How do I force myself to stop caring and just complete a project?First off there is code quality>tell Big Pickle to write js>it works, but it's written poorly>arrow functions everywhere>everything in one js file>bad variable names>not DRY enoughThen there is the scope creep>image search using hashes>oh how about a color filter>oh how about a nsfw filter>width, height, file sizeThen there are performance optimizations>0.5s queries? I want millisecond queries>render time is 0.5s? I want milliseconds for that too.All this makes a hobby project extend from a day into 1 or 2 months
this dude using big picle lmao
>>109225063The key is to overcome your pron addiction. Also, you should read about "intrinsic motivation vs extrinsic motivation". If you don't do this stuff on your own will, it's doomed to fail.
>>109225063Just do it
>>109225681this, once you're enough into a project you can spend entire days without thinking about anything else. earlier this year i built a simple 3d engine from scratch, using xlib and glibc.i would easily work 16-18 hours a day, every single day for weeks because I really got caught in it.also you should add side challenges to make it more interesting and learn stuff. for this project I decided it had to compile using -wconversion and pass a retarded linter that made my code looks funny. I unironically learned a lot of good patterns that increase code quality
>>109226342>decided it had to compile using -wconversionAll code should compile cleanly with all warnings enabled, if it doesn't it isn't clean code.
>>109226665>all warningsdefiine all warnings-wall + -wextra + -wconversion is only a small subset of all warningsI bet 99.9% percent of your code doesn't compile with all warnings because most warnings are irrelevant, conflict with each others, or are mutually exclusive.for instance, in C, enabling all warnings means your code must conform to c++98, even if you use c23 because there are people who need to write C code that can be compiled against C++ code. similarly, some warnings force you to write headers that conform with c89 because people using that version of c may want to link against your library/programit also means you must always declare function prototype before function definitions.there are also warnings when you're structs are laid out in such a way that makes access slow on specific architectures, if you fix it, it will now complain that it is suboptimal for another architecture. you can't win in this situation.clang even has warnings when comments above functions/structs aren't proper doxygen formatted
>>109226342>a simple 3d engine from scratchIs the code available?I wanted to do one too at some point, looked at many projects to see how they did the rasterization and the edge functions, most use either floats or int32 with 4 bits for the fraction, none of them use 6 or 8 bits how the hardware does.
why filename is base64 or whateverstop putting embed in my browser cache
>>109226342That's pretty cool anon, good job!
>>109225681>>109229415While I'm at it: I had great success with pomodoro timers. Just sitting down for 30 minutes and do something greatly helps you to follow through on it.
>>109225063do you consume nicotene? last time i quite i was programming for weeks straight like 12 hours a day
>>109229916>>109225681>>109225435I wrote my post poorly.I wanted to inquire about getting a quick MVP versus taking ones precious time to craft (over-engineer?) their projectsWhich approach do you prefer?When there is a chance I make things public facing, or release code to the public, I want good code. Performant code. And with enough features.
>>109230137I plan things for the best version I can do but if there are some problems I make the part that troubles me in the easiest form I can think of. No one will go to your first release and laugh about how much worse it is than what is available now
>>109230137>Which approach do you prefer?Chino
>>109230137i dont really plan things i have a rough idea in my head if i get stuck i break it down into a bullet point list of things i need to implement then just work through that list
>>109230208>>109230194does it often take you 10x longer than what you initially expected?I'm looking at an image search project where I had most of the tech developed. I thoughts it would take 2 days to polish, but I was way off. I'm looking at 10 more days (I only have 4 hours for this per day)
>>109225063>Python
>>109230258I was waiting for this
>>109230243Depends on how much experience I got in the field. For example when trying out a new programming language it can take much longer simply because I wanna take advantage of some of the new features
>>109230243things always take me longer than i think kek
>>109226342NICE! How does this work? Is it OpenGL/some other api? Or did you do your own rasterizer from scratch?
>>109226342>even casting enum values to the enum typeWhat the hell.>>109226665False. Compiler devs add new warnings all the time. Some useful, some retarded. It's why it's common knowledge that -Werror is a shit idea.
>>109230258I look like that
>>109230422>>109230398good lads itt
>>109228572i didn't publish the code, but it's a raycasting engine. I wanted to use 16.16 fixed points, but didn't have the time, so I settled on floats. that said, my program is very slow compared to other raycasting engines, because I wanted nicer visuals.>>109230568it's xlib, so basically 100% cpu, I create an image in memory, then put pixels in it. finally I call a function that uploads it to gpu and displays the image. It's extremely inefficient>>109230595that's wconversion for you, but it's a very good flag. it helped me a lot because I have functions that take floats, but work with integer under the hood for performance.
