In Oct’23 a basic Text widget with Subprocess to run shell commands allowed me to take my command line fu into a different arena because Tkinter gave me special tricks. This tool (shell functions) allows me to type up GUI apps or any Python script with less effort, for my fingers and brain.
wc newide; ksh newide | wc
513 1671 11018 newide in pnk
604 1987 29867 newide in python
The blocks of color are capital letters colored using tkinter methods (tag_add & tag_configure) with a bg and fg of the same color to make it look like a lego block, it’s all ASCII. On the right, an idle clicker game/toy made with pnk.lang, also just ASCII/UTF-8. The IDE you see in the pic also was coded using functional shell language I call pnk.lang and the original first iteration of this specification is in the legacy folder in the repo below. Just me learning how to code faster in Python but in shell but in neither at this point.
Don’t be put off by Ksh because Bash can also interpret it, since it’s just shell functions that print Python code. But I have plans to use an Xbox controller so I can move away from building with the keyboard at piecemeal rates. I will do this by abstracting away identifiers and all data we use as engineers into “dictionaries” to then transform it into a spatial system. I already done this first step! Using a basic grid with a maximum of 10 rows x 3 cols, indexing row[0-9] col[0-2], as you see there on the left side, and by using a letter I then categorize the functions of pnk(shell+python) so I can have 30 x 26[a-z] available slots; a00 b11 k22 and so on…by making it a visual shortcut that reduces cognitive load and typing for me.
Take a closer look at my repo without dismissing it too quickly. It could seem unnecessary but maybe someone else can see what I have made as useful and how we gonna take it to higher levels of abstraction and create a new realm for making computer applications in an abstract game/IDE of art and code. Perhaps it’s that creator effect that happens when you make something for the first time that makes me see its future utility and appeal. All this was made incrementally using my own software I built from scratch. I do have a great vision and would love to speak to anyone who is interested. I also demonstrate the utility of this small tool on youtube[link in repo]. Thanks.
I write python in shell. Literally. I made a shell function file (pnk.lang) that prints correct python code. The capital letters in the shell function’s name, my Tkinter IDE turns into color as you see above. So I let the color (capital letter) abstract away indentation for my brain. The second letter is for category of python code according to me. Again using color to allow me to think. I dont like Python syntax. I looks ugly to me. So I made this tool to make coding colorful, succint and fun for me. It reduces my eye strain among many other personal benefits.
I’m still annoyed with how verbose Objective-C is. Just check out what one has to do to create and concatenate a string. Madness:
<span style="color:#323232;"> NSString * test = [[NSString alloc] initWithString:@"This is a test string."];
</span><span style="color:#323232;"> NSString * test2 = [test stringByAppendingString:@" This value is appended."];
</span>
And god forbid you want to concatenate two things to a string:
<span style="color:#323232;"> NSString * test3 = [test1 stringByAppendingString:[test2 stringByAppendingString:@" Adding a third value."]];
</span>
Since the health is a float, yeah, it can create issues. A health of 0.000000001 is greater than zero, but that would almost assuredly be displayed to the user as simply 0, causing player confusion. The easiest solution is to have health and damage always be integers. A less great solution is to use a non-floating point decimal format. If such doesn’t exist in your language, you can emulate one by having health and damage both always be integers, but move the decimal point over, say two points, when displaying it to the user.
It sounds like the only concern you have with code is its bugginess, which is short sights. This is unfortunately better documented code than stuff I’ve seen in production. Obviously no one should do this, but let’s not act like there’s no benefit
Indeed, you can achieve a better result with less verbose naming convention. And choose better variable name to make it obvious than 0 Hp is death. While i don’t like having too verbose variable name (as it impacts the readability and quick understanding of the function), i’m not against that for the function name… without going too far of course!
Best is too have proper datamodeling of the object manipulated on top of some classic basic comments. Good interface contract is also a minimum. Best is to have full datamodeling of all the services, objects, in and out interactions between them, etc.
Documentation is a mandatory piece of the code delivery (with tests being the other important part) far too much forgotten if you don’t enforce it on your teams.
I decided to be wrong because the correct joke would be too convoluted. I’ll work on that implementation and then you can inject it at runtime via reflection.
Thanks for preparing your comment for my dependency injection! I agree that refactorability of comments is preferable over prematurely optimizing for performance.
Forever Ago I Ran A Minecraft Server And 2 Friends Joined And One Typed Everything They Said Like This and the other managed to misspell every single word with more than 2 letters in it. They misspelled the word “the!” According to Sir Capitalizes Every Single Word, its just much easier to type that way, which raised far more questions than answers…
I turn off LSPs during my train of thoughts. I don’t want all red and yellow underline bullshit to disrupt my thoughts. Like, calm the fuck down. I WILL write the correct code eventually; just give me some fucking time.
Well, I use Neovim, so turning off the LSPs or restarting them is sufficiently simple.
When I work on a new project, or on a new feature, I temporarily turn off the LSP, and rely on the compiler to figure out where the code errors. Plain white text gives me the freedom to write whatever the fuck I want without any disruption. Of course, I eventually turn on the LSP again to fix the little issues.
I also leave out little syntax errors and only only focus on the rough idea during my train of thoughts. And the variables, aren’t really supposed to be implied as private or unused – I do eventually meaningfully use them. If I have to prefix all my variables with a underscore to avoid the LSP, I might instead just disable the LSP. When I eventually turn the LSP back on, it tells me the actually unused variables and imports that I can now get rid of.
Because of the LSP, I used to write maybe three hundred lines of code per hour, but now I probably average at least five hundred or more.
programmer_humor
Newest
This magazine is from a federated server and may be incomplete. Browse more on the original instance.