thank you for the great editor! I acknowledge improvements might not be at scope at all, but I feel like sharing it here.
Jump mode and 2 letters jump mode in particular, is a great feature, working its best for structured, formatted code; in this case, there are reasonable amounts of tokens generated, thus user doesn't get overwhelmed. This results into really fast navigation!
Problem
On the other hand, single letter token mode doesn't seem like adding too much value on its own; that is: it's, in the best case, as good as two letters token mode.
And because the idea itself, of using simpler input for the navigation (which single keypress is), generates damn huge expectations (in myself, at least), along with the implementation, which fails to blow away a user's mind, it's rather experienced as slight disappointment, from the UX perspective.
So:
What if we just treated letters of words actual tokens?
Example:
We have the source line:
Beyond that, you're going to want to use Amp's jump mode to navigate. This mode
Upon single quote keypress, it's currently transformed to:
Isn't it more natural to have it like this:
Seen implications
Q: "But wait!... tokens won't be unique anymore!"
A: Is it an issue at all? By user's repeatedly pressing the same key, iterator would loop over duplicate tokens. Seems cool, isn't it?
Q: "Token distribution is text dependent!"
A: Me personally feeling just fine with lack of pre-defined ordering like in the second screenshot. Still, this is good point, as it's not that alphabetic ordering wouldn't have had any sense, so people might be benefitting from it.
The downside of it is that letters are not sorted alphabetically on keyboards since, more or less, since inception of typing devices AFAIK.
But wait... Who looks at keyboard while typing those days, really? :) Whatever ordering it is, not elaborating into its historical origins, it's captured in our heads, right?
And here we come to one more variant of my proposal:
What if we just... QWERTY?
and when you need it, you'll use n as reserved and forbidden for token use, to iterate thru the tokens!
Feature Request: improve single-letter token jump mode
Hi,
thank you for the great editor! I acknowledge improvements might not be at scope at all, but I feel like sharing it here.
Jump mode and 2 letters jump mode in particular, is a great feature, working its best for structured, formatted code; in this case, there are reasonable amounts of tokens generated, thus user doesn't get overwhelmed. This results into really fast navigation!
Problem
On the other hand, single letter token mode doesn't seem like adding too much value on its own; that is: it's, in the best case, as good as two letters token mode.
And because the idea itself, of using simpler input for the navigation (which single keypress is), generates damn huge expectations (in myself, at least), along with the implementation, which fails to blow away a user's mind, it's rather experienced as slight disappointment, from the UX perspective.
So:
What if we just treated letters of words actual tokens?
Example: We have the source line:
Upon single quote keypress, it's currently transformed to:
Isn't it more natural to have it like this:
Seen implications
Q: "But wait!... tokens won't be unique anymore!"
A: Is it an issue at all? By user's repeatedly pressing the same key, iterator would loop over duplicate tokens. Seems cool, isn't it?
Q: "Token distribution is text dependent!"
A: Me personally feeling just fine with lack of pre-defined ordering like in the second screenshot. Still, this is good point, as it's not that alphabetic ordering wouldn't have had any sense, so people might be benefitting from it.
The downside of it is that letters are not sorted alphabetically on keyboards since, more or less, since inception of typing devices AFAIK.
But wait... Who looks at keyboard while typing those days, really? :) Whatever ordering it is, not elaborating into its historical origins, it's captured in our heads, right?
And here we come to one more variant of my proposal:
What if we just... QWERTY?
and when you need it, you'll use
n
as reserved and forbidden for token use, to iterate thru the tokens!At least to me seems like nice one :)
Thank you for going thru it!