|
The Fool posted:We took sand and lightning and made a thinking machine. Based on the lowtax memorial thread in GBS right now I would say both simultaneously
|
#
?
Nov 13, 2021 03:02
|
|
|
|
| # ? Jun 2, 2024 21:02 |
|
|
Any time I think of anything analog it breaks my brain. How the gently caress did we discover radio? And it was analog..? Just seems so much harder than digital. Records? How? How do you store sounds in the grooves of a record? Magic.
|
|
#
?
Nov 13, 2021 03:02
|
|
|
How do magnets work?
|
|
#
?
Nov 13, 2021 03:09
|
|
|
I always think about what if I got sent back to the 1400's. How did electricity work? gently caress if I know. I'd be so screwed.
|
|
#
?
Nov 13, 2021 03:09
|
|
|
GreenNight posted:I always think about what if I got sent back to the 1400's. How did electricity work? gently caress if I know. I'd be so screwed. This is why I got that shirt that explains a whole bunch of basic processes and how to do them with no technology, and I wear it whenever I think there is a chance of randomly getting teleported back in time to the 1400s so I can have super advanced knowledge and live like a god.
|
|
#
?
Nov 13, 2021 03:26
|
|
|
Probably be burned at the stake. Although if you're in the US and you go back to the 1400's, maybe not.
|
|
#
?
Nov 13, 2021 03:29
|
|
|
Internet Explorer posted:Any time I think of anything analog it breaks my brain. How the gently caress did we discover radio? And it was analog..? Just seems so much harder than digital. Records? How? How do you store sounds in the grooves of a record? Magic. There's an old BBC television series called Connections. Episode 3, "Distant Voices" talks about... well a lot of things... but back in 1670, a Frenchman named Jean Picard was walking home after observing this new thing called a barometer he saw a spark inside his mercury-filled glass tube. https://www.youtube.com/watch?v=eCp8h9RkaSw
|
|
#
?
Nov 13, 2021 03:38
|
|
|
GreenNight posted:Probably be burned at the stake. Although if you're in the US and you go back to the 1400's, maybe not. Joke's on them, everyone dies from me spreading our futuristic super diseases.
|
|
#
?
Nov 13, 2021 04:01
|
|
|
SyNack Sassimov posted:Now do N-key rollover So N key rollover is the ability to depress multiple keys and have them be reported successfully before the previous key has been fully released. This is something we're all very used to on keyboards today, but it really didn't become a thing anyone implemented till the 70s, and not popular till the 90s. It's an important feature because it dramatically reduces the number of errors you make as a typist. There's a few different implementations, but basically it depends on whether or not you're using USB or not, and whether or not your keyboard is setting an encoded data (i.e. the ASCII integer representing your character) or not. For older keyboards (60s/70s), you mostly had encoded data, and this was long before USB was a thing. So do this at the circuitry level by only keeping the circuit closed for a brief fraction of a second (like 0.1ms). It doesn't matter if you keep it held down after that so long as your circuit design doesn't prohibit the circuits completed by other keys from completing and sending their electrical signal. This is called pulse output, which you can achieve either electrically or mechanically... but nobody wanted to pay for the cost of implementing it at the circuit level. Instead, with reed or mechanical keyboards, the way the design worked using "level switches" meant you inevitably got more keyboard collisions when two keys were activated at once. The keyboard's logic circuitry would detect collisions and try and report them in the correct order while blocking further operation until the collision cleared. The problem with this is that it could only support 2 key rollover. If one key is actuated while the other is depressed, you get a collision and you can't send data to the register containing the ASCII encoded value of the key until the first key is released. This means that you can recover from a situation where two keys are depressed, but if three keys are depressed (say, CTRL+ALT+Delete), then you always lose the middle key. We solve this by moving to "Matrix Scanning". For simpler input devices, you can just connect to one pin on your microcontroller and read the voltage level. But when we have more switches that need to be reported at once than we have free pins, we need to start multiplexing our inputs. I really don't understand exactly how that is implemented at the circuit level but this page here goes into some detail about it: http://www.openmusiclabs.com/learning/digital/input-matrix-scanning/index.html Exactly how many keys your matrix scanning circuit can detect at once depends on the implementation, which <electrical engineering jargon funny brain noises>. ANYWAYS, this is how we do things today, but in order to prevent "ghost signals" (erroneous keypresses due to conductive sensing matrix implementations having a weakness where the current can flow backwards through the switches and exit somewhere it's not supposed to), we have to use additional diodes (devices that let current flow only one way) in our circuit which means