How Microsoft Research can save Excel

Google just announced BigQuery-backed spreadsheets that can handle a billion rows. At Faraday, we love BigQuery and we are already putting this into use. 

But not everybody can jump to BigQuery and Google Sheets - yet. So we would like to take this opportunity to offer Microsoft some advice if it wants to keep Excel as everybody's favorite tool:

1. Your overriding priority should be that Excel, on the desktop, load gigabyte-sized files instantly. Obviously it would be lazy loading, you could build it in Rust, I'm sure there are some extreme algorithms that could be applied. Isn’t this a perfect project to show off Microsoft Research’s prowess?
2. Please handle dates like a sane person, e.g., YYYY-MM-DD ISO 8601. You can keep your special support for bizarre American-style M/D/YY, but consider a different default when exporting to CSV.
3. If a cell contains leading zeroes - like 05401, zip codes - treat them as strings and keep the leading zero. Dropping the leading zero is an ancient assumption from fixed-length files, and it’s wrong.

Microsoft Research has some of the world's most impressive computer scientists. Rebuilding Excel's backend to handle 100x more data would do a lot for maintaining the viability of Excel. And the other fixes mentioned here would make Excel feel more modern.