| modern syntax [message #92326 is a reply to message #92325] |
Wed, 18 November 2015 03:47   |
greg.addr
Messages: 160
Registered: May 2007
|
Senior Member |
|
|
> If you're running 8.4 or later, I encourage more modern syntax, for example
> IDL> x='Run0'+variable1.tostring()+'_Seq0001_Scan'+variable2.tost ring()
> IDL> x
> Run012345_Seq0001_Scan345
>
> The .tostring() static method on IDL_Variable types implies trimming of both leading and trailing blanks.
>
> For more on static methods, see http://www.exelisvis.com/docs/IDL_Variable.html
>
> Jim P.
I've been wondering about this recently. I admit it's rare that I use any language but IDL, so I'm sure I'm behind the times. Consistency is a good thing in programming, so I'm uneasy with a pick and mix approach to the new syntax: it can only leave code less readable. I understand the sense of the object approach, so I'd go with the new, but the notation - to my eyes - is getting worse:
variable1.tostring() vs string(variable1)
- the parentheses have become a relict, symbolically enclosing an argument which has gone elsewhere. To compensate the loss of meaning, we need to tack on a new preposition, 'to'. Ugh.
In some cases, though, the new syntax is elegant:
variable.tname vs size(variable,/tname)
although the old notation was unnecessarily blighted by the need to access this through the size function. I haven't yet grasped why variable.tname doesn't need (or allow) the following parentheses, since it appears to me to return a value and therefore qualify as a function. But I like it better without.
IDL> a={b:0,tname:4}
IDL> a.tname
4
IDL> (a).tname
% Object reference type required in this context: A.
I don't know what to do here, though. Isn't 'a' also an object now?
Then there's the grotesque,
PRINT, var1.Equals(var2)
I don't want to write that.
I'd be interested to know how others choose which to use.
cheers,
Greg
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
