|
as my career drags on, i notice that some of the most infuriating computer disasters seem to be created by people who couldn't count higher than one for example, there is Dr. Honked's story of the drugged-out coder who kept only one source file on his computer, compiled it, then cleared it so he could write more code some other examples from my experience: - my employer's technology was implemented redundantly across three languages. to streamline development, parts of it were ported to javascript and executed with runtimes for those respective languages. but in the process of porting, our architect dumped several classes of code into one javascript file (and not as javascript classes, btw). it stayed that way for over a year until we started shipping bugs because the file was so unwieldy. at that point i figured nobody could stop me from dividing it up, and did so. it still has a further to go but nobody will ever bother. - there was a project to gather telemetry from the product to support billing. part of the design was a server to receive the telemetry information. the dev who worked on the server side delivered an application that ran (web server and database) on one machine, to support all of our customers and their millions of end-users. the project never saw daylight, tg - another web service, much less complicated than the telemetry server, was put into production and supports customers from one machine per continent. i don't work on that one but this week i'm hearing rumors that a large customer's traffic is killing it, everyone involved on our side (esp. the deployment team) is just shocked and doesn't know what to do - there's an internal web UI that, as we were bringing in a second coder to work it, the original dev quit without notice. our current candidate wasn't very good but we hired him in the pinch and left him to work on the UI alone. there was no effective code review and he ignored whatever feedback we gave, so naturally he made the biggest mess you ever saw. but one of his tasks was to add unit tests to the UI. And he did -- all in one file spanning thousands of lines, cut and pasted all to hell post ur examples of disasters involving "one"
|
# ? Jun 2, 2018 22:25 |
|
|
# ? May 4, 2024 12:23 |
|
|
# ? Jun 2, 2018 23:10 |
|
didnt read voting 1 as directed
|
# ? Jun 2, 2018 23:22 |
|
pram posted:didnt read voting 1 as directed
|
# ? Jun 2, 2018 23:25 |
|
guess i should have stopped with one example, i expected too much from yospos
|
# ? Jun 2, 2018 23:35 |
|
remember when microsoft almost called it windows one
|
# ? Jun 2, 2018 23:44 |
|
kinect
|
# ? Jun 3, 2018 03:19 |
|
my XBONE Kinect is a piece of poo poo.
|
# ? Jun 3, 2018 03:25 |
|
Silver Alicorn posted:remember when microsoft almost called it windows one youll find thats Windows 2
|
# ? Jun 3, 2018 03:26 |
|
Arrays in R
|
# ? Jun 3, 2018 15:39 |
|
pram posted:kinect https://www.youtube.com/watch?v=fgJqliVWSSg&hd=1
|
# ? Jun 3, 2018 15:50 |
|
quote:post ur examples of disasters involving "one" In the first phase of a software project, implementing high availability would be serious over engineering / YAGNI crap. Then there's the next phase, where the project is too complicated to implement HA without massive, development halting refactors. Somehow I've never seen these two phases fail to overlap by anything less than six months.
|
# ? Jun 3, 2018 15:52 |
|
What is the correct number of successful tests to run before calling a critical task 'automated' and cutting headcount? Apparently in my org, it's one.
|
# ? Jun 3, 2018 17:21 |
|
Lol
|
# ? Jun 3, 2018 17:26 |
|
Agile Vector posted:youll find thats Windows 2 Good joke. 1010/1010
|
# ? Jun 3, 2018 18:13 |
|
the matrix, chugged along fine with some minor glitches until it got ruined by some guy named the "1"
|
# ? Jun 3, 2018 23:34 |
|
well BAM there it is
|
# ? Jun 3, 2018 23:38 |
|
idk what this is about kinect, does it have anything to do with unscalable software development or is it just some lame trademark joke
|
# ? Jun 3, 2018 23:55 |
|
kinnect ruled for the homebrew stuff people made. some of it went off to museums, and used at artists shows
|
# ? Jun 4, 2018 00:11 |
|
worth the money then. for the art work.
|
# ? Jun 4, 2018 00:44 |
|
tbh at an absurdly late stage of my career i still have no idea how much traffic one machine can serve because i've worked at places that either (1) deploy across multiple systems by default and don't allow devs to hand off unscalable code for any reason or (2) just throw code onto one machine and never communicate back anything about QoS
|
# ? Jun 4, 2018 00:53 |
|
I laugh every time I hear 'our production machine is down'. I also enjoy when some POS cluster software is completely unusable because it lost one cluster member and had a cascading failure.
|
# ? Jun 4, 2018 01:11 |
|
Gazpacho posted:tbh at an absurdly late stage of my career i still have no idea how much traffic one machine can serve because i've worked at places that either (1) deploy across multiple systems by default and don't allow devs to hand off unscalable code for any reason or (2) just throw code onto one machine and never communicate back anything about QoS 1
|
# ? Jun 4, 2018 01:27 |
|
|
# ? May 4, 2024 12:23 |
|
1 git repo containing 15 different projects all deemed business critical by the business with no tests, build pipeline or code reviews done.
|
# ? Jun 4, 2018 15:26 |