bc: add a line about interactive usage with mathlib

[kalebo]

Usually bc isn't very useful if it doesn't return floating point numbers where appropriate. e.g., 3/4 returns 0 when started with -i but for most users they would expect 0.75 to be returned so the tldr should mention that the argument -ldoes floating point arithmetic.

[mention-bot]

@kalebo, thanks for your PR! By analyzing the history of the files in this pull request, we identified @igorshubovych and @rubenvereecken to be potential reviewers.

[waldyrious]

LGTM 👍

[agnivade]

I am a bit confused on this.

The man page for bc on my ubuntu box says -l, --mathlib- Define the standard math library. for this option. But the PR mentions in interactive mode with floating point support. First of all, will -l invoke interactive mode too ? And are floating point support and math library one and the same ? If yes, then we should use the later to denote [l].

[kalebo]

@agnivade Yes, both -i and -l put bc into interactive mode. And yes, -l really means load the std math library. I agree the wording could be clearer.

The main issue that I see is that most people assume that a calculator should handle things like floating point division by default, but bc does not do that without specifying -l. I would almost say it's better to remove the line about -i and replace it with a line that uses -l and perhaps state that it puts in into interactive mode with standard math libraries and floating point support. Or something to that effect.

[agnivade]

No worries, just use f[l]oating and we should be good.

[waldyrious]

IIUC the "l" in -l refers to "lib(rary)", right? How about "Run calculator using the standard math library, for floating point support (this enables interactive mode)"?

[kalebo]

I reworded the entry as I realized that floating point support does not come from mathlib.You can run scale=20 and get the same result in normal mode. However, enabling mathlib support also sets scale=20 automatically. Ideally if a user just wants to run some quick calculations, including -l is easier than including scale=x which is more of a feature you'd want if you were writing a script. For this reason I replaced the line about -i with the one about -l since most users would expect their calculations not to be truncated at the decimal.

[waldyrious]

What exactly is the scale? Number of decimal places for calculation results?

[agnivade]

Hey @kalebo - Could you update the scale example to include the word 'decimal' too ? 'decimal precision' sounds much more clear than just 'precision'.

[kalebo]

I switched up the wording in the scale example. I'm a little torn between leaving in that example and ditching it. In short I don't want to make the tldr too in depth but bc's defaults are a bit of a stumbling block that I think would be nice to warn people about.

[agnivade]

Sure, LGTM.