Country Flags

I just had a great idea: Use a spritemap for country flags, using the ISO2 codes as coordinates.

Like AL is x=0 y=12 (approx) or DE is 4|5

I just told our codex gpt about this and she completely implemented it on her own for an app within three minutes.

This saves like 242 requests for the first visit to a profile page.

A simple idea, and if i had to implement this, i would have taken like an hour.
I estimate my token cost to $0.01 for this little gem

Golden Times!
- giz

Another nice idea i had a few weeks ago:

Display the Timezone Info as current time instead of offset.

Way cooler

I am looking for a volunteer that manages the spreadshirt shop.
Win wil be yours.
Make a few more colors and re-ship maybe.
- giz

Quote from gizmore

Aug 13, 2026 - 07:51:52

This saves like 242 requests for the first visit to a profile page.

It does not, as a profile page only loads a single country flag. Worse, it loads a huge flag sprite sheet when only a few bytes flag is required ;P

Only benefit would be on pages rendering a load of different flags at once, but even then, if caching headers are set properly, single files are probably better. Modern browsers load that via one connection in HTTP/2 anyway (or keep-alive on 1.1). And once it's loaded it can be stuck into cache, basically forever. You cannot properly cache a sprite sheet, clientside, as any change to any country flag (or, a new one, or removing one) would have to invalidate it.


(This reply was AI and brainrot free)

On my phpgdo7 framework I did not want completion and was too lazy for lazy loading. the spritemap is tiny... way better.

I am not talking about gwf3

Quote from livinskull

Aug 15, 2026 - 09:22:40

You cannot properly cache a sprite sheet, clientside, as any change to any country flag (or, a new one, or removing one) would have to invalidate it.

But country flags don't need to be changed/added/removed often nowadays as far as I know.
So it would probably still be worth it because cache invalidations would be few and far between.

A further refinement would be to split country flags into N groups in alphabetical order (in English) and create a sprite for each group.
Instead of one big sprite, we now have N smaller sprites.
Then, when a country flag changes/is removed/is added, most of the times only the cache for 1 sprite corresponding to that country's group would get invalidated, or at most 2 in case that country changes its name (the sprite corresponding to the group it leaves and the one corresponding to the group it joins).
Country groups must be allowed to grow/shrink freely for this to be possible.
This means group size differences would accumulate over time, meaning some groups could get big enough that their probability of invalidation could become significantly higher than the others due to increased membership alone.
Also, there is of course the matter of tuning the N parameter.
However, again, countries don't change all that often, I don't think it's a problem if it becomes necessary to regenerate all the groups in 100+ years .

By taking into account more data about actual countries (stability, geopolitical connection, ...), sprite groups can be created in smarter ways so as to optimize for different aspects of cache performance.
OK, enough rambling now

Have to admit, I am more on livinskull's side here (, yeah, I know!!):

Don't feel that is a problem right now, and the impact will be rather small (now, on new WC hardware/software).

EDIT: wrong Thread

Sprites are from another era, I think there should be better solutions now, like full CSS flags =)

Because HTML does not allow markup in <option> the UTF8 Flags are a nice solution as well... but that does not solve the markup problem for other <select> 'D'oh