additional cost. Thus, you get X key rollover based on the number of diodes the keyboard circuit board manufacturer has decided to implement. In theory this could be enough to support full N key rollover, where you can independently detect any key, regardless of how many keys are depressed, if you have a diode for each circuit switch. So at the end of all of this, we have the number of keys that can be reliably depressed at once as a function of how much money manufacturers are willing to pay. So how do these electrical signals make their way to the device, in sequence and in order? This is where the protocol comes into play: If you're not using USB, you're probably using an interrupt driven sequence of presses, which means you can send as many keys in sequence as you have reserved memory locations. 2-key rollover is 2 bytes, 6 is six, after that is excessive for us ten fingered humans. Anyways we store the keys pressed during the last scan event (which produces an interrupt telling the CPU to read the data stored at our keyboard's memory registers). Again, this depends on how many registers we are willing to implement and the amount of memory we're willing to reserve to hold the scancode, which is another cost function. Since USB uses a polling design rather than an interrupt design (though I think we still use a kernel interrupt to signal that there's data in a HID event to be read), this means we can report all the keys currently held, but not the individual press and release events. A USB HID device is ultimately going to be sending a stream of Input Reports to the host (via the USB controller), which describe the current state of the device's inputs (keys). This is the data part of the binary data I was talking about earlier. The structure of this data is specific to the HID input device, but it will ultimately comply with a HID device format specification, the data spec to be determined by the Report Descriptor sent from the device to the host. So this is how we interpret data structures from an unstructured binary stream. In other words, the binary stream of input reports generated from a joystick's or a mouse != a keyboard's input reports. This data forms a "key-pressed bitmap", and there's nothing stopping us from, say, declaring that this 106 key keyboard has a 106 field bitmap, with one bit representing every key. This gets us the state of every key on every polling cycle, thus achieving N key rollover. There's apparently a myth that USB keyboards only support 6 key rollover - this is because there's a limited implementation of the HID spec designed for BIOS writers, which sets a specific format for the report descriptors so they don't need to implement the full HID spec. This is called the boot protocol and all keyboards implement it, and in this format, other than the 8 modifier keys (LCtrl, RCtrl, LShift, RShift, LAlt, RAlt, LWindows, RWindows), the depressed keys are packaged into a fixed 6 element fixed length array of scancodes. Thus, the boot protocol only supports reporting the state of 6 keystrokes at once, unless the BIOS writer chooses to use a different and more complete implementation of the HID spec. Sources: https://telcontar.net/KBK/tech/rollover https://www.devever.net/~hl/usbnkro http://www.openmusiclabs.com/learning/digital/input-matrix-scanning/index.html https://forum.arduino.cc/t/how-to-program-n-key-rollover/890461
|
|
#
?
Nov 13, 2021 04:14
|
|
Cat Army
|
MY ceo is on vacation this week and is just now catching up the emails where the CISO deleted a production server due to a default RDP NSG. He specifically replied to my only contribution to that 20+ deep email thread where I had said "Couldn't we have just modified the NSG and then not deleted anything?" and nothing else from the thread. He is pretending to not know what a NSG is (he does) and is asking me to elaborate why modifying the NSG would have a better solution. I am now writing why making a very simple, basic change was the right move instead of swinging dicks in prod. He is having me, her subordinate, explain to the room why she done hosed up. I can tell you with full confidence that somewhere in the midwest my boss is cursing my name with every foul thing she can think of. Every scenario where I have had contrary opinions than her in the cloud space has enraged her to the most extreme degree. Me doing this in front of her peers in email, directed by her boss, might make her quit. Funny thing is that it seems that this server might not have been as prod as people are reporting, but people are settling grudges right now it seems.
|
|
#
?
Nov 13, 2021 04:15
|
|
|
I'm thinking your zero percent chances of being offered the CISO position is going up.
|
|
#
?
Nov 13, 2021 04:17
|
|
|
GreenNight posted:I'm thinking your zero percent chances of being offered the CISO position is going up. I will at best be invited to talk to possible replacements if she actually fucks off. I might sit in a few committees until he hire someone. I don't have the social status for clevel consideration.
|
|
#
?
Nov 13, 2021 04:22
|
|
|
I hear you on that. Something I've been trying to do at my own job is to better market myself to the c-levels and toot my own horn more.
|
|
#
?
