Misleading description for the -I argument of gmt info

Description of the problem

I find the description for -I argument of gmt info may be a little misleading for new GMT users (originally posted at https://github.com/GenericMappingTools/pygmt/issues/1035).

It seems that only when the -C argument is combined with -I, then the output will be rounded up/down for as many columns as there are increments provided in -I. Without -C, -I only outputs the min/max values for the first two columns no matter how many increments we input. So, we'd better always add -C when we use -I unless only the first two columns are needed.

Test

I use Iris flower dataset downloaded from https://github.com/mwaskom/seaborn-data/raw/master/iris.csv to test -I and -C.

# Download the sample iris data:
$ wget https://github.com/mwaskom/seaborn-data/raw/master/iris.csv

# Output with -I and -C:
$ awk -F "," 'FNR>1 {print $4,$1,$3}' iris.csv | gmt info -C -I1/0.2/0.3
0   3   4.2 8   0.9 7.2

# Output with only -I:
$ awk -F "," 'FNR>1 {print $4,$1,$3}' iris.csv | gmt info -I1/0.2/0.3   
-R0/3/4.2/8

# Output with only -C:
$ awk -F "," 'FNR>1 {print $4,$1,$3}' iris.csv | gmt info -C            
0.1 2.5 4.3 7.9 1   6.9

# Output with neither -I nor -C:
$ awk -F "," 'FNR>1 {print $4,$1,$3}' iris.csv | gmt info   
<Standard Input>: N = 150   <0.1/2.5>   <4.3/7.9>   <1/6.9>

GMT DOC

The description of gmt info at GMT DOC :

As another option, info can find the extent of data in the first two columns rounded up and down to the nearest multiple of the supplied increments given by -I.

This is right. When we only use -I without -C, only the rounded min/max values of the first columns will be output.

If -C is combined with -I then the output will be in column form and rounded up/down for as many columns as there are increments provided in -I.

This is exactly right. When we use both -I and -C, we can output the rounded min/max values for each column provided in -I.

The detailed description of -I at GMT DOC:

Report the min/max of the first n columns to the nearest multiple of the provided increments (separate the n increments by slashes), and output results in the form -Rw/e/s/n (unless -C is set). If only one increment is given we also use it for the second column (for backwards compatibility).

The first part of the first sentence assumes -C is used, while the second part assumes -C is not used.

Suggestion

It's better to revise the description of -I option at GMT DOC to explicitly indicate the difference for the two cases, i.e., -C is used and is not used. I'd like to help this issue if you think we need to revise a little about -I.

System information

Operating system: CentOS 7
GMT version (gmt --version): 6.1.1

GenericMappingTools / gmt

Misleading description for the -I argument of gmt info #4938