[somat]

it's weird but Visidata is very nearly my idea of a perfect spreadsheet. "But!", you exclaim, "Visidata is not a spreadsheet", I know, I know, that is what makes it so weird. Let me explain.

I am not fond of the the usual spreadsheet data model. "It's a big bag of cell" does not fill me with joy. And upon a bit of reflection I think it is the rows, I really hate how easy it is for rows to get out of sync. All I really want is row security.* And this is what visidata brings to the table.

* Relational databases provide this in spades. and in truth most of my spreadsheets have been replaced by them(I maintain a local postgres server on my desktop for all the small random prototype junk you would usually do in a spreadsheet using visidata as a pager) And while the database is great for analysis, random data entry sort of sucks. There are some great tools out there for this. I don't tend to use them, mainly because psql is always there, so I just sort of grumble when I have to do random entry without trying to fix anything. Why postgres instead of sqlite? I like the types and functions better.

[cykros]

It's all about picking the right tool for the job. I don't see easily constructing a Discounted Cash Flow model in a relational database making a ton of sense. Sure, you could have the data in one, and use a php script to handle the modeling, and output it on a webpage, but the reality is, a spreadsheet just does it a lot easier, and is what most people would want.

For handling structured data, spreadsheets can do it, but they're not always going to be the best tool for the job, as they do a lot of other things, too, which means if you're not careful, your data will lose its appropriate structure.