And progress has been quick from then on: Most of duit was written in the two weeks around last year’s Christmas. With applications and subsequent library improvements written in the evenings of January.
This is much faster than anything I can do. I’d be very interested to read more about this process. And also how much technical debt this incurs. (Although, from the description of its extensibility, it may not be an unusual amount.)
When I say UI, I don’t mean “user interface”. Instead I mean “a type implementing the duit.UI interface”. All of duits user interface “widgets” – a word I don’t like and avoid – implement the UI interface.
Why? This makes the rest of the documentation more confusing to read. Why replace a word that (seems to) match the author’s meaning by another one that already has a very different meaning?
It seems like duit.UI is the thing to be renamed to duit.Widget (or at least something that doesn’t have a very different meaning) instead of changing the meaning of words to match the library.
(I have to admit that I’m also guilty of naming classes UI when I mean widget in my own libraries.)
“Focus” and “hover” are the same concept in duit. That’s one concept less to worry about! No need to click to type – focus follows the mouse – as illustrated by this looping video:
The idea is to warp the mouse pointer to where you will focus your eyes.
I’m happy to see they are experimenting and I think hover to focus is the default in X. However, as a potential user I really this doesn’t become an irreversible baked in assumption. My mouse cursor already has a very different meaning than “where my eyes will be” and don’t want its position touched.
This seems like a decision usually left to the windows manager. (Sending position hints back to the WM and letting it decide to use them or not, on the other hand, could be a good idea.)
This looks very promising. As soon as I saw it I thought “plan9”, and sure enough, it’s somewhat related, using devdraw from plan9port.
The widget set is complete enough to create functioning apps. The developer has created several demo apps, including a local mail reader, acme clone, and a simple database browser and query tool.
It sounds silly to seriously try writing a new UI library. It would take an enormous amount of time which I didn’t have with a full-time job.
That’s been the problem I’ve seen with half of the UI libraries I’ve looked at. It is silly. “Halfway” is usually not good enough. If a key part you need for your interface (let’s say “dropdown boxes”) is not implemented or buggy it’s sometimes (for small projects) easier to just switch libraries instead of now becoming part of the upstream of your UI framework :(
That said, it looks good and I hope it gets somewhere :)
This looks really good from the video!
This is much faster than anything I can do. I’d be very interested to read more about this process. And also how much technical debt this incurs. (Although, from the description of its extensibility, it may not be an unusual amount.)
Why? This makes the rest of the documentation more confusing to read. Why replace a word that (seems to) match the author’s meaning by another one that already has a very different meaning?
It seems like
duit.UIis the thing to be renamed toduit.Widget(or at least something that doesn’t have a very different meaning) instead of changing the meaning of words to match the library.(I have to admit that I’m also guilty of naming classes UI when I mean widget in my own libraries.)
I’m happy to see they are experimenting and I think hover to focus is the default in X. However, as a potential user I really this doesn’t become an irreversible baked in assumption. My mouse cursor already has a very different meaning than “where my eyes will be” and don’t want its position touched.
This seems like a decision usually left to the windows manager. (Sending position hints back to the WM and letting it decide to use them or not, on the other hand, could be a good idea.)
This looks very promising. As soon as I saw it I thought “plan9”, and sure enough, it’s somewhat related, using devdraw from plan9port.
The widget set is complete enough to create functioning apps. The developer has created several demo apps, including a local mail reader, acme clone, and a simple database browser and query tool.
And it’s pure Go.
Very promising, indeed.
That’s been the problem I’ve seen with half of the UI libraries I’ve looked at. It is silly. “Halfway” is usually not good enough. If a key part you need for your interface (let’s say “dropdown boxes”) is not implemented or buggy it’s sometimes (for small projects) easier to just switch libraries instead of now becoming part of the upstream of your UI framework :(
That said, it looks good and I hope it gets somewhere :)
I saw some Dutch text in the email program demo video. “Duit” in Dutch means: coin.
So what you’re saying is: this is an initial coin offering? :-P