4 ms·

The hardest part of building software is not coding, it's requirements

sylware 3y ago
I disagree, the hard part is once it is "finished" is to "keep it that way". In most life cycle of many software components, coding is actually very small, to a point, spending more time coding in assembly several hardware ISA code paths is on the long run an advantage (independence from ultra-complex compilers is a massive "pro"). What's important in such life cycle of such software components is ownership and operation. Most if not all system components are like that.
- chiefalchemist 3y ago
  That's important. But if the application is not meeting the business' needs (as a finished product) what good is "keeping it that way"? In fact, as the business evolves, so must the application. That is "keeping it that way" means keeping your fingers on the pulse of requirements.
  - sylware 3y ago
    As I wrote in my post, not all software components are like I did describe and you are refering to a type of software which is not in that scope. The one you are refering to is a type of software where the "problem" is ever changing requirements. Then the "problem" to be solved is with the people making those ever changing requirements, good or bad reasons (usually, people are bad: planned obsolescence to justify income, etc).
    - chiefalchemist 3y ago
      I mean this because I'm trying to help, but you're difficult to understand. Ambiguous even. There's a comms heuristic: It's not what you say, it's what they here. You likely understand you. I'm not so sure we're understanding you in the same way. Again. This isn't a swipe; only some friendly food for thought.
      - sylware 3y ago
        ??? chatgpt?
        chiefalchemist 3y ago
        You? ChatGPT? Could be. It certainly comes across as if you could be. e.g. " Then the "problem" to be solved is with the people making those ever changing requirements, good or bad reasons " Good or bad reasons? Reasons for that? What does it have to do with changing requirements? We can't get in your head. You're going to have to speak in ways that presume the receive has *no idea* what you're talking about. Instead, there seems to be some leave of assumption, and that leave a gap.
        sylware 3y ago
        yep, chatgpt
        chiefalchemist 3y ago
        Be more vague, please. It makes having an intelligent conversation so much easier. FFS smh
chiefalchemist 3y ago
My heuristic is: Building is relatively easy. Deciding what to build? That's the 10x harder part. The experienced realize this. The less experienced, much less so.

I disagree, the hard part is once it is "finished" is to "keep it that way". In most life cycle of many software components, coding is actually very small, to a point, spending more time coding in assembly several hardware ISA code paths is on the long run an advantage (independence from ultra-complex compilers is a massive "pro"). What's important in such life cycle of such software components is ownership and operation. Most if not all system components are like that.

chiefalchemist 3y ago
That's important. But if the application is not meeting the business' needs (as a finished product) what good is "keeping it that way"? In fact, as the business evolves, so must the application. That is "keeping it that way" means keeping your fingers on the pulse of requirements.
- sylware 3y ago
  As I wrote in my post, not all software components are like I did describe and you are refering to a type of software which is not in that scope. The one you are refering to is a type of software where the "problem" is ever changing requirements. Then the "problem" to be solved is with the people making those ever changing requirements, good or bad reasons (usually, people are bad: planned obsolescence to justify income, etc).
  - chiefalchemist 3y ago
    I mean this because I'm trying to help, but you're difficult to understand. Ambiguous even. There's a comms heuristic: It's not what you say, it's what they here. You likely understand you. I'm not so sure we're understanding you in the same way. Again. This isn't a swipe; only some friendly food for thought.
    - sylware 3y ago
      ??? chatgpt?
      - chiefalchemist 3y ago
        You? ChatGPT? Could be. It certainly comes across as if you could be. e.g. " Then the "problem" to be solved is with the people making those ever changing requirements, good or bad reasons " Good or bad reasons? Reasons for that? What does it have to do with changing requirements? We can't get in your head. You're going to have to speak in ways that presume the receive has *no idea* what you're talking about. Instead, there seems to be some leave of assumption, and that leave a gap.
        sylware 3y ago
        yep, chatgpt
        chiefalchemist 3y ago
        Be more vague, please. It makes having an intelligent conversation so much easier. FFS smh

chiefalchemist 3y ago

My heuristic is: Building is relatively easy. Deciding what to build? That's the 10x harder part. The experienced realize this. The less experienced, much less so.