<rqou> also, (at least one of) the student records systems seems to be powered by an as/400
<lain> they used an as/400 to do student records at my HS
<mtp> same
<mtp> my school system was all AS400
<rqou> hmm i actually have no idea what my high school used
<lain> they also had a samba file share with the csv files that told the lunch system who got free lunch thanks to the government, among other things
<rqou> it might have been (partially) "paper"
<lain> and it had full read/write guest access
<rqou> lool
<lain> the temptation was real.
<rqou> grant it to everybody? :P
<lain> no, that would be noticed
<lain> just myself :P
<lain> but no, I didn't mess with it
<rqou> eh, school lunches are crap anyways :P
<rqou> thanks obama :P :P :P
<lain> I also didn't report it, because the climate was (probably still is) that I would've been accused of "hacking" and disciplined in some way
<lain> lol
<rqou> yup, climate at my HS was similar
<lain> this was long before obama
<rqou> wait really?
<lain> I graduated HS in 2005
<rqou> i had the impression you were a lot younger than that
<rqou> because of the "soylent ant fast food" thing :P
<lain> combination of squalor and efficiency, that :P
<lain> speaking of which
<lain> I'm currently waiting on a 33 page fax to send
<lain> this could take all night
<rqou> fax? wtf?
<lain> yeah
<lain> faster than mail!
<rqou> us gov or japan? :P
<lain> government stuff
<lain> US lol
<lain> trying to get medicaid
<rqou> oh god
<lain> they requested a massive amount of documentation to prove I'm poor
<lain> we'll see how this goes lol
<rqou> i'm still mooching insurance from my parents :P
<lain> lucky
<lain> I'm paying $380/mo to Anthem
<lain> and it's getting untenable :P
<rqou> thanks obama :P
<lain> yeahhh
<lain> I remember paying like $45/mo
<lain> those were the days
<rqou> so are you actually manually feeding a fax machine or using a modem to slowly spit tiff files out?
<lain> haha
<lain> we have some fancy big office fax/copy/scanner
<lain> so it already scanned in the whole stack, now it's just slowly transmitting it
<rqou> i don't understand why they still use fax
<rqou> i bet the other end is also a modem connected to a pc
<lain> yeah
<lain> likely
* qu1j0t3 still uses fax
<lain> I used to have this great online fax service, I wonder if those still exist
<lain> you could just email them PDFs
<lain> with the destination number
<lain> and they'd send it out
<lain> and you'd get a fax number, and they'd email you pdfs of anything sent to you
<lain> very handy
<lain> though I hesitate to use a service like that for something like this, which has my SSN plastered on every page lol
<rqou> yeah i wanted to try to set one of these up for lulz at some point but never got around to it
<qu1j0t3> as long as it's all encrypted...
<rqou> fax isn't encrypted
<qu1j0t3> i know. i meant all the email part.
<lain> they'd still have to decrypt to fax so it'd only be encrypted to them, and it's them that I don't really trust with that info
<qu1j0t3> what I'd REALLY like is for accountants and lawyers to be trained to use encryption.
<lain> ^^^
<rqou> anyways, i wanted to try to set up a fax gateway for lulz but it turns out that apparently "pbx software" is another one of the "batshit insane" software categories
<qu1j0t3> lain: right, threat modelling ftw.
<lain> the Security Now podcast can be a bit.. not great, but they have some great anecdotes from time to time from listeners about lawyers and such trying to use "secure documents" online
<qu1j0t3> oh god
<lain> like one recent story iirc was a place that sent an email with a "secure link" to download it
<qu1j0t3> i worked on some medical coding stuff recently
<qu1j0t3> and let's just say
<rqou> i don't understand why "move some samples of a waveform around and/or convert them back to electricity" is so hard
<lain> and when you click this secure link, you have to create a username and password, and then it lets you download the file
<azonenberg> yeah i've seen systems like that
<qu1j0t3> there are a lot of them
<lain> but the person was like "wait..." and clicked the link again, and again was prompted to create a username and password, and was able to download
<lain> so it's like
<qu1j0t3> lain: Yep!
<lain> it didn't
<lain> do
<lain> anything
<qu1j0t3> that's how they mostly work!
<lain> like wtf
<lain> lol
<qu1j0t3> it's insane.
<lain> :|
<azonenberg> lain: here's another one
<rqou> but you have to have the link first
<azonenberg> client of $WORK
<qu1j0t3> very few of the people i worked with in medical coding had any clue about keep data secure.
<rqou> just don't put that link into bit.ly or whatever :P
<lain> rqou: yeah pbx software is maddening
<azonenberg> they set up a "secure email" system for us to use
<azonenberg> When you get a new "secure message" you get a link to click
<rqou> did i tell my father's "ftp server" story here?
<azonenberg> that lets you view the message after entering a username and password
<azonenberg> only problem is, you can reset the user/pass with access to the nonsecure email the secure box is tied to...
<azonenberg> So basically if you can get into my regular corp email (no access to PGP keys etc required) you can get into the secure box
<rqou> anyways, so my father had a great story about ftp servers at a former employer
<azonenberg> meanwhile we had a perfectly serviceable PGP setup running on our corp boxes that the client could have used...
