Me: Optimize a sort & match method in backend because users complain it's a bit slow.

Coworker: These algorithms are both O(n), so they're identical *closes PR*

Me: *start zoom call* "Heeeeeeeeeey Iiiiiiiiiii wouuuuuuuld liiiiiiiiike toooooo diiiiiisscuuuuus thaaaaaaaat puuuuuuulllll reeeeeequuuueeest yooooouuuuu cloooooossseeeed"

Coworker: "wtf are you doing, why are you talking so slow"

Me: "No matter whether I talk fast or slow, the information still reaches you in O(n) time, so why are you complaining"

I fucking hate it when people misunderstand the purpose of (or abuse) big O notation. It's an estimate of how an algorithm SCALES once the set increases in size, in which case you leave out both less significant terms and constant factors.

But those terms and factors are important when you're talking about the DIRECT PERFORMANCE of the algorithm on fixed-size sets, instead of SCALING to larger sets.

1n and 10n are both O(n), but 10x performance on a job that used to take 10 minutes is still significant.

When I was in 7th grade, my neighbor (a DoD programmer) challenged me to write a sorting algorithm for a hypothetical super limited environment (he said a satellite). It didn’t have any built-in sorting methods, had very limited memory, slow processor, etc. so I needed to be clever about it.

It took me a few nights before i found a solution he liked. The method I came up with counted the number of occurrences of each number in the array and put them in the appropriate spots in a new array. This way it only required O(2n) running time and 2n memory.

I just learned today that this is called the “counting sort” 😄

I’m proud of little 11 year old me.

Just thought I'd share my current project: Taking an old ISA sound card I got off eBay and wiring it up to an Arduino to control its OPL3 synth from a MIDI keyboard. I have it mostly working now.

No intention to play audio samples, so I've not bothered with any of the DMA stuff - just MIDI (MPU-401 UART) and OPL3.

It has involved learning the pinout of the ISA bus connectors, figuring out which ones are actually used for this card, ignoring the standards a little (hello, amplifier chip that is wired up to the +12V line but which still happily works at +5V...)

Most of the wires going to it are for each bit of the 16-bit address and 8-bit data. Using a couple of shift registers for the address, and a universal shift register for the data. Wrote some fairly primitive ISA bus read/write code, but it was really slow. Eventually found out about SPI and re-wrote the code to use that and it became very fast. Had trouble with some timings, fixed those.

The card is an ISA Plug and Play card, meaning before I could use it I had to tell it what resources to use. Linux driver code and some reverse-engineering of the official Windows/DOS drivers got me past this stage.

Wired up IRQ 5 to an Arduino interrupt to deal with incoming MIDI data, with a routine that buffers it. Ran into trouble with the interrupt happening during I/O and needing to do some I/O inside the handler and had to set a flag to decide whether to disable/re-enable interrupts during I/O.

It looks like total chaos, but the various wires going across the breadboard are mainly to make it easier to deal with the 16-bit address and 8-bit data lines. The LEDs were initially used to check what addresses/data were being sent, but now only one of them is connected and indicates when the interrupt handler is executing.

There's still a lot to do after that though - MIDI and OPL3 are two completely different things so I had to write some code to manage the different "channels" of the OPL3 chip. I have it playing multiple notes at the same time but need to make it able to control the various settings over MIDI. Eventually I might add some physical controls to it and get a PCB made.

The fun part is, I only vaguely know what I'm doing with the electronics side of this. I didn't know what a "shift register" was before this project, nor anything about the workings of the ISA bus. I knew a bit about MIDI (both the protocol and generally how the MPU-401 UART works) along with the operation of a sound card from a driver/software perspective, but everything else is pretty new to me.

As a useful little extra, I made some "fake" components that I can build the software against on a PC, to run some tests before uploading it to the Arduino (mostly just prints out the addresses it is going to try and write to).

I think I'm falling in love. With TDD.

I used to be very skeptic about it. You know, the usual reasons: it takes longer to deliver, constant "flow" interruptions, etc, etc. But ever since I've tried it I'm nothing but happy about my choice :)

I'm moving forward, I'm not making any regressions, I'm no longer afraid to make any changes in my code as I know tests will show what exactly I break,.. And most importanty, I have all use-cases with corner-cases defined and "explained" in the code... No more do I have to search in Confluence for how this exact scenario should behave. Everything is here. Everything's in the tests.

It's amazing!

