Hacker News Reader

Comment by zozbot234 | Hacker News Reader

Excel is a disaster. It will randomly corrupt input that happens to fit certain patterns by parsing it into "dates" or "numbers", unless you go to the trouble of explicitly setting the right "cell properties" via the incredibly clunky UI. It hides the code that's input in cells, lets it randomly mix with input data and does not support any kind of source control for it. It's not even fit for purpose as a prototyping tool, you're literally better off writing some code in FORTRAN IV and key-punching it into a fresh deck of punched cards. At least you're going to have proper code/data separation.

View on HN

Yup, but you can make it do the job and that's enough for the majority.

It's only "enough" to them because they don't grok the failure modes. Which is basically the story of almost all "no-code" or "low-code" tools.

You can do "low-code" and even "no-code" right, but it basically has to be a DSL that's cleanly defined in terms of some other programming language with sensible semantics. That way when the DSL fails to do what you expect it to (as it invariably will at some point), you at least get to pick up the pieces and carry on. But that's not how Excel is built, or most similar tools.

Preaching to the choir, but mostly they don't care about any of that.

I mean, we're renaming genes because of how much excel has been corrupting data.

https://www.theverge.com/2020/8/6/21355674/human-genes-renam...

I agree that there's tons of value in the paradigm (more than a lot of people on hn would admit), but excel particularly has way more sharp edges than people know.

It's not just that it can do the job, but that the majority has access to Excel. There may be better tools/platforms to solve a given problem, but in most cases the rank and file don't have access to them or permissions to do what they want.

"If the only tool you have is a hammer, you tend to see every problem as a nail."

Reply on news.ycombinator.com