Suggestion: Save possible seeds to speed up subsequent runs

[rda2]

When I used this tool, I put in a few data points, searched for a seed, put in a few more, searched again, etc.

It appears that each time it searched through all possible seeds, even though most of them had already been excluded from previous data.

I'd suggest limiting the search to previously found seeds if the user adds data without deleting or editing other data points.

If for instance a user puts in enough data to narrow down the search to 30 seeds, then adds another data point, it should search those 30 seeds first.

[IceLitty]

Let me say the easiest way to find correct seeds:

1. Need today have a heavy shower, and best is not change the time.
2. Ingame take screenshot when each shower dropped, about 10 minutes / 50 showers.
3. Use a card reader to copy the screenshot file to PC, then check the "File Created Time", or just use a timestamp in the filename. I just fill in this information, no another.
4. It must be done for only one seed.

[Colin5887]

I personally believe this is a useful but redundant function. This function revolves too many works:

• We need to save these seed data
• We need to detect if any of the previous weather data is changed.
• We need to automatically update the saved data.

By far, the program works and the most time-consuming part is to input enough weather data instead of calculating. I believe it worth is to wait for a few more minutes (or maybe seconds if you have a powerful CPU) when you've carefully recorded weathers for a couple of days.

[Treeki]

This is why I haven't done it yet - even on my phone from 2017, the algorithm is fast enough that it's not too much of an issue.

It does get a fair bit slower when making heavy use of shooting star times (due to the extra calculations involved), so there's possible room for improvement there.

It feels like incremental scanning would only help so much, though, as if you change any existing data the scans will become invalid, and in many cases the scan will simply fail early if your data returns too many possible seeds.

I am not entirely opposed to adding this, but I would rather focus on other features first