5 ms·

Show HN: A touch interface for tree editing

17 points by hcs 9y ago

Micoloth 9y ago
Wow, this is great good job! There is a small bug, where you can get a 2-wide block with only one 1-wide child, and when you pull it down a white square remains where it shouldn't. Anyway I really like this, it is so comfortable to use!
- Micoloth 9y ago
  It happens when you take a white square and pull it diagonally to the bottom right or bottom left
  - hcs 9y ago
    Good catch, it shouldn't be possible to drag the white sqaures ("handles") any direction but up.
fiatjaf 9y ago
Great interface.
vanderZwan 9y ago
Very cool, I like the simplicity of it. The Fibonacci example makes me wonder if it would make more sense to make the interface expand down instead of up, since that matches the reading direction of Lisp syntax more. Also, could we "remix" this to work for concatenative languages? That is, postfix notation. So something like Forth, Joy or Kitten?
- hcs 9y ago
  Thanks! A big part of the idea is representing data flow, I don't think that makes as much sense with a concatenative language where things are irregularly moving on and off the stack. You'd have to show the relation of words to the stack, somehow, or rework it to be centered around control flow. I have only a passing knowledge of Forth, though, more modern concatenative languages might have a different texture. I did, however, take the RPN approach, which is why operations appear after/below their operands, which has always seemed more natural to me.
  - vanderZwan 9y ago
    > I did, however, take the RPN approach Hah! I feel dumb for not noticing this now :) Regarding concatenative languages, for most words you how many elements of the stack it consumes and then drops back on the stack, so those should be easy. You're right that when that varies things get complicated. Not sure how to address that - perhaps manual annotations? But would it still be worth it then? I suspect something like Joy[0] or Cat[1] would be easier to handle though, being a more predictable, functional language. Kitten[2], while a lot cooler and the only one still being actively developed, requires type annotations, so would probably a pain to write out. BTW, sticking to the topic of programming language input (even though this interface is more general than that) if you add some kind of support for often and recently used keywords (kind of like the predictions in mobile soft-keyboards) that could save a lot of typing too! Maybe have a top three of both, one on the left, the other on the right. [0] https://en.wikipedia.org/wiki/Joy_(programming_language) https://en.wikipedia.org/wiki/Joy_(programming_language) [1] https://web.archive.org/web/20120723021444/http://www.cat-language.com https://web.archive.org/web/20120723021444/http://www.cat-la... [2] http://kittenlang.org/ http://kittenlang.org/
beastcoast 9y ago
The demo doesn't work on iOS. Otherwise looks cool!
- hcs 9y ago
  Sorry about that, it works OK on my iPad and in the iOS simulator, if you want to help me debug your system drop me a line!

17 points by hcs 9y ago

Wow, this is great good job! There is a small bug, where you can get a 2-wide block with only one 1-wide child, and when you pull it down a white square remains where it shouldn't. Anyway I really like this, it is so comfortable to use!

Micoloth 9y ago
It happens when you take a white square and pull it diagonally to the bottom right or bottom left
- hcs 9y ago
  Good catch, it shouldn't be possible to drag the white sqaures ("handles") any direction but up.

fiatjaf 9y ago

Great interface.

vanderZwan 9y ago

Very cool, I like the simplicity of it. The Fibonacci example makes me wonder if it would make more sense to make the interface expand down instead of up, since that matches the reading direction of Lisp syntax more. Also, could we "remix" this to work for concatenative languages? That is, postfix notation. So something like Forth, Joy or Kitten?

hcs 9y ago
Thanks! A big part of the idea is representing data flow, I don't think that makes as much sense with a concatenative language where things are irregularly moving on and off the stack. You'd have to show the relation of words to the stack, somehow, or rework it to be centered around control flow. I have only a passing knowledge of Forth, though, more modern concatenative languages might have a different texture. I did, however, take the RPN approach, which is why operations appear after/below their operands, which has always seemed more natural to me.
- vanderZwan 9y ago
  > I did, however, take the RPN approach Hah! I feel dumb for not noticing this now :) Regarding concatenative languages, for most words you how many elements of the stack it consumes and then drops back on the stack, so those should be easy. You're right that when that varies things get complicated. Not sure how to address that - perhaps manual annotations? But would it still be worth it then? I suspect something like Joy[0] or Cat[1] would be easier to handle though, being a more predictable, functional language. Kitten[2], while a lot cooler and the only one still being actively developed, requires type annotations, so would probably a pain to write out. BTW, sticking to the topic of programming language input (even though this interface is more general than that) if you add some kind of support for often and recently used keywords (kind of like the predictions in mobile soft-keyboards) that could save a lot of typing too! Maybe have a top three of both, one on the left, the other on the right. [0] https://en.wikipedia.org/wiki/Joy_(programming_language) https://en.wikipedia.org/wiki/Joy_(programming_language) [1] https://web.archive.org/web/20120723021444/http://www.cat-language.com https://web.archive.org/web/20120723021444/http://www.cat-la... [2] http://kittenlang.org/ http://kittenlang.org/

beastcoast 9y ago

The demo doesn't work on iOS. Otherwise looks cool!

hcs 9y ago
Sorry about that, it works OK on my iPad and in the iOS simulator, if you want to help me debug your system drop me a line!