i also recommend you pack images and audio in rpa files for distribution. could probably save some space with that.
As has been pointed out by
Smrdgy already, packing the game assets into an rpa file won't do much for compression; it mostly leads to said assets being harder to access. Which may or may not be seen as a boon.
Another valid point they brought up is the importance of rpyc files. These should only ever be recreated in very rare circumstances, by the developer, after having populated the
You must be registered to see the links
. If you don't feel like reading the
Under the hood article they linked: Ren'Py compiles a script by associating each line with a name – in case of labels it's said label's name, for regular lines it's basically a number. These names are in turn referenced by the save files as well as by persistent storage. Apart from the values of the defined variables, the former tracks each visited line of your playthrough allowing you to continue or rollback from the saved point, while the latter tracks each line ever visited allowing you to
CTRL-skip through already played parts. Having the rpyc files be recreated on the user's end will lead to new names on all but labels, thus breaking most saves and the ability to skip over already seen lines (given default settings for the latter).
However to make a drastic change in size, the dev should consider converting the images into webp instead of jpeg.
No.
WEBP's strengths lie in supporting transparency and still looking fine on low quality settings. And while It does well on less detailed, uniformly coloured areas, photorealistic renders aren't it's forte. In fact, on most quality setting above it's default (see also
You must be registered to see the links
), modern JPEG encoders like mozjepg an it's de-facto successor jpegli have it beat; meaning lower file sizes at the same DSSIM values or lower DSSIM values at the same file sizes. So, unless everybody here's fine with Sylvie loosing a part of her freckles – which I am not – WEBP should only be used for the actual compression – and
even then…
