confusing example, rename data.sizes??

Hi @Anirban166 I read your doc blog about expect_x_time and that is helpful.

  expect_loglinear_time(PeakSegPDPA(rpois(data.sizes, 1), rep(1, length(rpois(data.sizes, 1))), 3L), data.sizes = 10^seq(1, 4, by = 0.5))

what is data.sizes in this example? It appears three times so I think this is potentially confusing. In the first argument data.sizes is ONE of the values that are specified by the named data.sizes= argument, right? Because it is plural (sizes) I would think it is a vector of more than one number. To avoid that confusion can you change it to N instead, as we discussed?

  expect_loglinear_time(PeakSegPDPA(rpois(N, 1), data.sizes = 10^seq(1, 4, by = 0.5))

Also to clarify the example can you please use a variable name as argument to PeakSegPDPA, and omit the second argument? e.g.

expect_loglinear_time({
  data.vec <- rpois(N, 1)
  PeakSegPDPA(data.vec)
}, data.sizes....)

does that make sense to you?

Hi @Anirban166 I read your doc blog about expect_x_time and that is helpful.
  expect_loglinear_time(PeakSegPDPA(rpois(data.sizes, 1), rep(1, length(rpois(data.sizes, 1))), 3L), data.sizes = 10^seq(1, 4, by = 0.5))
what is data.sizes in this example? It appears three times so I think this is potentially confusing. In the first argument data.sizes is ONE of the values that are specified by the named data.sizes= argument, right? Because it is plural (sizes) I would think it is a vector of more than one number. To avoid that confusion can you change it to N instead, as we discussed?
  expect_loglinear_time(PeakSegPDPA(rpois(N, 1), data.sizes = 10^seq(1, 4, by = 0.5))

Yes your right, N would look less confusing than the repeated usage of data.sizes in the expression. There are two simple ways I can go about this in asymptoticTimings based on the arguments/parameters, so I would like to know which one is better:

Parameters remain the same, and internally I can assign N as data.sizes, and then subsequently create the function accepting the expression with N as parameter: (I think this is the better way due to no additional parameters)
```
asymptoticTimings <- function(e, data.sizes, max.seconds)
{ ...
N <- data.sizes
fun.obj  <- function(N)
eval(lang.obj)
...
}
```
Add N as a parameter, assign it to data.sizes as a default: (thought of this because user might not know to put only N in the expression, and I can add that in the roxygen documentation for the parameter N)
```
asymptoticTimings <- function(e, data.sizes, N = data.sizes, max.seconds)
{ ...
fun.obj  <- function(N)
eval(lang.obj)
...
}
```
Both will work, but are one of these what you wanted, or should I think of something else?

Also to clarify the example can you please use a variable name as argument to PeakSegPDPA, and omit the second argument? e.g.
expect_loglinear_time({
  data.vec <- rpois(N, 1)
  PeakSegPDPA(data.vec)
}, data.sizes....)
does that make sense to you?

Yep, perfectly does! but in that code above we would require to give a value to max.segments atleast right (and not just include data.vec)? Because otherwise it would throw an error for max.segments (like its not an integer), since in your code the default is set to NULL

works fine:

will make the changes to code & blogs but please check if anything else needs to be improved as well, also I guess I should make the same changes for the cDPA test case as well? (and other functions as well)

Anirban166 / testComplexity

confusing example, rename data.sizes?? #46