3 ms·

I have to pretty much reject this post wholesale. True the post it's disagreeing with picked weak examples. But other than that pretty much all of its overall p

I have to pretty much reject this post wholesale. True the post it's disagreeing with picked weak examples. But other than that pretty much all of its overall points were actually very accurate. The original article was not about json/scala/etc. It was about this: To summarize, we have to deal with a lot of accidental complexity these days, for multiple reasons. First of all, we adopt techniques and abstractions without analying and evaluating their costs and their usefulness in our specific situation. Moreover, we love solving technological problems.

lmm 8y ago
> True the post it's disagreeing with picked weak examples. But other than that pretty much all of its overall points were actually very accurate. That's what I initially thought. But how can the overall points possibly be right if they lead to exactly the wrong conclusions in most or all of the specific examples? At best the part you quote is vacuous and unhelpful for actually making technical decisions.
- alexandercrohde 8y ago
  Just logically, a point can never be disproven with bad examples. Bad examples may be unpersuasive, may bore the reader, but they don't make a true point false. As best I can tell when you call his point "vacuous" you mean you think it's too abstract to be practical. I disagree. What I take a away from his article is: - Before you build something, ask yourself "Is this REALLY a technical problem? Or is it better solved non-technically?" - What are the costs, especially complexity, of this change? Weigh those against the benefit. Seems concrete and useful to me.
  - lmm 8y ago
    > Seems concrete and useful to me. It sounds useful. But Westheide followed this methodology, presumably put time and effort into it, and yet made completely wrong technical decisions. So I don't believe the method can possibly be useful; he'd've got better results by saving his time and flipping a coin. (Put another way, this methodology doesn't outperform a placebo)
    - alexandercrohde 8y ago
      You know that's a bad argument. You can't definitively say "one person tried to follow a piece of advice and it didn't help one person therefore the advice is false/wrong/worthless." I'm sure if you thought about this though that would be pretty clear to you.
      - lmm 8y ago
        > You know that's a bad argument. You can't definitively say "one person tried to follow a piece of advice and it didn't help one person therefore the advice is false/wrong/worthless." But this isn't just one person trying - this is the example given by the person giving the advice, so it's presumably the strongest case they can make in favour of their advice.

lmm 8y ago

> True the post it's disagreeing with picked weak examples. But other than that pretty much all of its overall points were actually very accurate. That's what I initially thought. But how can the overall points possibly be right if they lead to exactly the wrong conclusions in most or all of the specific examples? At best the part you quote is vacuous and unhelpful for actually making technical decisions.

alexandercrohde 8y ago
Just logically, a point can never be disproven with bad examples. Bad examples may be unpersuasive, may bore the reader, but they don't make a true point false. As best I can tell when you call his point "vacuous" you mean you think it's too abstract to be practical. I disagree. What I take a away from his article is: - Before you build something, ask yourself "Is this REALLY a technical problem? Or is it better solved non-technically?" - What are the costs, especially complexity, of this change? Weigh those against the benefit. Seems concrete and useful to me.
- lmm 8y ago
  > Seems concrete and useful to me. It sounds useful. But Westheide followed this methodology, presumably put time and effort into it, and yet made completely wrong technical decisions. So I don't believe the method can possibly be useful; he'd've got better results by saving his time and flipping a coin. (Put another way, this methodology doesn't outperform a placebo)
  - alexandercrohde 8y ago
    You know that's a bad argument. You can't definitively say "one person tried to follow a piece of advice and it didn't help one person therefore the advice is false/wrong/worthless." I'm sure if you thought about this though that would be pretty clear to you.
    - lmm 8y ago
      > You know that's a bad argument. You can't definitively say "one person tried to follow a piece of advice and it didn't help one person therefore the advice is false/wrong/worthless." But this isn't just one person trying - this is the example given by the person giving the advice, so it's presumably the strongest case they can make in favour of their advice.