@Karen I guess that means zero chance of it being fixed in Maple 2017 then?  Good that it's looked after.

Well this is interesting ..

I suppose the evalf(convert(2.59/3.7),fraction) is a workaround.

**

parse(`2.59/3.7`)   #parse is also an alternative for a workaround
              .7000000000

I've always used Firefox never had a problem except in the past it has been slow but it's much faster now.  Some days, recently, it's been very quick.  Today is no exception, seems quick right now for me.  I'm using Firefox quantum 64 bit version 57.0.2.  I think Firefox updated to quantum a couple of months ago - ironically I didn't notice it being any faster.  Even before quantum, mapleprimes at some times was fast, and then on and off it's been fast and back to normal which I like to refer to as being the slow days of the past - not painfully slow, but just not quick.  Could be my internet connection but it's pretty quick.

Sometimes it's tough to avoid a title that might be percieved by others as not being legitimate, which is why it's important to skim the content before making a snap decision. 

Looks like it's off by 1/2

Ah, I didn't realize typsetting level changed.

Maple 2017.1  Build 1238644 Typesetting level=Extended
Maple 2016.1  Build 1132667 Typesetting level=Maple Standard

..and thanks for the explanations although 2017 changing to Maple standard produces output b:= without the () while 2016 just produces a blank line.  Not a problem, but was perplexed as to why the behavior changed.

kernel density looks like what I'm after, thanks.

**edit added** Actually KernelDensitySample is more like what I was looking for but thanks for pointing out the KernelDensity

Of course small sample sizes decrease the statistical power.  Small sample sizes increase the likelihood of Type II errors skewing results which decrease the power of the study. 

If researchers are constrained to small sample sizes they may have to settle for less conclusive results.  Whether this issue is important or not depends ultimately on the size of the effect being studied.  ie. A small sample size would give more meaningful results in a poll of people living near an airport being negatively affected by air traffic than it would in a poll on their education levels.

@mapleatha 

Yes I see.  The empty symbol `` is not printed in the output.  If you wanted to continue using the output then strange behaviour is expected as you've mentioned. 

use lprint on the output and you will see why it is counted by nops.  I made g our modified output.

Notice the empty symbol in the output of g.  I can't think of a way around using that output without Maple reverting your manipulated equation back to the way it was originally.

I offer one possible idea .... that is to only use the ``output for display purposes, and just continue to use f (the original output) for further calculations.

@mapleatha ok no problem.  Just wanted to make sure you were able to get it to work on Maple 13.  On my M13 op(2,f)*``op(1,f) works.  

Not sure if this will work for you but moving `` in front of everything works as well  ``op(2,f)*op(1,f)

@mapleatha You forgot `` in front of the second op.  Put `` in front of the second op and it will work in Maple 13

@mapleatha .. no you don't need a special package.  Different versions of Maple sometimes do slightly different things with the same operations (ie order and sorting) sometimes it's not cut and dry to try to manipulate Maples output to the way you want. 

I used Maple 2017 when I came up with my answer above.  Just tried it on M12, unfortunately my solution above leaves it in the expanded form. 

To me it seemed they could be used interchangebly.  More of an unsure query than a statement.  I wasn't too sure but I think a viable answer, no?

..or symbol.  Funny I've always used symbol, but seems name and symbol can be used interchangeably. 

It is done here in mathematica

https://mathematica.stackexchange.com/questions/9327/is-it-possible-to-produce-anaglyphs-with-mathematica

Yes, I recall _Maxim_ wrapping call suggestion.  Although, I think if a large error is going to pop up, an internal routine (like the try catch) should take over and display only the first bits (maybe 10 lines - maybe that's too much) of information and terminate the rest as a returned display.