* azonenberg facepalm
<lain> azonenberg: haha, that's better though - with this one, it didn't even save the user/pass, it just discarded it and let you create a new one every time XD
<lain> so if you just have the URL, you have the file, period
<rqou> some customer (telco) had a problem with their box and needed to send a log file
<azonenberg> yeah this was a little better
<rqou> support says "email it to support@foocorp.com"
<azonenberg> also it wasnt just an email
<rqou> customer says that's not allowed
<azonenberg> it looks like it was a big blob of base64
<rqou> insists on using ftp
<lain> ftp lol
<lain> I hope it's at least sftp
<azonenberg> but the creds were checked server side
<rqou> eventually support guy doesn't want to deal with this and spins up a random box and puts an ftp server on it
<rqou> and says "use user foocorp password foocorp" or something similar
<rqou> and gets the log file, resolves the customer issue, and forgets about it
<azonenberg> so maybe clientside decryption using a key provided by the server after user/pass auth?
<rqou> some months later engineers/it are consolidating servers and notices this ftp box
<rqou> took one look
<rqou> and threw out the harddrive immediately
<lain> inb4 now being used to trade warez by various people on the internet
<rqou> warez, pr0n, possibly cheese pizza, you name it
<lain> wheeeee
<lain> do you knowwww what the most popular usenet server was for binary file trading back in the day?
<lain> it was one of the TLAs, I'm forgetting which one though
<lain> fbi? cia? one of them
<azonenberg> loool
<rqou> lol wow
<lain> they ran the most high-performance binary-tolerant usenet server lol
<lain> or so I'm told by people older than I am who have used usenet for a long, long time :P
<lain> I have to wonder if that's a similar case of server left in closet and forgotten, or if it was intentional (keep track of what people are trading?)
<rqou> fbi tracking cheese pizza with honeypots?
<azonenberg> yeah i was wondering the same
<lain> rqou: wouldn't surprise me
<rqou> i mean, they still do that today
<azonenberg> how many people got busted trading stuff on that server?
<azonenberg> has to be nonzero
<lain> who knows
<lain> <3 usenet though
<rqou> hmm i'm far too young for usenet
<lain> it's still alive and well today
<rqou> for spam :P
<lain> naw
talsit has joined ##openfpga
<lain> there's active newsgroups, but also the alt.bin.* groups :P
<lain> torrents are.. problematic
<lain> I'll leave it at that
<lain> :P
* azonenberg idly wonders what %age of tor traffic is torrents
<lain> shame the nsa doesn't publish such statistics
<azonenberg> lolol
<lain> come on you know it's in a ppt somewhere
<mtp> tor isn't really good for torrents
<rqou> right, the CCC network is better :P
<lain> I've never really messed with tor, but I hear it's really slow
* rqou is planning to bring a 40gbe card to 34c3
zino has quit [Ping timeout: 260 seconds]
<lain> lol
<rqou> although i was told that >10gbe is "bring your own optics"
<rqou> which will be pretty pricey just by itself
<lain> briefly playing with usenet on a gigabit line, I ran into the issue that verifying (par2) the files took longer than downloading them in the first place
<rqou> how does the architecture of usenet work?
<rqou> who actually is responsible for storing files?
<mtp> hierarchichal distribution
<lain> there's various providers, they all interconnect
<lain> what mtp said
<mtp> everyone is responsible for storing the section of the hierarchy that they want
<mtp> and purging it on their own schedule
<digshadow> azonenberg: sem is getting moved on Thursday
<rqou> so if you want to be your own provider, you have to find an existing one willing to send you stuff?
<digshadow> hopefully will be back up shortly thereafter
<lain> many of the larger providers these days are capable of adding storage faster than content is uploaded (I wouldn't be surprised if this is in part due to old data getting DMCA takedowns), so their retention is effectively infinite
<mtp> yep
<mtp> or you can have an independent feed
<lain> here's a map from 4 years ago
<lain> it's useful to have an up to date map like this if you're choosing providers, because often you will want a primary and a backup or "fill" provider to fill in gaps in your primary
<lain> and you'd like those to be served by different backbones, otherwise they're likely to be missing the same pieces
<rqou> wow it's pretty small
<rqou> where did all the old .edu usenet servers go? disappeared?
<rqou> i know my $FANCY_SCHOOL's usenet server was decommissioned in the late 90s or so
<lain> this map may only cover servers that are willing to host binaries :P
<lain> I don't know for sure though
<rqou> lol ok
<azonenberg> Soooo guys
<rqou> how do the providers not end up with "cheese pizza" problems?
<lain> rqou: any provider that wants to stay alive for more than a day has a means of reporting said pizza
<azonenberg> they probably take stuff down when asked
<lain> and they do takedowns swiftly when reported
<azonenberg> FBI etc publishes databases of hashes iirc
<azonenberg> of known pizza
<lain> and alert the authorities, presumably
<azonenberg> So you can just hash all uploaded files and if you get an exact match, remove it silently
<azonenberg> and possibly log
<lain> also there's legal precedent for usenet to exist as a neutral content storage thing
<lain> which is to say
<lain> they're allowed to exist and continue to host /whatever/, so long as they respond to relevant legal requests
<azonenberg> rqou: So for starshipraider's probing arch
<lain> and as long as they never, but NEVER, advertise illegal content
<azonenberg> My current plan is for each line card to mate via a q-strip to the host board
<rqou> right, need to think about what pins are needed there
<azonenberg> (screwed down with standoffs between them, for robustness)
<rqou> i was supposed to do that a month ago :P
<azonenberg> Then have a second q-strip to the I/O card
<azonenberg> the connector card*
<azonenberg> Which will be a purely passive PCB that just breaks a q-strip out to probes
<azonenberg> There will be a low-speed flying leads one
<azonenberg> then a second one designed for high speed that mates to solder-down 1K resistive probes
<azonenberg> How do you suggest connecting said probes?
<azonenberg> SMA is both large and expensive
<azonenberg> u.fl has poor mating cycles
<mtp> azonenberg, orinoco MC-card?