Yeah, it DOES take longer to deliver so if you're hardcore Agile living by "Ship it as soon as it compiles" TDD might be too slow. But if you prefer knowing when your code is covering all the use cases w/o any errors -- TDD is the way.

A few weeks ago, I was kept up until the wee hours of the morning trying to figure out how in the hell the Monty Hall problem works. After finally getting it (I'm slow, okay?), I decided to write a program to run simulations of it.

First incarnation of program took user input. User enters what door they choose (1, 2, or 3), then is told what door Monty opens, then given the decision of staying with the door they originally chose or switching, then informed how that worked out for them.

Second incarnation of program ran on a loop. At the start of each loop, a random door is picked for the user guess. Then the door Monty opens is calculated from the remaining doors (excludes user guess and prize door). Then user switches doors (choosing the door that was not their original door or the door Monty opened). At the end of each loop, if the door they switched to was the prize door, it would increment a win counter, else increment a loss counter. After running the loop 1000000000 times, it printed to console `You always switched doors, resulting in ${wins} wins and ${losses} losses`.

THEN I decided to write a variation to run a while loop on the outside of the loop to increase the number of total doors until the point where the decision to switch doors hurt more often than it helped. At this point, I decided to incorporate file I/O and write to a file rather than a console. And that was neat!

And then I decided it would be cool to go back to the three door variation, printing on each loop the original door, the door Monty opened, the door that was switched too, the result of the switch (win or lose) and what the prize door was.

But for the life of me, I couldn't seem to get the file to write properly. It would, like, always crash my terminal. I tried open + append, I tried append. I tried createWriteStream. Still just failure.

And then I changed it to an appendFileSync and happened to look at one of the files that I was writing to. "Huh, over a gig seems a lot."

"Well, how much are you writing each loop? Did you forget to keep in mind how many bytes that would be?"

TLDR: If you're going to write a program that's going to write data to a file on a loop, you might want to figure out how much it's going to end up writing .... before trying to run it. And running a loop 1000000000 times may be a little excessive.

*face palm*

I want a boring software developer job. I’ve been working for software consulting companies since the beginning. And is just so stressful. Clients always ranting, the need to always be in the cutting edge, or even the complete opposite. There’s always pressure to get certified in X o Y. I don’t want that anymore. I don’t want to be constantly catching up with the latest stack or framework. I want a boring job. A slow-paced job maybe maintain some old hunk of software that does not give too much trouble. I’m tired of putting down fires all the time. Of running against the clock to deliver a meaningless app. Because all this apps don’t contribute to anything in the world. Just more clutter, more bloat. I just want to work 8 to 5 and be done with it. Just throw myself in the couch after it and play some games. Maybe do some gardening. Or bread. I love bread. Don’t you love bread?

So recently I installed Windows 7 on my thiccpad to get Hyperdimension Neptunia to run (yes 50GB wasted just to run a game)... And boy did I love the experience.

ThinkPads are business hardware, remember that. And it's been booting Debian rock solid since.. pretty much forever. There are no hardware issues here. Just saying.

With that out of the way I flashed Windows 7 Ultimate on a USB stick and attempted to boot it... Oh yay, first hurdle to overcome. It can't boot in UEFI mode. Move on Debian, you too shall boot in BIOS mode now! But okay, whatever right. So I set it to BIOS mode and shuffled Debian's partitions around a bit to be left with 3 partitions where Windows could stick in one more.

Installed, it asks for activation. Now my ThinkPad comes with a Windows 7 Pro license key, so fuck it let's just use that and Windows will be able to disable the features that are only available for Ultimate users, right? How convenient would that be, to have one ISO for all the half a dozen editions that each Windows release has? And have the system just disable (or since we're in the installer anyway, not install them in the first place) features depending on what key you used? Haha no, this is Microsoft! Developers developers developers DEVELOPERS!!! Oh and Zune, if anyone remembers that clusterfuck. Crackhead Microsoft.

But okay whatever, no activation then and I'll just fetch Windows Loader from my webserver afterwards to keygen my way through. Too bad you didn't accept that key Microsoft! Wouldn't that have been nice.

So finally booted into the installed system now, and behold finally we find something nice! Apparently Windows 7 Enterprise and Ultimate offer a native NFS driver. That's awesome! That way I don't have to adjust my file server at all. Just some fuckery with registry keys to get the UID and GID correct, but I'll forgive it for that. It's not exactly "native" to Windows after all. The fact that it even has a built-in driver for it is something I found pretty neat already.

