3 ms·

The problem with configuration options is that they need maintaining indefinitely to accommodate a small proportion of the user population. That means testing b

The problem with configuration options is that they need maintaining indefinitely to accommodate a small proportion of the user population. That means testing behaviour with the flag on/off, and taking into account the possibility of that flag being on when dealing with bug reports. In this case it's a minor thing which probably wouldn't require much maintainance, but you have to draw the line somewhere.

blktiger 10y ago
Agreed. Especially in a case like this where it's actually pretty trivial to write an extension for this behavior. This change decreases the maintenance of chromium, removes a non-obvious data-destroying keyboard shortcut, and still gives users the option of bringing the feature back if they want it.
shkkmo 10y ago
In most cases that makes complete sense... but this behavior has been around so long. Adding a configuration option means they can track data on if this feature is being used and worth continuing to support. That said, I'm glad they finally did something to fix this bad UX design.
- notatoad 10y ago
  >Adding a configuration option means they can track data on if this feature is being used and worth continuing to support leaving it enabled also allows them to do this, which they have done for the last couple years and clearly their data says that it's not worth supporting, because now they're dropping it.
_pferreir_ 10y ago
This. Configuration flags increase complexity. Every time you change the feature in question (or add something that uses it) you will have to take into account every possible flag combination.
- TeMPOraL 10y ago
  That's not true. It's like whenever you change a line of code you need to take into account every other line of code in the project. If option for backspace behaviour suddenly starts interfering with some unrelated feature, it means somebody seriously fucked up the architecture of the project. Nothing else. Configuration flags aren't scary, and should not be used as a rationalization for dumbing down software.
- joeyespo 10y ago
  This is true in general. Although "decrease complexity" shouldn't exactly trump "build the right thing." In this case, the right thing really is contextual, so a configuration or extant setting / default is worth the cost in complexity. The quote "Everything should be made as simple as possible, but not simpler" comes to mind here.
stymaar 10y ago
> The problem with configuration options is that they need maintaining indefinitely This is a shortcut button … The maintenance argument is really a straw-man in this case … > but you have to draw the line somewhere. And this shows that they chose to place this line at «this product is our product, there is no customization options»
parenthephobia 10y ago
It's inconceivable that this should require any maintenance. I don't need a "backspace key goes back" checkbox, I just need to be able to customize the keyboard bindings. Decoupling the actions a program can perform from the key combinations which trigger those actions is basic stuff. From the maintainer's point-of-view, having the option to bind "backspace" to "window.history.back" shouldn't require any more maintenance than having the option to bind "alt-left" to "window.history.back" does. It is Google's fault that they didn't make Chrome work this way. It also means I can't unbind "ctrl-shift-q" from "quit_immediately_without_asking", because Google refuse to add an option to make it so that I can't accidentally fat-finger a C-w into something that ends my entire session. All keyboard shortcut problems disappear instantly if key bindings are configurable, however, without adding new options for each command.

blktiger 10y ago

Agreed. Especially in a case like this where it's actually pretty trivial to write an extension for this behavior. This change decreases the maintenance of chromium, removes a non-obvious data-destroying keyboard shortcut, and still gives users the option of bringing the feature back if they want it.

shkkmo 10y ago

In most cases that makes complete sense... but this behavior has been around so long. Adding a configuration option means they can track data on if this feature is being used and worth continuing to support. That said, I'm glad they finally did something to fix this bad UX design.

notatoad 10y ago
>Adding a configuration option means they can track data on if this feature is being used and worth continuing to support leaving it enabled also allows them to do this, which they have done for the last couple years and clearly their data says that it's not worth supporting, because now they're dropping it.

_pferreir_ 10y ago

This. Configuration flags increase complexity. Every time you change the feature in question (or add something that uses it) you will have to take into account every possible flag combination.

TeMPOraL 10y ago
That's not true. It's like whenever you change a line of code you need to take into account every other line of code in the project. If option for backspace behaviour suddenly starts interfering with some unrelated feature, it means somebody seriously fucked up the architecture of the project. Nothing else. Configuration flags aren't scary, and should not be used as a rationalization for dumbing down software.
joeyespo 10y ago
This is true in general. Although "decrease complexity" shouldn't exactly trump "build the right thing." In this case, the right thing really is contextual, so a configuration or extant setting / default is worth the cost in complexity. The quote "Everything should be made as simple as possible, but not simpler" comes to mind here.

stymaar 10y ago

> The problem with configuration options is that they need maintaining indefinitely This is a shortcut button … The maintenance argument is really a straw-man in this case … > but you have to draw the line somewhere. And this shows that they chose to place this line at «this product is our product, there is no customization options»

parenthephobia 10y ago

It's inconceivable that this should require any maintenance. I don't need a "backspace key goes back" checkbox, I just need to be able to customize the keyboard bindings. Decoupling the actions a program can perform from the key combinations which trigger those actions is basic stuff. From the maintainer's point-of-view, having the option to bind "backspace" to "window.history.back" shouldn't require any more maintenance than having the option to bind "alt-left" to "window.history.back" does. It is Google's fault that they didn't make Chrome work this way. It also means I can't unbind "ctrl-shift-q" from "quit_immediately_without_asking", because Google refuse to add an option to make it so that I can't accidentally fat-finger a C-w into something that ends my entire session. All keyboard shortcut problems disappear instantly if key bindings are configurable, however, without adding new options for each command.