<azonenberg> Not enough res to tell
<azonenberg> mtp: ?
<azonenberg> ink?
<rqou> you don't need res to tell, these hp/agilent probe pods have been around for ~3 decades :P
<lain> ohh man, orinoco cards
<lain> I haven't heard that name in /ages/
<azonenberg> mtp: looks big and expensive
<mtp> it's smaller than SMA
<lain> yeah it's very small
<mtp> expensive, maybe
<lain> it's on the right there
<mtp> there you go
<rqou> so the way i was thinking to connect 1k resistive probes was just "diy the probes from a sma pigtail and a resistor"
<azonenberg> Hmm, interesting
<azonenberg> rqou: Well the issue is
<azonenberg> eight SMAs are not exactly small
<lain> azonenberg: have you considered v-bite connectors btw? I don't know if they make v-bites for smaller than SMA but they're nice pcb-edge-mount connectors with good mechanical strength and excellent rf properties
<rqou> hmm
<rqou> that's why i'm curious how the agilent pods can handle the signal speeds they are supposed to be able to handle
<azonenberg> lain: Never heard of them
<azonenberg> let me see
<lain> actually those all look big :(
<rqou> aren't the agilent pods supposed to be good for 500mbps?
<azonenberg> i dont really know much about rf connectors in general
<rqou> my experience is that "they're all huge (ham, test equipment) or fragile (mini pcie cards)"
<balrog> azonenberg: what's that really tiny connector that's smaller than uFL that's used in some devices?
<balrog> I've seen it in Apple stuff
<rqou> that's even more fragile
m_w has quit [Quit: Leaving]
<balrog> I'm just curious :P
<azonenberg> yeah i dont know whats out there
<azonenberg> one option is to have a small connector like uFL broken out to SMA by pigtails
<azonenberg> then SMA to each probe?
<azonenberg> well sorry, adapter cables
<azonenberg> not pigtails
<azonenberg> Just to avoid having a zillion SMAs in a tiny place
<lain> azonenberg: you could use uhhh
<azonenberg> i'm just thinking how massive this thing is gonna be if it has 32 SMAs on the front, plus power reference inputs for tracking VCCO etc
<lain> azonenberg: samtec makes some things that terminate to SMA
<azonenberg> oh?
<lain> ISORATE on that page
<lain> guessing it's $$$ lol
<lain> balrog: it seems w.fl is smaller than u.fl
<azonenberg> digikey doesnt even seem to stock them
<lain> azonenberg: in my experience samtec is happy to sell direct
<lain> at low volume
<rqou> balrog: ipex mhf4?
<lain> unless that's changed, it has been a number of years since I bought from them :P
<azonenberg> interesting
<rqou> here's a nice ugly as sin connector idea :P https://www.serialcables.com/images/products/SA-36H-D.jpg
<rqou> oh that's not 50ohm though
<rqou> lain: do you know how to rf? i assume 100ohm twinax != 2x 50ohm coax?
<rqou> you can't just do that, right?
<lain> rqou: yeah you couldn't use it like that
<cr1901_modern> Did Twitter die for anyone else?
<lain> cr1901_modern: says something is techincally wrong for me
<azonenberg> rqou: so here's an option for probes
<cr1901_modern> Okay, so it's not just me. Very good
<azonenberg> then add a 1K to the test point
<cr1901_modern> I'll try to live without it for now
<lain> rqou: twinax internally has two lines in the middle, possibly twisted together, surrounded by dielectric and then the shield around that. coax only has one line in the center
<azonenberg> umcc has limited mating cycles, of course
<lain> rqou: if you tried to use the two inner lines of a twinax they'd couple strongly to each other and you'd have crosstalk for days
<rqou> like 10?
<azonenberg> rqou: but you could make the expansion board a "consumable"
<rqou> so when these small connectors break does the cable piece break or does the pcb piece break?
<rqou> or "unspecified?"
<azonenberg> i.e. when a umcc wears out you replace the breakout and get a new q-strip to umcc adapter
<azonenberg> well q-strip to 8x umcc
<lain> azonenberg: if this is for connecting board to board, there are connectors that just have lots of coax in them, or connectors that have good rf properties without necessarily being coax, so you can connect many RF things with one go
<rqou> are q-strips even controlled impedance?
<azonenberg> rqou: yes
<azonenberg> they're specified out to 20+ GHz
<rqou> oh wow
<azonenberg> sorry, 28 Gbps
<azonenberg> 14 GHz
<rqou> no wonder "<foobar> module" manufacturers love them :P
<rqou> yeah, i've seen them
<azonenberg> well i meant the brochure etc on the page
<rqou> i'm even sitting on some samples that i haven't used yet :P
<azonenberg> they're available as either 50 ohm single ended all positions loaded, or 100 ohm differential with pairs of pins then gaps
<azonenberg> Suffice it to say, they will not be our limiting factor
<rqou> so for the probe board
<rqou> when i was at brcm they had a really ugly qsfp breakout pcb that was tree-shaped
<rqou> can't you just make a tree-shaped adapter that can then fit a lot more smas?
<lain> azonenberg: just use type-c cabling lol
<rqou> that's also differential though
<lain> it's rated for high insertion count, cheaply available in bulk, has low loss at high datarates, ...
<azonenberg> rqou: the issue is
<rqou> not single ended
<lain> mmm
<lain> true
<azonenberg> there will be four of these 8-channel boards
<azonenberg> in close proximity
<rqou> otherwise you could have just used a qsfp like i suggested
<rqou> ah, so they won't fit when you have more than one
<azonenberg> Exactly
<azonenberg> Unless i move them very far apart
<azonenberg> in which case the carrier board will be massive
<lain> stack them vertically
<rqou> so what if you take 8x sma to coax pigtails and solder the other end directly onto the adapter pcb?
