Most frequently viewed/copied characters

[eugenesvk]

Another category in addition to manually added Favorites could be automatically generated "frequent chars" requiring 0 clicks

[Mercury13]

May I refuse?

It’ll require lots of persistent data and slow down the startup. Maybe it’s possible to speed up then, but will require very tedious programming.

Also, >20 random characters are just a mess.

RECENTLY copied — it’s possible, less data and not so random chars.

[eugenesvk]

But it won't be random characters, it will be the 20 characters user has clicked on to view the most number of times (0 clicks means 0 additional clicks to save to favorites, otherwise we do need user behavior to decide which symbols to save)

And you could have user configuration of the max number of chars to remember frequency of interaction for, so it shouldn't be much data at all

And if using frecency (frequency + recency) the long tail of less used chars will be removed from the dataset eventually to free up space for the new candidates...

But recently copied should also be fine, I think

[Mercury13]

One of reasons why I gave up later Qt is slow startup!

[eugenesvk]

I agree with you that startup speed is important, thanks for keeping it in mind!

I just don't understand why adding another list of chars would affect it. The frecency data could even be loaded after startup?