Nov 13, 2021 04:23
|
|
|
Has anyone mentioned that you ask really good questions ? Because I have Linux support as a responsibility, have been in IT for ~30 years, make six figgies plus bonus and equity, and... I know what most of those words mean. You are going to soar past me in technical expertise and, ideally, compensation.
|
|
#
?
Nov 13, 2021 09:17
|
|
|
mllaneza posted:Has anyone mentioned that you ask really good questions ? Because I have Linux support as a responsibility, have been in IT for ~30 years, make six figgies plus bonus and equity, and... I know what most of those words mean. You are going to soar past me in technical expertise and, ideally, compensation. You know hardware people don’t get along with software people right?
|
|
#
?
Nov 13, 2021 09:48
|
|
|
Has anyone else gotten a "Threat actor in systems" email from fbi.gov, passes SPF/DKIM/etc and is sent from a .gov IP? It's obviously fake, but it just got spewed into all the whois/arin/ripe and domain contacts I can see, opening up thousands of tickets.
|
|
#
?
Nov 13, 2021 10:22
|
|
|
Internet Explorer posted:Any time I think of anything analog it breaks my brain. How the gently caress did we discover radio? And it was analog..? Just seems so much harder than digital. Records? How? How do you store sounds in the grooves of a record? Magic. Records are easy, it's literally just one form of vibrations (grooves in shellac/vinyl) translated to another form of vibrations (sound waves in air), with can be done with no electricity or electronics at all. The electronics only serve to make it louder and have a wider frequency spectrum.
|
|
#
?
Nov 13, 2021 10:51
|
|
|
SopWATh posted:There's an old BBC television series called Connections. Connections 1/2/3 and The Day The Universe Changed are seriously some of the best tv docs ever made and probably the only media i've ever packratted. That and The Secret Life of Machines, which is on archive.org: https://archive.org/details/Secret_Life_Of_Machines
|
|
#
?
Nov 13, 2021 13:20
|
|
|
JehovahsWetness posted:Connections 1/2/3 and The Day The Universe Changed are seriously some of the best tv docs ever made and probably the only media i've ever packratted. I haven't watched them all but Tim Hunkin has been uploading remastered versions and also did a few new ones in his workshop as The Secret Life of Components during the last year: https://www.youtube.com/c/timhunkin1/videos They're a good watch!
|
|
#
?
Nov 13, 2021 14:18
|
|
|
Bitcoin posted:Has anyone else gotten a "Threat actor in systems" email from fbi.gov, passes SPF/DKIM/etc and is sent from a .gov IP? I have not but I just saw the discussion going on here: https://old.reddit.com/r/sysadmin/comments/qsun7o/email_from_fbi_looks_odd/
|
|
#
?
Nov 13, 2021 15:18
|
|
|
KozmoNaut posted:Records are easy, it's literally just one form of vibrations (grooves in shellac/vinyl) translated to another form of vibrations (sound waves in air), with can be done with no electricity or electronics at all. The electronics only serve to make it louder and have a wider frequency spectrum. If you want to really get into the breakdown of how things work, the TechnologyConnections channel is great. His earlier stuff is dry and discusses more of the analog to digital stuff, but his latest videos are a great blend of incredibly informative and sassy. Pro-tip: Turn on the captions for his videos. There's often jokes embedded.
|
|
#
?
Nov 13, 2021 18:56
|
|
|
Yo, technology connections is great. If you like his channel, check out Applied Science. On my phone so no link. But his stuff is some of the best and most informative content I have seen on YouTube Have this oldie but goodie https://m.youtube.com/watch?v=vvzWaVvB908
|
|
#
?
Nov 13, 2021 23:49
|
|
|
I watched a looot of Technology Connections over the past few months sitting with the newborn. Good poo poo, puts her right to sleep. Look up the one about the toaster Never heard of Applied Science, thanks for the link! CloFan fucked around with this message at 23:58 on Nov 13, 2021 |
|
#
?
Nov 13, 2021 23:56
|
|
|
There was a discussion the other day about working for company A or B, what is the most ethical/unethical. I think the result was basically "get paid". But imagine working for these loving guys: https://twitter.com/ericisaac/status/1459225065636171788?s=20 loving kill me at that point
|
|
#
?