<rqou> so like in the picture i linked earlier (https://www.serialcables.com/images/products/SA-36H-D.jpg) but replace the qsfp with pcb+qstrip
<azonenberg> So soldering down eight short pigtailis
<azonenberg> and just have SMAs hanging off?
<azonenberg> That's totally doable
<rqou> yeah, should work unless someone somehow accidentally rips one off :P
<azonenberg> Lol a little epoxy will help but you have to expect people to be somewhat careful
<azonenberg> Anyway, this is actually good because it's flexible
<azonenberg> For inputs you could hook up a homebrew resistive probe
<azonenberg> for 50 ohm outputs, you drive directly to the DUT
<azonenberg> And for low speed (less than say 50-100 MHz) you can just use flying leads on 0.1" headers bridged to a qstrip
<azonenberg> like a saleae type input
<azonenberg> all of these will be compatible with the same io stage
<rqou> hmm what if you skip the sma connectors altogether?
<rqou> just have qstrip to pads breakout
<rqou> and a roll of rg316 or whatever
<rqou> and you assemble a new adapter on each victim
<azonenberg> Hmmm
<azonenberg> Can you buy double ended pigtails prestripped?
<rqou> can't you just buy a whole roll and strip it manually?
<azonenberg> Yeah, just thinking about what a pain it'd be to do it by hand :p
<azonenberg> But totally an option
<cyrozap> nats`: If no one has mentioned it already, there are FOSS (CC0) implementations of the xvc protocol here: https://github.com/tmbinc/xvcd https://github.com/Xilinx/XilinxVirtualCable
<azonenberg> And i guess you'd need to manually solder an 0402 or so sized 1k resistor down to the signal being probed
<azonenberg> Also, if i have a qstrip to pads breakout
<azonenberg> there's no reason why you couldn't put a SMA pigtail on it
<azonenberg> rather than soldering straight to the dut
<rqou> although i have a very similar one and ime it barely even works for rg59
carl0s has quit [Quit: Leaving]
<azonenberg> rqou: oh also
<azonenberg> My plan is for the io boards to have an i2c eeprom on them
<azonenberg> (as well as the probe boards)
<rqou> yeah definitely
<azonenberg> So you can determine what's hooked up and automagically ID things
<rqou> you always need "random glue memory" everywhere
<azonenberg> for example, the low speed probes will be 50 ohm native inputs
<azonenberg> The high speed resistive probes will attenuate
<rqou> i was thinking also for e.g. calibration settings
<azonenberg> So i have to scale the Vt DAC appropriately
<rqou> if there eventually is a sdr module or something
<azonenberg> We can come up with some kind of discovery protocol
<rqou> not edid? :P :P
<rqou> least working discovery protocol ever
<azonenberg> lol
<rqou> offtopic: i should buy myself an "actually f*cking works" crimper
<azonenberg> rqou: also i added you to the starshipraider github if you want to look at the wiki (there isnt much there)
<azonenberg> i want to start coming up with mechanical and electrical standards for probe boards
<azonenberg> the qstrip isnt super mechanically robust with all those small pins so i want to design the host board to have PCB under the I/O board
<azonenberg> with a qstrip at the inboard end and say 4-40 screw holes at the outboard ebnd
<azonenberg> end*
<azonenberg> s.t. you can mate the I/O module then screw it down on standoffs to provide more rigidity
<rqou> hmm i wonder if minipcie connectors can be used in some places?
<rqou> they're probably cheaper and can use clips
<azonenberg> ?
<azonenberg> for what
<rqou> maybe for the breakout card
<rqou> so base<->i/o can be q-strip and huge
<rqou> and i/o<->probe can be minipcie-like
<rqou> so it's faster to switch
<azonenberg> Certainly plausible
<azonenberg> Start writing up notes on that wiki page
<azonenberg> for candidate connectors etc
<rqou> hmm the problem is that this makes the probe card require a particular pcb thickness
<azonenberg> Yes, that will have to be standardized anyway if we do screw-down with standoffs
<azonenberg> Target 1.6mm
<azonenberg> that's a de facto standard and most batch fabs etc can do it
<rqou> eh, but you can futz around with washers if you use standoffs :P
<azonenberg> Both probe and io boards will likely be 4 layers
<rqou> hmm in that case 4layer 0.8mm is really expensive for some dumb reason
<azonenberg> the probe boards will be signals around internal grounds
<rqou> (on batch fabs at least)
<azonenberg> IO boards will probably have to have an internal power plane
<azonenberg> But still shouldn't be hard to do
<azonenberg> the carrier board will most likely be 1.6mm 8 layer
<azonenberg> or 6 if i can squeeze it into that
<azonenberg> i really hope one day we'll see lower-end FPGAs that have integrated DRAM in the package
<azonenberg> If we had that, i could probably have done this on a 4L board :p
digshadow has quit [Quit: Leaving.]
digshadow has joined ##openfpga
uwe_ has quit [Ping timeout: 276 seconds]
uwe_ has joined ##openfpga
digshadow has quit [Quit: Leaving.]
amclain has quit [Quit: Leaving]
<lain> or even PoP
<lain> we should make a little carrier board with an artix7 and some dram, have it epoxy encapsulated and BGA'd, and sell it as a thing :P
<azonenberg> Hey, somebody did that with an am335x
<lain> yeah I was just going to say
<azonenberg> if i ever needed a cortex-a8, i'd buy that
<azonenberg> because it lets me use a cheap pcb
<rqou> wait, i thought you could do a ram on a 6 layer pcb?
