I had the following code:
set swapDict [fillSwapDict]
set swapDict [mergeSwapDict ${swapDict} ${mergeList}]
set swapList [sortedList ${swapDict}]
displayList ${swapList} ${end}

Each call does use the data of the call before it. So I thought it is
better to write it like this:
displayList [sortedList [mergeSwapDict [fillSwapDict] ${mergeList}]] ${end}

Which is the 'better' way, or is another way better?

--
Cecil Westerhof
Senior Software Engineer
LinkedIn: http://www.linkedin.com/in/cecilwesterhof

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Cecil Westerhof <[email protected]> wrote:

I had the following code:
set swapDict [fillSwapDict]
set swapDict [mergeSwapDict ${swapDict} ${mergeList}]
set swapList [sortedList ${swapDict}]
displayList ${swapList} ${end}

Each call does use the data of the call before it. So I thought it is
better to write it like this:
displayList [sortedList [mergeSwapDict [fillSwapDict] ${mergeList}]] ${end}

Which is the 'better' way, or is another way better?

There isn't. You are slowly moving closer to merely a "stylistic"
issue on the order of whether this:

if (...) {
...
}

or this

if (...)
{
...
}

Arrangement of the openbrace for a C if clause is the "better way".

The second is more LISP like, and avoids having to allocate, and
ultimately destroy, three variables. But you will find people who
argue with convincing reasons why either is 'better' than the other.

Use whichever you prefer, and don't fret over it. Myself I use the
LISP like layout mostly, but will use the tempvar version on occasion
(and esp. if one of the tempvar version can avoid processing the same
data into the same result two or more times).

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Rich <[email protected]d> writes:

Cecil Westerhof <[email protected]> wrote:

I had the following code:
set swapDict [fillSwapDict]
set swapDict [mergeSwapDict ${swapDict} ${mergeList}]
set swapList [sortedList ${swapDict}]
displayList ${swapList} ${end}

Each call does use the data of the call before it. So I thought it is
better to write it like this:
displayList [sortedList [mergeSwapDict [fillSwapDict] ${mergeList}]] ${end}

Which is the 'better' way, or is another way better?

There isn't. You are slowly moving closer to merely a "stylistic"
issue on the order of whether this:

if (...) {
...
}

or this

if (...)
{
...
}

Arrangement of the openbrace for a C if clause is the "better way".

Personally I prefer the first, but if I would work somewhere where
they prefer the second, I would use the second.

The second is more LISP like, and avoids having to allocate, and
ultimately destroy, three variables. But you will find people who
argue with convincing reasons why either is 'better' than the other.

Yes, that is why I prefer the second. But when sharing your code (what
I intend to do) it is best to use a style that most people understand
better.

Use whichever you prefer, and don't fret over it. Myself I use the
LISP like layout mostly, but will use the tempvar version on occasion
(and esp. if one of the tempvar version can avoid processing the same
data into the same result two or more times).

I added extra functionality and for that I put the first tempvar back.
Shaved about a third of the execution time.

--
Cecil Westerhof
Senior Software Engineer
LinkedIn: http://www.linkedin.com/in/cecilwesterhof

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

At Sat, 7 May 2022 12:39:07 -0000 (UTC) Rich <[email protected]d> wrote:

Cecil Westerhof <[email protected]> wrote:

I had the following code:
set swapDict [fillSwapDict]
set swapDict [mergeSwapDict ${swapDict} ${mergeList}]
set swapList [sortedList ${swapDict}]
displayList ${swapList} ${end}

Each call does use the data of the call before it. So I thought it is better to write it like this:
displayList [sortedList [mergeSwapDict [fillSwapDict] ${mergeList}]] ${end}

Which is the 'better' way, or is another way better?

There isn't. You are slowly moving closer to merely a "stylistic"
issue on the order of whether this:

if (...) {
...
}

or this

if (...)
{
...
}

Arrangement of the openbrace for a C if clause is the "better way".

The second is more LISP like, and avoids having to allocate, and
ultimately destroy, three variables. But you will find people who
argue with convincing reasons why either is 'better' than the other.

Use whichever you prefer, and don't fret over it. Myself I use the
LISP like layout mostly, but will use the tempvar version on occasion
(and esp. if one of the tempvar version can avoid processing the same
data into the same result two or more times).

