Rational Rose etc. could generate code from UML diagrams, then you “only” needed architects.
In reality it only gave a little help during the design phase, as soon as someone touches the generated code, you have to manually merge changes to UML.
Rational Rose etc. could generate code from UML diagrams, then you “only” needed architects.
In reality it only gave a little help during the design phase, as soon as someone touches the generated code, you have to manually merge changes to UML.
And 42 seconds in jython.
That’s easy you just use the huge number of test cases to ensure against introducing new bugs.
/S
I was looking into learning COBOL some years ago, because i found that verbosity interesting.
And it seemed like there’s not many libs and toolboxes out there, compared to the major languages that has libs for everything, so I couldn’t really use it for small projects.
Editing grub.cfg from an emergency console, or running grub-update from a chroot is a close second.
Adding the right Modeline to xorg.conf seemed more like magic when it worked. 🧙🏼
Correct, and your point is?
Could be I was not clear when I wrote performance, I am talking about High Performance Computing, where you want to spend all CPU cycles on solving your problem. While taking Amdal’s Law into account. https://en.wikipedia.org/wiki/Amdahl’s_law
🤦🏽♀️ Thanks for explaining, my brain must have corrected the race condition.
Regarding threads: I have had good experience with using thread safe queues everywhere to exchange data between threads, it’s the right tool in many cases, but I doubt queues to be useful when coding for performance.
Until they find out that the way to descriptive variables or functions needs to be extended with new business logic requiring renaming of functions again and again.
I think maintaining code with this level of verbose naming, will be a pain over time. If they don’t let the naming slip, and then they could as well use cryptic 3 letter names.
Ah I see now that I mixed it up with classless functions in a header.
Yep, first thing that comes to mind is that header only classes need to use inline functions.
Then you open the core file with GDB and hope the stack is not smashed.
That’s Weyland-Yutani
Where did you find that? They may be right if it’s a closed fork, but if it’s from something on GitHub i can’t see how it violates GPL?
Edit: just scrolled a bit down and saw another comment, with the full image, nvm.
Reminds me of my lottery number randomizer in VB 3.0 I calculated 10 rows and checked each input box against the others without any loops, and then each row was just copy pasted. 🤦
I meant to write 3.5" floppy drives, and yes the 3.5" and 2.5" form factors are still going strong, even if the NVMe’s probably will reduce the use of 2.5"
They sadly don’t have 3.5" [floppy] drives anymore, and both the ISA and PCI busses are nowhere to be found 😔
I used pcpartpicker for my latest build it’s a good help when assembling and can help avoid those incompatible parts.
My first thought was, can you flush the buffer?
Perfect
I prefer the multi thread problems that can be solved using queues.