<rqou> or even 4 layers if you bend the rules a bit?
<azonenberg> Yes but an am335x could be broken out on a 4L board comfortably if you didn't need the ram pins
<azonenberg> i.e. ethernet and some GPIO
<azonenberg> plus consider the cost of engineering time it takes to lay out a big ram interface
<azonenberg> it's nontrivial
<rqou> wait is the am335x not pop like omap?
<azonenberg> Nope
<azonenberg> Look at a beaglebone
<rqou> the old beaglebones are pop?
<azonenberg> not the black
<azonenberg> lain: and PoP FPGAs are unlikely to happen any time soon for thermal reasons
<azonenberg> maaybe a low end zynq
<lain> ahh true
<lain> ok no I've got it
<rqou> <troll>PoP intel cpu?</troll>
<lain> we rout a hole under the bga and use one of those perimeter-populated bgas for the fpga
<lain> and then we do package-under-package
<azonenberg> looool
<lain> and solder the ram to the underside, upside-down
<lain> :D
<rqou> lol
<lain> I'm sure pick&place machines will have /no problem/ with this whatsoever
<azonenberg> rqou: a PoP atom is totally plausible
<lain> they might exist
<azonenberg> lain: just use high melt solder
<lain> but you didn't hear that from me
<azonenberg> and attach it before you start
<azonenberg> Just like the passives under e.g. PGA packaged pentiums
<lain> actually that may be public knowledge
digshadow has joined ##openfpga
digshadow has left ##openfpga [##openfpga]
<azonenberg> rqou: btw
<azonenberg> rather than trying to cut and solder probe pigtails
digshadow has joined ##openfpga
<azonenberg> I'm starting to lean back toward u.fl
<azonenberg> the limited mating cycles arent a big deal if you pretend the probe is permanently attached
<lain> zip-tie them down :D
<azonenberg> But being able to replace it even a few times would be convenient
<azonenberg> Basically, u.fl to SMA cables on each input
<azonenberg> then SMA to the probes
<lain> we had to zip-tie a plug-on module on a board in the panel meters I worked on in a past life. they were used in trains, and uh. turns out vibration will unplug anything.
<azonenberg> lolol
<azonenberg> I was going to screw it down
<azonenberg> this wont move
<lain> :D
<azonenberg> Especially not if you add loctite
<azonenberg> In other news, i love this error message
<azonenberg> "ERROR: Don't know what argument "reset" is supposed to mean, but it's not a valid port number"
<lain> lol
DocScrutinizer05 has quit [Disconnected by services]
DocScrutinizer05 has joined ##openfpga
<rqou> so i'm currently trying to learn (again) how various side channel attacks work
<rqou> and what i've found is
<rqou> people who do these attacks tend to use the "statistics" terminology to explain their math
<rqou> but i don't have a traditional statistics background
<rqou> i learned it all as "machine learning"
<rqou> which uses different words for everything because naming things is hard
<rqou> once you match up "wtf are they actually talking about" it makes a lot more sense :P
<azonenberg> lol
<azonenberg> i found the same thing about machine vision and multivar calc
<rqou> lool
<azonenberg> i honestly learned multivar differential calc
<azonenberg> by analogy to discrete machine vision operations
<azonenberg> i visualized a 2D filter on an image and it made sense
<azonenberg> then just took the intuitive limit as pixel size hit zero
<rqou> heh i remember accidentally discovering moment generating functions when taking intro probability
<rqou> there was a homework problem to find the distribution of the sum of two discrete random variables
<rqou> i had no idea how to do it, so i thought "i can write a brute-force summation and see if i can simplify it"
<rqou> and then i saw "hey, this equation has the same form as a convolution"
<rqou> "i can just take the DTFT of the distributions and multiply them"
<rqou> and then one of my friends showed me how that's actually a legitimate thing to do
<rqou> :P
<rqou> (convergence/boundary condition issues did not occur in this problem)
mifune has joined ##openfpga
mifune has joined ##openfpga
<lain> most of the time if I'm struggling to understand some high-level concept, it's usually because someone has decided to use pure math and/or boatloads of domain-specific terminology to explain it
<lain> like when I was implementing a reed-solomon encoder
<rqou> <insert "why not both" meme here :P >
mifune has quit [Ping timeout: 240 seconds]
<lain> but like
<lain> all this talk about polynomials and finite fields and stuff
<lain> it's all trivial
<lain> but the terminology makes it sounds SOOOOOO complicated lol
<lain> and the notations and stuff
* rqou has to admit that i learned it via the "polynomials and finite fields" method
<lain> I kept reading papers on the subject
<rqou> i should teach myself the BCH view at some point; we didn't cover that
<lain> and books
<lain> and like none of it made any friggin sense because I didn't know the first thing about this terminology they were using
<lain> but it turned out to be lots of concepts I was already familiar with, just didn't know the names of
<rqou> hmm i guess it was much better for me because i learned it in the context of a discrete math class
<lain> all it took was one paper that actually used PICTURES and DIAGRAMS instead of "pure math" and it all made sense lol
<rqou> that was specifically covering e.g. finite fields and polynomials
<lain> yeah I didn't go to college
<lain> :P
<rqou> yeah when i was in HS i found this material impenetrable as well
massi has joined ##openfpga
<rqou> i don't understand why we can't have discrete math at the HS level
<rqou> iirc my HS administration explicitly vetoed a linear algebra course
<rqou> that a teacher was already stepping up to be willing to teach
<rqou> actually, clarification
<rqou> why can't we have discrete math other than the random shit on permutations that gets dumped in somewhere along the way
<rqou> and that everybody wonders "why do we need to learn this?"
