4 ms·

The idea in ZAMM that has stuck with me the longest, and affects my daily thinking most, is gumption traps. Those little things, little frustrations, that make

by mjb 4y ago

The idea in ZAMM that has stuck with me the longest, and affects my daily thinking most, is gumption traps. Those little things, little frustrations, that make you lose enthusiasm for a project or task. Hard things can be enjoyable. Even chores like cleaning, and maintenance, or documentation. But we all have those little things that eat away at our gumption. I've learned that an important way to improve my enjoyment of life is to remove those gumption traps. For example, a lot of people give the advice to buy a cheap tool, and replace it with an expensive one if it breaks. That's OK advice, as far as it goes. But I find that working with bad tools, and breaking a tool, is a huge gumption trap for me. So I buy fewer quality tools instead. If that means I have to do some tasks the hard way, that's OK, because I still find that enjoyable.

hinkley 4y ago
> But I find that working with bad tools, and breaking a tool, is a huge gumption trap for me. This is my current thesis for why software continues to improve at such a slow rate from the customer's perspective. We as developers spend the least amount of time and energy on our tools. Having marinated in a large pile of mediocre tools, we then set about designing our software to be 'good'. We're standing in a cognitive hole with respect to what 'good' even is and then we get nominally the same results over and over. Jetbrains frustrates me constantly, in particular because it frustrates me in the same ways it frustrated me fifteen years ago when I first started using it. It's held up as some pinnacle of developer tooling, instead of the table stakes/median yardstick we should be applying to their tools at this point, 15-20 years after they were designed. Until we expect better of each other the customer is always going to be disappointed with what they get. It takes a massive amount of energy to produce gemstones from garbage.
- ljw1001 4y ago
  In my opinion, Jetbrains IDEA was a better tool ten years ago, before they decided to make versions for every language imaginable. In some ways they’ve improved and occasionally still delight, but in others their Java support is remarkably weak or non existant. Support for working with JNI is one that comes immediately to mind
  - hinkley 4y ago
    The leaky abstraction definitely hurts. Having not worked there it's hard to tell if the team shifted to maintenance mode because the dreamers that started the whole thing left, or because the code is so work-hardened that only maintenance can be done on it. But it definitely feels like it's a product in maintenance mode. I recall too, just before TeamCity launched, that there was a notable uptick in bugs and regressions. As if someone had been working on a new internal product that poached all of the R&D-compatible people and left things in the hand of more junior members.

by mjb 4y ago

hinkley 4y ago

> But I find that working with bad tools, and breaking a tool, is a huge gumption trap for me. This is my current thesis for why software continues to improve at such a slow rate from the customer's perspective. We as developers spend the least amount of time and energy on our tools. Having marinated in a large pile of mediocre tools, we then set about designing our software to be 'good'. We're standing in a cognitive hole with respect to what 'good' even is and then we get nominally the same results over and over. Jetbrains frustrates me constantly, in particular because it frustrates me in the same ways it frustrated me fifteen years ago when I first started using it. It's held up as some pinnacle of developer tooling, instead of the table stakes/median yardstick we should be applying to their tools at this point, 15-20 years after they were designed. Until we expect better of each other the customer is always going to be disappointed with what they get. It takes a massive amount of energy to produce gemstones from garbage.

ljw1001 4y ago
In my opinion, Jetbrains IDEA was a better tool ten years ago, before they decided to make versions for every language imaginable. In some ways they’ve improved and occasionally still delight, but in others their Java support is remarkably weak or non existant. Support for working with JNI is one that comes immediately to mind
- hinkley 4y ago
  The leaky abstraction definitely hurts. Having not worked there it's hard to tell if the team shifted to maintenance mode because the dreamers that started the whole thing left, or because the code is so work-hardened that only maintenance can be done on it. But it definitely feels like it's a product in maintenance mode. I recall too, just before TeamCity launched, that there was a notable uptick in bugs and regressions. As if someone had been working on a new internal product that poached all of the R&D-compatible people and left things in the hand of more junior members.