The only other issue is source code readablity. *Sometimes* you end up with either a really long line (which does not "fit" on your screen, unless you
have an insanely large edit windows and/or are using a really small font), or an insanely large number of continue lines (and possibly really large indenting). Sometime breaking up a large and complex expression into subexpressions results in code that is easier to read and understand (and in some cases easier to debug).




--
Robert Heller -- Cell: 413-658-7953 GV: 978-633-5364
Deepwoods Software -- Custom Software Services
http://www.deepsoft.com/ -- Linux Administration Services
[email protected] -- Webhosting Services

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 5/7/2022 7:32 AM, Robert Heller wrote:

The only other issue is source code readablity. *Sometimes* you end up with either a really long line (which does not "fit" on your screen, unless you have an insanely large edit windows and/or are using a really small font), or an insanely large number of continue lines (and possibly really large indenting). Sometime breaking up a large and complex expression into subexpressions results in code that is easier to read and understand (and in some cases easier to debug).

Cecil:

I totally agree with Robert, especially with debugging.

I will often go even further and create more specific temporary
variables, since it often results in more self-explanatory code:

set filledDict [fillSwapDict]
set mergedDict [mergeSwapDict ${filledDict} ${mergeList}]
set swapList [sortedList ${mergedDict}]


But the primary reason is that if you create one functional style
statement with nested proc/command calls, it becomes quite difficult to
tell which of the intermediate steps isn't working correctly when the
final result is not as expected.

To that end, with my editor, I can do 4 drag and drop copies of those
variables onto a single line, like so:

filledDict mergedDict swapList mergeList

then a triple click on the line selects the 4 and 1 keystroke launches
an external editor tool where the 4 variables are sent to a tcl program
and returned with this text to replace the 4:

puts "filledDict= |$filledDict| mergedDict= |$mergedDict| ..."

I might then add some newlines to this, or [lsort] them etc.

Given the speed of computers these days, I seldom write more than a few
lines of code before I hit my launch key in the editor which will write
out the file and launch it as the arg to a tclkit. In about 1 second, I
will have a console window in front of me with the output of that puts
right there. A [vwait ::ffff] can work like a breakpoint there as well.

I find that this is far easier and quicker than to navigate to the code
in some sort of single stepping and break-point setting debugging tool.

And when done, I comment them out, and usually right indent the puts
past say column 50 so they're still right there if needed again while
not making the real code hard to see.

I tend to build bottom-up and so I develop small functional procedures
that I can grow and debug independent of the rest of some program. Later
I glue the pieces together after I know they are all working correctly.

Hope this helps.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

This is Tcl. If you don't like the way it looks, create commands to make it look the way you want. Perhaps the wiki page on pipeline programming has some ideas of one way to deal with deeply nested commands.
https://wiki.tcl-lang.org/page/Pipeline+programming.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 5/7/2022 8:02 PM, Robert Heller wrote:

The only other issue is source code readablity. *Sometimes* you end up with either a really long line (which does not "fit" on your screen, unless you have an insanely large edit windows and/or are using a really small font), or an insanely large number of continue lines (and possibly really large indenting). Sometime breaking up a large and complex expression into subexpressions results in code that is easier to read and understand (and in some cases easier to debug).

While I agree with the above in terms of readability, note that
temporary variables can have an significant impact on performance when
dealing with even moderately large data. It's not the temporary variable
itself but rather Tcl's reference counting and COW implementation means
that the additional variable reference does not allow Tcl's in-place modification optimization. Consider for example,

% proc p {} {lrepeat 10000 x}
% timerate -calibrate p; # Tell timerate to ignore overhead of p - not
relevant
46.01236794640557 µs/#-overhead 48.2837 µs/# 40428 # 20710.9 #/sec
% timerate {linsert [p] end x}
4.052869 µs/# 19974 # 246738 #/sec 80.952 net-ms
% timerate {set l [p]; linsert $l end x}
59.1994 µs/# 9505 # 16892.1 #/sec 562.690 net-ms

The temp version is slower by an order of magnitude. Of course, this is
only true when the value returned by the procedure does not have another reference elsewhere in which case the additional temp reference will not matter. It's just another consideration to keep in mind.

/Ashok

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)