4 ms·

The Zen of Missing Out on the Next Great Programming Tool (2016)

8 points by bhalp1 9y ago

lafar6502 9y ago
Modern ui is a mess and frameworks + stackoverflow make it messier. Old desktop APIs , even Win32, were at least built around a model that made some sense and allowed you to use correct abstractions, and there was a logical path to a working solution. Today you try to glue together some random SO answers and some ugly jquery plugins, with no idea why it has to be done this way or how to approach a task in a correct way. Look at basic things like drag&drop, keyboard shortcuts, menus, tab order, field focus management and all stuff that’s important in application ui - why it’s always an ugly and painful hacking with messy and half-baked js , we already had much better tools and replaced them with something that only pretends to be general purpose ui runtime
- megaman22 9y ago
  It's a breath of fresh air when I get to get away from our "modern", overwrought JS web stack and go bang out some internal tools with good old WinForms. So much less friction and magic all around.
  - lafar6502 9y ago
    Maybe this is the only reason why new js ui frameworks are created- to run away from the mess and unintentionally find yourself in even bigger pile of junk
luckydude 9y ago
Couldn't agree more as a manager for ~25 years. Engineers, and I am one, a decent one, love the new shiny stuff. Me too. As a manager who has to ship product that works? Not so much. That made me switch gears. My team used to joke that if it happened in Unix after 1980 I don't like it. That's not true, we used mmap, and a bunch of other stuff that happened later but the docs I used to program were old. I let new stuff in grudgingly. I think it worked. We're dead so there is that (BitKeeper) but we had an almost 20 year profitable run with no VC in the valley so there is that too. Our support was over the top, I would challenge you to find a company that had as many users as we did and had the support that we did. Our 24/7 average response time was something like 22 minutes and the only reason it was that long was we went to bed. When we were awake the response time was measured in seconds or single digit minutes. I credit the response time to being boring. No shiny stuff, just stuff that works, be boring. The mantra was "drive towards zero support costs, that's how we scale". That didn't mean not support people, it meant make the product so good that there were no support issues. We didn't get there but we got closer than a lot of companies. tl;dr: don't beat your management / leadership up so much for not wanting the shiny new stuff. You are in this to make money or change the world or something. All of those somethings are not well served by unproven new stuff. If you have to use new stuff choose very very carefully and think about the costs.

8 points by bhalp1 9y ago

Modern ui is a mess and frameworks + stackoverflow make it messier. Old desktop APIs , even Win32, were at least built around a model that made some sense and allowed you to use correct abstractions, and there was a logical path to a working solution. Today you try to glue together some random SO answers and some ugly jquery plugins, with no idea why it has to be done this way or how to approach a task in a correct way. Look at basic things like drag&drop, keyboard shortcuts, menus, tab order, field focus management and all stuff that’s important in application ui - why it’s always an ugly and painful hacking with messy and half-baked js , we already had much better tools and replaced them with something that only pretends to be general purpose ui runtime

megaman22 9y ago
It's a breath of fresh air when I get to get away from our "modern", overwrought JS web stack and go bang out some internal tools with good old WinForms. So much less friction and magic all around.
- lafar6502 9y ago
  Maybe this is the only reason why new js ui frameworks are created- to run away from the mess and unintentionally find yourself in even bigger pile of junk

luckydude 9y ago

Couldn't agree more as a manager for ~25 years. Engineers, and I am one, a decent one, love the new shiny stuff. Me too. As a manager who has to ship product that works? Not so much. That made me switch gears. My team used to joke that if it happened in Unix after 1980 I don't like it. That's not true, we used mmap, and a bunch of other stuff that happened later but the docs I used to program were old. I let new stuff in grudgingly. I think it worked. We're dead so there is that (BitKeeper) but we had an almost 20 year profitable run with no VC in the valley so there is that too. Our support was over the top, I would challenge you to find a company that had as many users as we did and had the support that we did. Our 24/7 average response time was something like 22 minutes and the only reason it was that long was we went to bed. When we were awake the response time was measured in seconds or single digit minutes. I credit the response time to being boring. No shiny stuff, just stuff that works, be boring. The mantra was "drive towards zero support costs, that's how we scale". That didn't mean not support people, it meant make the product so good that there were no support issues. We didn't get there but we got closer than a lot of companies. tl;dr: don't beat your management / leadership up so much for not wanting the shiny new stuff. You are in this to make money or change the world or something. All of those somethings are not well served by unproven new stuff. If you have to use new stuff choose very very carefully and think about the costs.