I'm really excited about the interactive widgets. I've been curious about the idea of creating a spreadsheet widget to provide excel-like functionality in Jypyter.
The thing about Excel is there are many 50~80% use case implementations. But none with all 30 years of Microsoft's spreadsheet experience baked in. The long tail use cases are where Excel excels, and in spreadsheets there are a lot of snowflake use cases.
I don't really see why you need those advanced features. In my experience, as soon as I'm requiring non-basic features, I actually shouldn't be using spreadsheets.
The biggest thing about "Excel" and spreadsheets is the work you do is not reproducible and it is the main reason why you should program all your work. What people do with macros and spreadsheets is very dangerous and unless it is documented it has too much human error capability.
The business case for using Excel overwhelms engineering best practices up to somewhere around five or six sigmas. The big ball of mud software architecture is good enough to make money. Anyway explaining business logic to a technical programmer is also fraught with human error. TANSTAAFL.
Not necessarily. There have been a number of disasters in the last few years caused by errors that have been directly attributed to spreadsheet usage. Many financial firms are reining in excel use by their business users. Prototyping a solution is fine, but for production use they are requiring an auditable solution that has proper source code control and can't be tweaked by the end user.
Spreadsheets are great products but for production use they are really two edged swords.
Sure, but that shift is happening under the same logic that leads businesses to run ERP applications on big iron mainframes even here in the 21st century.
It's a matter of degree, but human error remains. The subprime clusterfuck wasn't based on using Excel.