Nov 14, 2021 12:04
|
|
|
Bitcoin posted:Has anyone else gotten a "Threat actor in systems" email from fbi.gov, passes SPF/DKIM/etc and is sent from a .gov IP? Yeah, followed by a drat phishy email supposedly from ARIN. Thankfully I was able to send the security team slack the spamhaus tweet and verify that we got the same email they did : https://twitter.com/spamhaus/status/1459452609979371520 The ARIN email was sent to the filter as Potential Phishing and we'll see how that turns out
|
|
#
?
Nov 14, 2021 12:57
|
|
|
The issue that allowed this to happen turns out to be remarkably stupid, per Brian Krebs: https://krebsonsecurity.com/2021/11/hoax-email-blast-abused-poor-coding-in-fbi-website/ quote:Much of that process involves filling out forms with the applicant’s personal and contact information, and that of their organization. A critical step in that process says applicants will receive an email confirmation from eims@ic.fbi.gov with a one-time passcode — ostensibly to validate that the applicant can receive email at the domain in question.
|
|
#
?
Nov 14, 2021 14:25
|
|
|
I haven’t seen it posted in the thread, for today only, this book on how powershell came to be is free/donation ware https://twitter.com/concentrateddon/status/1438226483210772484
|
|
#
?
Nov 14, 2021 17:27
|
|
|
orange sky posted:There was a discussion the other day about working for company A or B, what is the most ethical/unethical. I think the result was basically "get paid". But imagine working for these loving guys: I can tell you a little bit about it. I'll start at the beginning for context so sorry if I'm saying things people already know. So prisoners get access to a phone, they can buy phone credit to make calls and the rates are wildly inflated compared to what you would pay in the community. I'm sure the justification is the increased overheads of working in a secure environment, I am sure the real reason is the captive audience. I'll try and keep away from ethics  These days, in my country where prisoners are usually kept in their own cells or double cells, the phone companies will pay for the cabling to go into the cells (which is a fun job, let me tell you) because they know they will make the money back on the install due to the captive audience element of it. The headache for the prisons is that all these calls have to be monitored (not every prisoner gets 100% monitoring but it's still a big undertaking) so the phones are only switched on at certain times, basically dictated by the staffing levels of the prison and the ability to keep up with monitoring the calls. So, this means with semi-limited access to phones AND prisoners having to follow the routine of going to work, exercise and so on, it means that it is not always possible for prisoners and families to be on the phone at the same time. This created a gap in the market for people like the example you've highlighted to middle man voicemail systems and capitalise on the situation. Over here, I know some prisons block these services as they cannot guarantee who is leaving/listening to the voicemails therefore the risk is the prison may accidently facilitate the breaching of a restraining order.
|
|
#
?
Nov 15, 2021 11:40
|
|
|
I'll happily discuss the ethics of it - there shouldn't be such a thing as a "gap in the market" or an idea of a captive audience. People in prison are losing their freedoms already, the stated aim of prisons is rehabilitation, letting prisoners talk to their families without putting cost barriers in place is a basically zero-cost way to help with this. Put the phones in a room that is only open in waking hours so you can monitor calls, make the calls free because your actual cost is about 2p a minute and it's not worth worrying about. Monitor the system for abuse and put limits on the duration of each call, or the amount of time spent on the phone each month if you want, maybe have the recipient confirm they want to be connected before putting the call through. Charging people 13c a second to hear a voicemail and putting it behind a system that charges $8 to process a credit top-up payment is indefensible.
|
|
#
?
Nov 15, 2021 12:23
|
|
|
I think this was brought up in passing a few pages back. But tell me about the horribleness of Xenserver. My company wants to switch from ESXI and VMWare Horizon Apps to Xenserver and Citrix because of licensing costs and the terrible support for Horizon Apps. But let's be honest it's probably 90% the costs and 10% the upper echelon pushing for it, there is probably some grifting involved. While I luckily seldom have to touch Horizon Apps right now, I probably also won't have to deal with Citrix. But what horror awaits me in having to deal with Xenserver in tyol 2022? Because from my limited research nobody uses that since at least an decade.
|
|
#
?