Fast-forward a few hours and it's time to Re Boot.. drivers from Lenovo that required reboots and whatnot. Fire the system back up, and low and behold the network drive doesn't mount anymore. I've read that this is apparently due to Windows (not always but often) mounting the network drive before the network comes up. Absolutely brilliant! Move out shitstaind, have you seen this beauty of an init Mr. Poet?

But fuck it we can mount that manually after every single boot.. you know, convenient like that. C O P E.

With it now manually mounted, let's watch a movie! I've recently seen Pyro's review on The Platform and I absolutely loved it. The movie itself is quite good too. Open the directory on my file server and.. oh. Windows.. you just put db.thumb on it and db.thumb:encryptable. I shit you not, with the colon and everything. I thought that file names couldn't contain colons Windows! I thought that was illegal in NTFS. Why you doing this in NFS mate? And "encryptable", am I already infected with ransomware??? If it wasn't for the fact that that could also be disabled with something as easy as a registry key, I would've thought I contracted ransomware!

Oh and sound to go with that video, let's pair up some Bluetooth headphones with that Bluetooth driver I installed earlier! Except.. haha nope. Apparently you don't get that either.

Right so let's just navigate the system in its Aero glory... Gonna need to flick the mouse for that. Except it's excruciatingly slow, even the fastest speed is slower than what I'm used to on Linux.. and it's jerky as hell (Linux doesn't have any of that at higher speed). But hey it can compensate for that! Except that slows down the mouse even more. And occasionally the mouse driver gets fucked up too. Wanna scroll on Telegram messages in a chat where you're admin? Well fuck you mate, let me select all these messages for you and auto scroll at supersonic speeds! And God forbid that you press delete with that admin access of yours. Oh maybe I'll do it for you, helpful OS I am!

And the most saddening part of it all? I'd argue that Windows 7 is the best operating system that Microsoft ever released. Yeah. That's the best they could come up with. But at least it plays le games!

Story Time!

Tittle: About Larry.

Fun Game: Tell me if / when in this story you know the plot twist.

Setting: Years ago, non coding job.

I work with Larry a lot, Larry works remote. In technical terms Larry is senior to me and I escalate some technical issues that get assigned to Larry. I've never met Larry in person.

Larry can be hard to work with, but he's plenty good at his job and I don't mind his prickly side. Sometimes it takes telling Larry something a few times before it sinks it, but that's not a big deal. Sometimes it seems like Larry doesn't remember his cases entirely, but he has a lot of cases. Also Larry has good reason for how he works considering the land of scubs who usually escalate to him without any thought / effort.

Larry's escalation team is short staffed and they're trying to hire folks, but that's been like that forever.

So one day I get an email that Larry is going to be out of the office for a few weeks. Nothing unusual there.

My current case that I share with Larry sort of floats in limbo for a while. The customer is kinda slow to respond anyhow and there's nothing that I need Larry for.

Finally I get automated notice that my case has had a new escalation engineer. Laura. Laura is much more positive and happy compared to Larry. Understandably Laura isn't up to date on the case so we go back and forth with some emails and notes in the case.

The case is moving along just fine, we're making progress, but it's slow because of the customer's testing procedures. Then we hit a point where this customer's management pushes on sales for a solution (this customer's management is known for doing this rando like for no reason).

Down the management chain it goes and everyone wants a big conference call to get everyone up to date / discuss next steps (no big deal).

Now I really don't want to do this with Laura and throw her into the deep end with this customer, she doesn't have the background and I'd rather do this call with Larry & Me & Laura. Also according to the original email Larry is due back soon.

I start writing an email to Laura about "Let's try to schedule this for when Larry gets back."

Then I stop ... I don't really know why I stop but when it is a "political case" I want some buy in on next steps from management so I go talk to my manager.

-Plot Twist Incoming-

Long story short, my manager says:

"Laura IS Larry..."

O

M

G

I had no idea. Nobody told me, nobody told ANYBODY, (except a couple managers).

Back up a few months Larry apparently went to his managers and told them he was going to transition, surgery and all, in a few months.

Managers wondering how to address this went to HR and some new hire very young to be a manager HR manager drone logiced out in her bonkers head that "Well it shouldn't matter so don't tell anyone."

ARE YOU FUCKING KIDDING ME!!??

Thank god I didn't send that email...

I did send an email to Laura explaining that I had no idea and hoped I didn't say anything stupid. She was very nice about it and said it was all good.

After that incident made the management rounds (management was already fuming about being told not to tell anyone) things came to another critical point.

