README.adoc: quantify how bad agg is
This commit is contained in:
+12
-3
@@ -15,14 +15,23 @@ Background
|
||||
----------
|
||||
First, I thought something as simple would already exist.
|
||||
But there's basically only _agg_, which doesn't care about keypresses,
|
||||
and it uses an inefficient file format.
|
||||
and GIF is an inefficient file format:
|
||||
|
||||
[options="autowidth,header",cols="<,<,<,>,>"]
|
||||
|===
|
||||
| Format | Encoding | Tool | Filesize | Relative size
|
||||
| WebP | Lossless | libwebp preset 9 | 401884 | 100 %
|
||||
| GIF | Palette | magick | 825280 | 205 %
|
||||
| APNG | Lossless | apngasm + oxipng -z | 939822 | 234 %
|
||||
| GIF | Palette | gifski (used by agg) | 1046027 | 260 %
|
||||
|===
|
||||
|
||||
Second, I thought this would be reasonably doable in something other than C++.
|
||||
Meanwhile, all the good important libraries have a C API.
|
||||
|
||||
Text rendering could still see some improvements, notably with Unicode
|
||||
box-drawing characters, and keypress parsing could learn many more sequences,
|
||||
but I've already spent enough time on this project for now.
|
||||
box-drawing characters, and keypress parsing could learn many more sequences.
|
||||
I've already spent enough time on this project for now.
|
||||
|
||||
Packages
|
||||
--------
|
||||
|
||||
Reference in New Issue
Block a user