Nov 15, 2021 13:32
|
|
|
Thanks Ants posted:I'll happily discuss the ethics of it - there shouldn't be such a thing as a "gap in the market" or an idea of a captive audience. People in prison are losing their freedoms already, the stated aim of prisons is rehabilitation, letting prisoners talk to their families without putting cost barriers in place is a basically zero-cost way to help with this. Put the phones in a room that is only open in waking hours so you can monitor calls, make the calls free because your actual cost is about 2p a minute and it's not worth worrying about. I should probably highlight where these initiatives end up as it is relevant. Typically, the government are very very slow on the uptake of new technology, hence why private providers even have the opportunity to identify gaps in the market and try and get their systems in place. There is similar example of this happening with the method used by families for sending money into the prison (for commissary) - the non-tech process was very open to exploitation by money launderers so a private provider made a digital version that happened to charge families for spending money in for the convenience After ~a while~ the government looked at this and said "its absolutely wrong that people are being charged for this service, we need to make our own system and not charge people for it" (rightly so and they did invent their own system which closed the market for the private provider and put that particular product in its grave) We ~think~ that the government typically dont like to "be the first" to put any such systems in place because they are so scared if something goes wrong there would be reputational damage, but typically, they will let private providers test the waters, then if nothing goes wrong, they take over and make it a free service. It's quite bizarre but that's generally how the machine works.
|
|
#
?
Nov 15, 2021 14:09
|
|
|
angry armadillo posted:I should probably highlight where these initiatives end up as it is relevant. And yet, charging an arm and a leg for something where there's absolutely no alternative and the population in question is already extremely vulnerable to exploitation is still extremely unethical. Just because the system is designed to allow people to grift doesn't mean they should.
|
|
#
?
Nov 15, 2021 15:01
|
|
|
To be clear, I'm not condoning it. I am giving you a first hand account of what happens in practice, as I work in the industry.
|
|
#
?
Nov 15, 2021 17:07
|
|
|
Maybe not condoning it, but I think the issue that people have with your explanation is that it implies that regulation catches up and fixes the issue. That it's a lag due to slow turning wheels of government rather than flat out corruption and graft. It isn't and it won't because the people in the government are the ones making the money off of the private exploitation of people.
|
|
#
?
Nov 15, 2021 17:15
|
|
|
Friendly reminder that Angry Armadillo doesn’t work for an American prison, if I remember right.
|
|
#
?
Nov 15, 2021 17:17
|
|
|
Regulation on prison phone calls is happening in the US, but it's moving slowly.. the FCC has put a bunch of caps in place. But there's a lot more to be done: https://www.prisonpolicy.org/phones/state_of_phone_justice.html
|
|
#
?
Nov 15, 2021 17:24
|
|
|
I am just posting my experience of a system that was referenced given it pertains to my job. If you want to criticise the government then go for it. As it happens apparently we are one of the last prisons here not to have this voicemail system - I get a pushy email off the CEO every 6 months or so and he follows up with the prison boss to say I'm unhelpful and he tells him to go away as well. We won't take it due to the restraining order issue but a lot of prisons do because they can put "increased family ties helps rehabilitation" in an inspection report somewhere The Iron Rose posted:Friendly reminder that Angry Armadillo doesn’t work for an American prison, if I remember right. Correct
|
|
#
?
Nov 15, 2021 19:24
|
|
|
Hate to play backseat mod but maybe this ethics debate should go into the it ethics thread (https://forums.somethingawful.com/showthread.php?threadid=3971776) rather than the generic working in it one
|
|
#
?
Nov 15, 2021 20:00
|
|
|
My thoughts with starting that thread were basically, if there's enough interest in it it will sustain itself and people could be pointed there, if not, then those types of discussions could stay in their individual threads as long as it doesn't make the thread unreadably lovely. Given that it fell off, I wasn't going to worry about it too much. I didn't really do my part in trying to keep it active, either. There's also a thread in D&D about tech topics which actually sees quite a bit of use. https://forums.somethingawful.com/showthread.php?threadid=3763277 From a mod perspective, I'll say that I think these tangential conversations are usually good ones to have and unless it gets out of control or people are being super lovely, I'm inclined to let them go. Lord knows I started enough of these types of derails. In the end, I think they're a net positive. Happy to hear feedback though if anyone wants to PM me and have a discussion about it. Internet Explorer fucked around with this message at 20:13 on Nov 15, 2021 |
|
#
?
Nov 15, 2021 20:09
|
|
|
|
| # ? Jun 2, 2024 21:02 |
|
|
I like having these tangents in the core threads for a couple reasons: 1. Having the conversational topics flow helps build the sense of a mini-community that I like. 2. I only ever visit threads that I have bookmarked and very rarely update my bookmark list. Having topics get split off from the main threads often results in me just missing the content altogether.
|
|
#
?
Nov 15, 2021 20:13
|
|