<lain> calculus was a big barrier for me
<lain> but then I read Optics
<lain> and then a lot of calculus concepts just made sense, learned as I went along
<lain> because it put them in context
<rqou> hmm i personally found calculus easy enough
<lain> the one class I took on calculus was awful though
<lain> the professor taught it as rote memorization
<lain> "just plug this into that"
<lain> which
<lain> is not at all how2calculus
<lain> that's how2algebra, and maybe geometry, but...
<rqou> oh berkeley's single-variable calculus is terrible
<rqou> it's a weeder class
<lain> calculus is all about understanding how shit relates
<rqou> multivariable is slightly better
<rqou> i took ap calculus in HS and skipped straight into multivariable calculus
<lain> I didn't advance much in math in HS ... I took some AP courses but kept doing poorly because they were homework-heavy and ain't nobody got time for that
<lain> I had shit to do :P
<rqou> i did shit and ap courses in HS
<rqou> mostly because i didn't do the HS part very diligently :P
<lain> so they stopped letting me do AP courses eventually
<rqou> my $FANCY_SCHOOL grades are higher than my HS grades, which is _highly_ unusual
<lain> school in general just.. there's so many problems
<rqou> imho our K12 system sucks
<lain> our uni system sucks too
<rqou> i felt i benefited a lot from university though
<lain> the whole thing is crap :P
<lain> it's highly variable
<rqou> <politics, troll>but betsy devos will make it all better!</politics, /troll>
<lain> but I think ultimately the traditional model is going to break down
<lain> especially in tech fields
<rqou> tech fields and education have a really strange and complicated relationship it seems
<rqou> i wonder how it will go under the current administration's leadership?
<rqou> erm, "leadership"
<rqou> lain: i can definitely agree with "it's highly variable"
<rqou> when i was taking operating systems i got put with the "stragglers" project group
<rqou> in the end i ended up doing 99% of the debugging
<rqou> i learned a lot, but i don't know if the teammates did
<rqou> and this is $FANCY_SCHOOL, so $GENERIC_SCHOOL might be even worse
<lain> most of my university experience was people telling me I can't possibly know X or I can't possibly do X
<lain> at one point in CS101 (after being told by everyone that I can't test out of it, nobody ever can)
<lain> the prof was like "would you be my lab assistant and help the others, at least until I can get you the paperwork to test out of this class? you shouldn't be here"
<lain> and I was like lol sure >_>
<lain> he was a pretty chill dude.
<rqou> hmm at $FANCY_SCHOOL you can't test out of the first cs class either
<rqou> you can test out of data structures
<lain> iirc he even mentioned the administrator people gave him shit for getting me the test
<rqou> but our intro cs course is rather famously unique
<lain> this course was prettymuch "how2c"
<rqou> ours is a python port of MIT's famous "structure and interpretation of computer programs" course
<rqou> we didn't even have a "how2c" course :P
<rqou> very sink or swim
<rqou> er, we did actually
<rqou> it sucks horribly
<rqou> and isn't required
<lain> I was unnecessarily cocky back then :P I turned in the first homework in pen, the day it was assigned >_>
<lain> and he was like "how do you know it compiles?"
<lain> and I said "if it doesn't, I'll turn in the rest of the homework on the expected date, in the expected format, from now on."
<rqou> lool
<lain> (it compiled, and was apparently the best solution he'd ever had a CS101 student hand in)
<lain> but that's not fair
<lain> I had been coding for years
<lain> most of these people never saw code in their life
<lain> I tried to do the same thing in my intro EE class
<lain> oh man did that go poorly
<lain> it was a simple quiz on ohm's law
<lain> he handed out the homework sheet at start of class, and then lectured on ohm's law
<lain> I was familiar, so I filled it out while also listening to him, in case there was anything I didn't know
<lain> he *blew up* when I handed it to him though
<lain> he was like YOU DIDN'T PAY ATTENTION TO THE LECTURE
<lain> YOU CAN'T POSSIBLY HAVE CORRECT ANSWERS
<lain> YOU ARE GOING TO FAIL MY CLASS
<lain> boy did I get an ear-full that day heh
<rqou> wow
<lain> I prettymuch fucked right off and dropped out the next day
<lain> >_>
<lain> I was like nah
<lain> I don't have time
<lain> for that
<lain> this is a waste
<lain> I'm going to just make some PCBs
<lain> and learn shit on my own
<lain> <_<
<rqou> IME $FANCY_SCHOOL doesn't mind cockiness, but if you screw yourself you will get no sympathy
<lain> slight overreaction! but also fuck that guy
<lain> :3
<lain> I have humbled since then, but lol
<lain> I did not foresee that reaction
<lain> man he was fuming
<rqou> my dad had a great story from a (good) professor in one of his intro EE courses in college
<davidc___> heh; I didn't ever get to test outta stuff (first two years which is all I did wasn't really like that)
<davidc___> so I spent my C course figuring out how to write programs with only one statement
<lain> hahaha
<rqou> that professor doesn't normally teach classes
<rqou> and he was teaching an intro ee lab
<davidc___> (turns out that GCC has a repeatable behaviour with setjmp and longjmp in a single expression.....)
<rqou> so lab 1 starts and he asks everybody "ok, let me see your prelabs"
<lain> davidc___: oh god
<davidc___> Note I didn't say defined... just predictable!
<rqou> professor looks through all of them briefly and almost all of them suck
<rqou> professor starts grilling students "how do you expect to perform measurement X?"
<rqou> student: uh, idk?
