power posted:Well, one of the biggest issue with the magnetic signal is that it's extraordinarily hard to measure something at the femtotesla level. It involves devices that require a super conductor to measure a macroscopic quantum phenomena (https://en.wikipedia.org/wiki/SQUID). To even have these things functions at the femtotesla level, the sensors have to be cooled by liquid helium -- liquid nitrogen isn't cold enough. So, toting around liquid helium and having the requisite insulation is already a huge barrier. Worse yet, the presence of any electrical device (or anything else that produces a magnetic field) can completely destroy the tuning on these things. One of the grad students I worked with accidentally walked into the MEG room with her cellphone in her pocket, which led to the medical imagining researcher we worked with having to manually recalibrate all 306 sensors. He wasn't too pleased with that situation! I'll refrain from further speculation about untapped ideas in your literal research area. Thanks for the links, those'll be interesting reading! It's exciting to hear "we still don't have a particularly great idea what level of information can even be decoded" rather than "we're pretty sure it's useless," to say the least.
|
# ? Nov 18, 2016 04:22 |
|
Ralith posted:Oh, I assumed that was the type of detection EEGs used. Can you tell I have no idea what I'm talking about? No worries, I was throwing out acronyms left and right with very little explanation. (E)EG measures electricity and (M)EG measures magnetic field. And yeah, I'm actually super excited to see how far these things will go. There's only a few papers even trying to use modern machine learning techniques in this field, so there's a lot of fairly low hanging fruit to explore from the CS side.
|
# ? Nov 18, 2016 05:04 |
|
Dairy Power posted:Well, one of the biggest issue with the magnetic signal is that it's extraordinarily hard to measure something at the femtotesla level. It involves devices that require a super conductor to measure a macroscopic quantum phenomena (https://en.wikipedia.org/wiki/SQUID). To even have these things functions at the femtotesla level, the sensors have to be cooled by liquid helium -- liquid nitrogen isn't cold enough. So, toting around liquid helium and having the requisite insulation is already a huge barrier. And good luck even getting access to an MEG machine. The cost of supplying the Helium alone is enough that only a handful of UK universities have one, though HyQUIDs are supposed to require much less Helium. There's also the fact that there's basically only one company currently selling MEG machines, and I've heard that they are very noisy and require constant re-calibration, so radiographers hate them. quote:Worse yet, the presence of any electrical device (or anything else that produces a magnetic field) can completely destroy the tuning on these things. One of the grad students I worked with accidentally walked into the MEG room with her cellphone in her pocket, which led to the medical imagining researcher we worked with having to manually recalibrate all 306 sensors. He wasn't too pleased with that situation! At least that just means the data is nonsense, if it were an MRI machine they could have had their phone tear through their leg.
|
# ? Nov 18, 2016 10:07 |
|
Dairy Power posted:Hopefully some breakthrough physics will allow engineers to develop some technology that can actually extract reasonable amounts of information from the brain and be mobile. I don't have any idea what that would be since electrical information is limited by the distance square law and the insulation of the scalp, and magnetic information is at the femtotesla (10^-15) level, which literally anything that produces a electricity/ a magnetic current will drown out. As it stands, I don't see BCIs being used outside of specific medical and research settings for a very long time. Decades of SciFi literature have promised we'll get room temperature superconductors any day now...
|
# ? Nov 18, 2016 15:55 |
|
Edison was a dick posted:The cost of supplying the Helium alone is enough that only a handful of UK universities have one, though HyQUIDs are supposed to require much less Helium. Oh wow, that's really cool. I wasn't familiar with HyQUIDs. Every time I see an advance like this, I get the urge to learn more physics. Also, to try to restart an MEG collaboration... quote:At least that just means the data is nonsense, if it were an MRI machine they could have had their phone tear through their leg. Ha, yeah. Though, I'd imagine she wouldn't have made the same mistake if she'd heard an active MRI machine. Munkeymon posted:Decades of SciFi literature have promised we'll get room temperature superconductors any day now... I'll settle for my general purpose AI assistant and daily use jetpack, personally.
|
# ? Nov 18, 2016 18:01 |
|
Dairy Power posted:I'll settle for my general purpose AI assistant and daily use jetpack, personally. We're scarily close to the former, honestly. It turns out that rearranging information is a lot easier than the complete wreckage of thermodynamics that is required for the less subtle sci-fi achievements like flying cars and safe, quiet jetpacks.
|
# ? Nov 18, 2016 18:15 |
|
This is definitely one of the more interesting coding horror derails
|
# ? Nov 18, 2016 18:22 |
|
Dairy Power posted:Oh wow, that's really cool. I wasn't familiar with HyQUIDs. I'm only familiar with them because I'm working with the group that's planning to make them. The fun bit is that electronics and chemistry advances mean you can theoretically do a combined MEG and MRI by pinging tuned parahydrogen instead of water and measuring the EM from that using a HyQUID, though you'd need an MEG scanner sampling at a higher frequency than anyone even had plans to make.
|
# ? Nov 18, 2016 20:56 |
|
https://www.youtube.com/watch?v=QM1iUe6IofM
|
# ? Nov 21, 2016 20:41 |
|
Sounds like this guy would really like zero capture lambdas in c++.
|
# ? Nov 21, 2016 22:03 |
|
theratking posted:Sounds like this guy would really like zero capture lambdas in c++. Seriously. "I'd love this but I haven't seen it anywhere ever" *describes lambdas*
|
# ? Nov 21, 2016 22:47 |
|
code:
|
# ? Nov 21, 2016 23:46 |
|
Gonna make a browser extension that replaces "Object-Oriented Programming" with "My Preconception of Java from Cookie Cutter Enterprises".
|
# ? Nov 21, 2016 23:55 |
|
OOP is a coding horror unless there is some stateless fashion of OO that I am unaware of
|
# ? Nov 21, 2016 23:58 |
|
xtal posted:OOP is a coding horror unless there is some stateless fashion of OO that I am unaware of Immutable objects?
|
# ? Nov 22, 2016 00:52 |
|
xtal posted:OOP is a coding horror unless there is some stateless fashion of OO that I am unaware of Don't be tiresome.
|
# ? Nov 22, 2016 01:10 |
|
|
# ? Jun 7, 2024 18:01 |
|
rt4 posted:Immutable objects? Immutable state is still state (unless you mean immutable bc no instance variables)
|
# ? Nov 22, 2016 01:17 |