Live data from Hacker News

You shouldn’t use a spreadsheet for important work

lemire.me

21–30 of 200 posts

Re: You shouldn’t use a spreadsheet for important work

#23
post #11

Honestly, I don't have a problem with the fact that Excel is used in so many cases - programming for non-programmers is important. My problem is that Excel is terrible . Its formula system is painful and the formulas are invisible. It's not that a light user DB/spreadsheet program is a bad idea, it's that Excel made a lot of terrible decisions 20 years ago and now they're married to them.

> Excel made a lot of terrible decisions 20 years ago and now they're married to them I can't really think of any off the top of my head, would be interested if you could share one or two that they're permanently married to. However, I could make a very long list of easy to implement fixes & features they could add that would offer different ways of doing things than the braindead way you have to do things now, while…

Excel doesn't support dates before 1900-01-01, and it incorrectly treats the year 1900 as a leap year[1]. This bug-as-a-feature started in the first Excel for the sake of compatibility with Lotus files.

[1] http://support.microsoft.com/kb/214326/en-us

EDIT: Apparently Wikipedia has a list of "quirks". I haven't vetted any of the citations, but the page claims issues with statistical function accuracy, failures with modulo operations, confusing use of numeric precisions, and an inability to open two identically-named files

Re: You shouldn’t use a spreadsheet for important work

#24
having worked for Investment Banking clients I can see what the author intended to address.

I have seen a lot of abuse/ over use of excel/VBA

I once had to write CRUD front-end with Excel/VBA activeX components retrieving data from Sybase / SQL Server databases.

Which could have been easily engineered as a simple web-app. "But NO .. IT HAD TO BE A FUCKING SPREADSHEET"

Re: You shouldn’t use a spreadsheet for important work

#26
post #25

A thought experiment: is it advisable to save mission critical, structured information as a loose collection of schemaless text files? If it is not advisable, why is it accepted for software engineering professionals to manage their source code this way?

well a programming language with a specification has a pretty well-defined schema.

The E?-BNF provides a syntactical schema for the code and the rest of specification provides the semantic schema.

Re: You shouldn’t use a spreadsheet for important work

#28
post #11

Honestly, I don't have a problem with the fact that Excel is used in so many cases - programming for non-programmers is important. My problem is that Excel is terrible . Its formula system is painful and the formulas are invisible. It's not that a light user DB/spreadsheet program is a bad idea, it's that Excel made a lot of terrible decisions 20 years ago and now they're married to them.

> Excel made a lot of terrible decisions 20 years ago and now they're married to them I can't really think of any off the top of my head, would be interested if you could share one or two that they're permanently married to. However, I could make a very long list of easy to implement fixes & features they could add that would offer different ways of doing things than the braindead way you have to do things now, while…

Little things like how terrible and confusing their formula language is. Anything beyond simple arithmetic requires two Googles per cell.

Now, replacing it with Python isn't better - any language that uses double-equals-signs is not fit for a layman-programming-platform.

Re: You shouldn’t use a spreadsheet for important work

#29

Data in arrays, exports to CSV, facile API... what's not to love?

I remember on job where I wrote a lot of tools to load a database by reading an excel spreadsheet from perl[1]. It was amazingly fun.

1) perl does have some good libraries to read Excel files without exporting

Re: You shouldn’t use a spreadsheet for important work

#30

Earlier quoted context omitted.

I loved this article [3] from 2008, that argues that Excel and VBA were responsible for the credit crunch That's a rather aggressive phrasing of what the article says, that many of the calculations involved in the credit crunch were done in Excel/VBA.

Yes, I agree its a bit aggressive. Ultimately it was human nature... spreadsheets don't cause apocalyptic financial crises, overzealous capitalists with spreadsheets cause apocalyptic financial crises! The author was (is?) a recruiter for quants and financial developers so he certainly had a unique perspective.

Is it just me, or does the financial system seem to be designed in precisely the opposite way of how we try to design good software? One huge intertwined web, where one small thing indirectly affects thousands of other things. Its seems to be the equivalent of a "big ball of mud" type piece of software.
Post reply on HN