<rqou> professor: "that's what the prelab was supposed to be for"
<rqou> and then the professor canceled the lab
<rqou> but my father had actually done the prelab properly
<rqou> and wasn't too happy about that
<rqou> the professor asks to see him after the lab
<rqou> and then the professor appreciated how he had actually done the prelab properly, but also meticulously went through and graded it very pedantically
<lain> (one time at one of the unis I attended briefly, my english teacher was a senile old lady and -- I shit you not -- she would often come to class asking for something she never assigned, and then accuse us of lying to get out of an assignment, and give us all F's)
<rqou> wtf
<lain> (in that same class, there were adults, people in their 30s and 40s, who... we were told to write an brief essay on our favorite piece of classic literature, and they wrote theirs on articles in Us magazine)
<lain> that was a quality university
<rqou> hmm i guess it's believable
<lain> very premium.
<rqou> my sister in HS had an english teacher that couldn't even fractions
<rqou> as in basic "how do i mess with fractions to appropriately tweak/compute grades"
<lain> lol
<lain> wow
<rqou> and she was the head of the english department
<rqou> on the other hand my HS science department had people that actually knew wtf they were doing
<lain> my HS physics teacher was awesome
<rqou> including "we got tired of the staff area copier being not working all the time, so we *gasp* bought our own copier"
<rqou> wow!
<lain> he learned physics to race cars
<lain> well, to better tune his race cars
<rqou> also, they installed a _lab-wide wifi network_ **with a password**!
<lain> lol
<rqou> my chemistry teacher in HS was a former civil engineer
<rqou> a lot of people in my HS didn't like that chemistry teacher though
<rqou> because she would go on tangents a lot when people who actually knew wtf was going on asked distracting questions
<rqou> people including me :P
<lain> haha
<lain> there was a math teacher at my HS, long before I attended there
<davidc___> we got a communications prof "removed from grading duties"
<lain> the chem teacher would tell a story about how this math teacher was a sucker for certain types of problems, and this was well-known in the student body
<lain> so students would write one of these problems on the chalkboard before he came into class
<davidc___> (blatant favouratism was most of the motive for her removal, but the thing that sold it to the dean was when she failed a dude that had _just_ finished a masters in english....)
<lain> and he'd come in, and before starting lecture he'd see this problem, and just start solving it and totally forget about class
<lain> and the students one time got a ladder
promach has joined ##openfpga
<lain> this class was on the 2nd floor, across the quad from the chem room
<lain> and they climbed out the window
<lain> into the quad
<lain> and all left
<lain> and then came back before he noticed they were gone
<lain> at least that's how the chem teacher tells it, he says he watched it all unfold :P
<lain> davidc___: lol
<rqou> lol
<rqou> my chem teacher was also known for relatively difficult exams
<rqou> but she had a policy (for ap chem) that if you scored a high enough score on the ap exam (vs your class grade) your grade would be raised
<rqou> but if you scored too low on your ap exam your grade would be lowered
<rqou> afaik the lowering of grades never/rarely happened, but raising happened all the time
<lain> one time I had a math teacher in HS state at the end of the year that you could turn in ANY homework you didn't do throughout the entire semester and she'd count it as being on time
<lain> what a mistake
<lain> two friends and I were up all night
<lain> we did basically every assignment, because we had not done any
<lain> she was SO MAD
<rqou> lool
<lain> she almost didn't accept it
<rqou> i had a math teacher that didn't even grade homework
<lain> but one of my friends was like "you gave your word!"
<lain> and she was like "dammit."
<lain> :P
<rqou> my father did that to in college for some intro class
<rqou> iirc it was an intro programming course
<lain> one of my friends would bring a pillow to math class and literally sleep through it
<lain> at first she woke him up and asked him really hard questions
<lain> at rnadom
<lain> but he always answered 100% correctly
<lain> and instantly
<lain> so she just gave up
<nats`> meanwhile in France University is "free"
<nats`> :D
<nats`> and engineering school are not that expensive :D
<rqou> <troll, trump>but you guys have radical islamic terrorists!</troll, /trump>
<nats`> there are even engineering school where you're payed to attends lesson
<nats`> it seems so fucked up to get years of debt to learn something :D
<lain> yeah
<rqou> <troll>but at least we don't have terrorists! we're going to make america great again!</troll>
<nats`> :D
<nats`> did you hear about teh jetl propulsion lab guy
<nats`> he was forced to give his phone to homeland security breaking rules of.... homeland security :D
<rqou> jpl guy who got his phone searched by cbp?
<rqou> yeah
<nats`> dumbass
<rqou> meanwhile i got waved through french border security with very minimal checking
<nats`> yep because it's not usefull
<nats`> you need to scare terrorist while on transportation to avoid big attack
<nats`> but once on the territory you can do nothing
<nats`> only hope the "spies" are efficient enough
<rqou> i've seen chinese border security checking less aggressively than what the US has been doing recently
<rqou> including one time where my father got called in for secondary inspection
<rqou> the inspecting guy looked and asked "is this your personal laptop?"
<rqou> "is this second laptop also your personal laptop?"
<nats`> TBH I think if a guy is not known from intel service previously of an attack there are like 0.01% chance to stop him
<rqou> "why did the guy outside send you in here? get out :P "
<nats`> if he already know I would let him enter the territory and clean the problem in private :D
<lain> I once got patted down and had my bags tossed :P
<lain> but
<lain> they did that to everyone that day, INCLUDING THE PILOT AND CREW
<lain> oh god they were pissed
<lain> I think it was TSA training day
<lain> they also interestingly lied to me about something
<rqou> one of my housemates had a great story about tsa security theater
<lain> they took me aside to swab my hands
<lain> I asked what it was for (honestly didn't know)
<lain> turns out it's a bomb residue test
<lain> but the guy said "it checks for sweat, to test if you're nervous. you look nervous."