Laura was going to visit the company HQ. Laura had been there before, as Larry, everyone knew her as Larry... nobody (outside some managers) knew Laura was Larry either. With nobody knowing shit Laura was going to walk in and meet everyone ...

One manager at HQ finally rebelled and held a meeting to tell his people. He didn't want Laura walking in and someone confused, thinking it was a joke or something horrible happening.

HR found out and went ballistic. They were on a rampage about this other manager, they wanted to interview me about how I found out. I told HR to schedule their meeting through my manager (I knew they didn't want my manager to know they were sniffing around).

Finally the VP in our department called up the HR head and asked WTF was going on / kind of idiots they had over there (word has it legal and the CEO were on the call too).

HR had a change in leadership and then a couple weeks later there were department wide meetings on how to handle such situations and etc.

A wild project appears!
The deadline is set in two months.

It's a 3D environment interactive app with some oil drilling models and other stuff, for a stand on a show. It needs to look nice, but The Company we're working for needs to figure out where the fuck their product is located on those machines. Think tiny pipes, O-rings etc.

I prepare a build in the first couple of days for The Company to figure shit out.

Management holds the build back because:
> the ocean waves are going the other way
> the underwater area doesn't look so nice
> the antialiasing could be better
> one pipe is 5cm off center
> the sky is not blue enough
> the drillship propellers are pointed the wrong way
> one icon is too far to the right
> the shadows could use some work
> there are shadows on the seabed
> some flickering on ambient occlusion
> it loads too slow
> one random object is flipped on it's Z axis
> it's too green
> camera locks up if you move about 2km out of the range
> the name of the build should represent the date of the build
> the name of the build SHOULDN'T be anything else than just a simple three-word name, no dates because their environment doesn't allow apps that are not allowed (by name) by admin
> lots more random things that won't prevent them from using the app

I'm only a month late, but it's good progress. In about a week I hope we can get some feedback if we can use those models at all and what to showcase.

Then I can work on the basic functionality. And then it's a simple case of time travel to meet the deadline.

Reminder that the best way to speed up tar is to not use the 'v' option.

You know what I really like? When apps override the scrolling behaviour...
Yes, I really want you to disable my scrolling settings and enforce your own goddamn scrolling logic, it's so great to have shit I'm used to being forcibly removed from me.
Biggest surprise is the latest place I've seen this is Unreal Engine :O Scrolling the news in the launcher is slow as fuck.

How do you explain to your client that no, you cannot have a perfect solution, because the algorithm is O(2^n)?

I mean, without requiring him to get a degree in CS. Also without making him think that you can't build efficient code because you're dumb. Or that the hardware is slow.

New upgrade of kernel 4.12 just awesome but startup is very slow. I am experiencing this shit in my centos7. Overall there is an improvement in I/O. Loving new kernel.

I was hoping it would be possible in a big international company to work (as a software developer) on my own laptop (MacBook Pro) - cause of better parameters = better performance = better efficiency. After I got hired, I was told that it is not possible to bring my own laptop. So I was given an old DELL laptop with Windows + a lot of security stuff in it from the company. The poor DELL is so slow - that even a single commit into the branch takes about 2 minutes because of the security stuff : -O ...I am soooo disappointed... :[ .... On the other hand, by working at home on my MacBook in compare with that DELL I feel about it like I work with some super ultra alien technology from the future :D what a feeling <3

Many smartphone cameras lack the ability to turn off burst shot mode.

The burst shot feature on smartphone camera software is almost always not helpful, only annoying. All it does is spam the storage with useless near-duplicate photos.

"Then simply don't hold the camera shutter button!"

Sometimes, this happens by accident. Or the phone has an I/O lag in the moment of releasing the shutter button, so the release of the shutter button is not registered and burst mode is initiated after the I/O lag.

The only purpose of burst shot seems to be making many low light photos to find one that is not shaken. Even then, there must be an option to turn it off.

Also, the point-and-shoot intuition of holding the camera shutter button to set focus and exposure, and releasing to capture a photo is far more convenient. On newer phones, that has been replaced with highly annoying burst shots.

"Then use a third-party app that does allow turning off burst mode."

The problem with third-party applications is that they are awfully slow, since they can not be optimized for a specific device like pre-installed camera applications are. This slowness, as one might expect, leads to missed moments.
On some smartphones, third-party applications can not even access all camera features, such as 2160p video recording. Some phones use a proprietary API that can only be accessed with the pre-installed camera app.