> and they left that many errors in.

How many of those errors are visible to end users in a way that actually matters? A 'suspicious' floating point conversion in a display aspect ratio check? A memory leak in a program that's almost entirely interactively driven? I don't see anything else terribly huge on the list either.

Software errors have an impact to the user and a cost to fix. These don't seem like errors where the ROI on the fix was likely to be positive, so maybe it's reasonable they were left in.

You should also consider the errors that Microsoft _removed_ from the calculator before passing too much judgement on their software development process. After all, this is the same company that replaced IEEE754 floating point in calculator with an arbitrary precision library. (At considerably more expense than any of the bug fixes indicaetd by the attached article)

https://blogs.msdn.microsoft.com/oldnewthing/20040525-00/?p=...

> considering this application Calculator could be one of the most-used programs on Windows

I'd seriously doubt that. I'm sure browsers and word processors are used far more often, and that's not even including programs that run every time the computer boots!

Agreed. Calculators in general, yes. The Calculator app on my PC, not so much. I’d also assume people using calculators that much would buy a physical one, it’s much more convinient.

I have a calculator button on my keyboard right above the ten-key. When I push it it opens my calculator app or brings it to foreground if it's already open. I consider it extremely convenient since I find I need a calculator often, but not often enough to justify leaving one always on my desk. Also I can copy and paste answers/intermediary results easily and it does conversions automagically (e.g. "5 cups in ounces").

Physical calculator? You mean a phone right? or just Googling the equation? or saying okay google and asking the equation to your phone so you can just say the words billion and not type like 9 fucking zeroes or something?, or use desmos? or wolframalpha?

Physical calculators are for classrooms.

A phone is in almost all regards worse than using a calculator with actual buttons. Voice is even less efficient. The major reason to use a physical calculator at your desk is because it's a dedicated device that's always ready, and fast. Anecdotally, being able to use it efficiently one-handed also plays a role.

BTW, I think that python/ipython cli is a much better calculator. I stopped using other calculators long time ago. Although I understand that it's probably too complicated for ordinary user.

To be frank, I do the same with the GNU Octave CLI, which is IMO much more suited to evaluating mathematical expressions, and very short programs. Out-of the box, that is.

The right answer is probably to use the right tool for the job, be it xcas, octave or python. Microsoft's calculator is probably better suited for unit conversion, for instance, although units(1) would be a contender :)

There is no task executed in Excel that I can't do better and/or faster in ipython.

- Pivot Tables

- Reach goal.

- Simple curve fitting

- Drawing charts in general (even with matplotlib categorical bat charts are a world of pain and then they tell you to learn Altair or something.

- Tables with conditional formatting

- ...