<rqou> as he was going through security, he notices tsa isn't paying much attention
<nats`> lol
<rqou> turns out the tsa agent was talking with his buddy about how to max out his 401k account (retirement savings account for non-US people)
<lain> lol
<rqou> :P
<nats`> and someday all those bomb residue test will be useless
<rqou> when i was traveling with my "big box of scary electronics"
<rqou> i got meticulously x-rayed at hamburg
<rqou> when i was entering the us i preemptively tell the security guy that it's full of random electronics
<rqou> security guy ignores me, making an "i don't give a damn" impression
<rqou> swabs it with the bomb residue test
<rqou> comes back negative
<rqou> and just tells me to go away :P
<lain> haha
<rqou> so i guess we only need the machine
<nats`> look at the wikipedia article
<nats`> hilarious CNN interview of this bitches from white house
<rqou> nats`: so what's the overall opinion of trump over in FR?
<nats`> I don't know but we have worse than him here
<nats`> all people I know are against brexit trump/poutine/marine lepen
<nats`> but I can't say for general view, it's a pretty messy era
<nats`> I have one problem with trump
<nats`> unlike poutine and other disgusting people he's motivated by something totally irrational
<nats`> and that's frightening
<nats`> it means the guy having the Bomb in hand is a fucking crazy rat
<nats`> 99% of what he says is motivated by his own ego being injured
<nats`> look at this bathrobe story
<nats`> who in the world who loose time to make an official statement from the white house to deny he is wearing bathrobe
<davidc___> lain: also drugs
<rqou> i thought our drugs came from mexico, hence why we need a wall to keep the bad hombres out :P :P
<lain> hm?
<lain> oh
<lain> davidc___: interesting
<nats`> for hair ?
<nats`> that's what you're talking about
<davidc___> just residue
Hootch has joined ##openfpga
Bike has quit [Quit: slep]
eightdot_ is now known as eightdot
rvense_ has joined ##openfpga
jix_ has joined ##openfpga
felix___ has joined ##openfpga
jix has quit [*.net *.split]
felix_ has quit [*.net *.split]
rvense has quit [*.net *.split]
trap15 has quit [*.net *.split]
trap15 has joined ##openfpga
<rah> ^ talks about open source FPGA tools
<rah> '"Call to Arms : Volunteer Wanted" University Consortium to develop open simple FPGA HW and SW'
<rqou> huh even I didn't know this PowerPoint existed
<rqou> it's pretty old
<rqou> and also I took FPGA design taught by Prof. Wawrzynek :P
<rah> rqou: oh really? :-)
<rah> small world :-)
<rqou> yeah I went to Berkeley
<rah> cool
<rah> he sent me that PDF but it was online previously (now 404 apparently)
<rah> "You'll see that my motivation was to create a simple hardware model so that we could foster a community tools effort. However, at Berkeley it was difficult to get it off the ground because the initial software effort required to map designs to our chip was too much. In the history of FPGAs, the tools always end up being the more difficult to get right. I still believe in the vision. Just need to find the right student(s) to help work on it. Maybe some
<rah> rqou: when were you at Berkeley? :-)
<rqou> fall 2012-spring 2016
<rah> plainly Prof. Wawrzynek is not being loud enough :-)
talsit has left ##openfpga [##openfpga]
kuldeep has quit [Remote host closed the connection]
kuldeep has joined ##openfpga
<rah> rqou: were you aware that Wawrzynek was interesting in open FPGA tools when you were at Berkeley?
felix___ is now known as felix_
<felix_> http://moscow.sci-hub.cc/761a234419347d6c0719b6f0d9b7fdd1/10.1109@ASPDAC.2005.1466599.pdf was developed at ntua and has a full design flow. sadly the pages only return 404
cr1901_modern has quit [Ping timeout: 276 seconds]
cr1901_modern has joined ##openfpga
pie_ has joined ##openfpga
pie_ has quit [Changing host]
pie_ has joined ##openfpga
openfpga-bb has quit [Ping timeout: 240 seconds]
azonenberg has quit [Ping timeout: 256 seconds]
azonenberg_work has quit [Ping timeout: 256 seconds]
openfpga-bb has joined ##openfpga
azonenberg_work has joined ##openfpga
azonenberg has joined ##openfpga
m_w has joined ##openfpga
promach has quit [Quit: Leaving]
digshadow has quit [Quit: Leaving.]
Bike has joined ##openfpga
mifune has joined ##openfpga
mifune has joined ##openfpga
amclain has joined ##openfpga
eduardo_ has joined ##openfpga
eduardo__ has quit [Ping timeout: 240 seconds]
kkiningh has joined ##openfpga
<azonenberg> rqou: interesting
massi has quit [Quit: Leaving]
pie_ has quit [Ping timeout: 256 seconds]
digshadow has joined ##openfpga
nickjohnson has quit [Remote host closed the connection]
pointfree has quit [Remote host closed the connection]
dingbat has quit [Remote host closed the connection]
nickjohnson has joined ##openfpga
dingbat has joined ##openfpga
kkiningh has quit [Ping timeout: 260 seconds]
pointfree has joined ##openfpga
Hootch has quit [Read error: Connection reset by peer]
dingbat has quit [Quit: Connection closed for inactivity]
mifune has quit [Ping timeout: 245 seconds]
digshadow has quit [Quit: Leaving.]
digshadow has joined ##openfpga
rvense_ is now known as rvense