strategys other than subroutine and OO?

[Ela]

I have 300 lines of codes that are used roughly twice. Some of the lines
have delicate difference in execution (e.g. ~6 variables have to be replaced
and conditional statements are also different) and therefore many arguments
need to be passed into it in order to differentiate the execution parts. OO
development is analogously also difficult. I wonder if there is any good way
to reuse the codes, e.g. some sort of "goto" rather than duplicating these
lines. Whenever there is any change, I find that is really error-prone.

 

[RedGrittyBrick]

I have 300 lines of codes that are used roughly twice. Some of the lines
have delicate difference in execution (e.g. ~6 variables have to be replaced
and conditional statements are also different) and therefore many arguments
need to be passed into it in order to differentiate the execution parts. OO
development is analogously also difficult. I wonder if there is any good way
to reuse the codes,

There are good ways and there are easy ways. Choose good.

e.g. some sort of "goto" rather than duplicating these
lines. Whenever there is any change, I find that is really error-prone.

"For a number of years I have been familiar with the observation that
the quality of programmers is a decreasing function of the density of go
to statements in the programs they produce. More recently I discovered
why the use of the go to statement has such disastrous effects, and I
became convinced that the go to statement should be abolished from all
"higher level" programming languages (i.e. everything except, perhaps,
plain machine code). At that time I did not attach too much importance
to this discovery; I now submit my considerations for publication
because in very recent discussions in which the subject turned up, I
have been urged to do so."

Edager W. Dijkstra. Go To Statement Considered Harmful. 1968.



In nearly 30 years of programming I have completely avoided use of GOTO
in a professional context. All I have seen continues to convince me that
Dijkstra was right on this matter. YMMV.

 

[Joost Diepenmaat]

I have 300 lines of codes that are used roughly twice. Some of the lines
have delicate difference in execution (e.g. ~6 variables have to be replaced
and conditional statements are also different) and therefore many arguments
need to be passed into it in order to differentiate the execution parts. OO
development is analogously also difficult. I wonder if there is any good way
to reuse the codes, e.g. some sort of "goto" rather than duplicating these
lines. Whenever there is any change, I find that is really error-prone.

Sounds like using closures/passing functions as arguments could be very
helpful here, but without seeing your code it's impossible to be sure.

 

[Willem]

Ela wrote:
) I have 300 lines of codes that are used roughly twice. Some of the lines
) have delicate difference in execution (e.g. ~6 variables have to be replaced
) and conditional statements are also different) and therefore many arguments
) need to be passed into it in order to differentiate the execution parts. OO
) development is analogously also difficult. I wonder if there is any good way
) to reuse the codes, e.g. some sort of "goto" rather than duplicating these
) lines. Whenever there is any change, I find that is really error-prone.

Maybe you can take the lines that are exactly duplicated and put those
into sub functions. If you define the sub functions inside the main
function they can access the main function's 'my' variables, so you
don't need to pass them as arguments.


SaSW, Willem
--
Disclaimer: I am in no way responsible for any of the statements
made in the above text. For all I know I might be
drugged or something..
No I'm not paranoid. You all think I'm paranoid, don't you !
#EOT

 

[Ela]

Maybe you can take the lines that are exactly duplicated and put those

into sub functions. If you define the sub functions inside the main
function they can access the main function's 'my' variables, so you
don't need to pass them as arguments.

From perlintro:
A Perl script or program consists of one or more statements. These
statements are simply written in the script in a straightforward fashion.
There is no need to have a main() function or anything of that kind.

So would you mind giving a very simple example? i guess your suggestion is
very close to the final solution since i never have the concept you suggest.

 

[ccc31807]

I have 300 lines of codes that are used roughly twice. Some of the lines
have delicate difference in execution (e.g. ~6 variables have to be replaced
and conditional statements are also different) and therefore many arguments
need to be passed into it in order to differentiate the execution parts. OO
development is analogously also difficult. I wonder if there is any good way
to reuse the codes, e.g. some sort of "goto" rather than duplicating these
lines. Whenever there is any change, I find that is really error-prone.

I have a suggestion. Since you didn't post a coding sample, my
suggestion may not be appropros, but maybe it will.

First, you surely don't have a function that is 300 lines long. If you
do, I don't think I can help. You probably have batches of lines that
are duplicated. The first thing I would do is abstract these batches
of lines into separate functions and call them separately, like this:

&batch_a;
&batch_b;
&batch_c;
sub batch_a { #first batch of lines
}
sub batch_b { #second batch of lines
}
sub batch_c { #third batch of lines
}

Second, you can pass different parameters into these functions to
account for the different variables. If you need to set a variable as
a result of some behavior, set the variable by calling a function,
like this:

$var = &setvar;
sub setvar { #code here sets a var which you would return
return $var; }

Third, you can replace your conditional statements with booleans and
set them outside the function call, like this:

my $conditional = 0
$conditional = 1 if &some_function_that_returns_true;
$contitional = 0 unless &some_function_that_returns_true;
if ($conditional) { #code to execute
}

Fourth, if possible, place your common code (our variables and
functions) in a PM and import the backage. This way, you can separate
the interface and the functionality, and this will also allow you to
refactor your code if and when you have the time to do so. You don't
need to create classes and objects to do this.

Fifth, as a strength building exercise, you might attempt to place ALL
your code, every line of it, in user defined functions and execute the
program by the judicious call of these functions. Of course, you would
still need to declare your lexical variables, but you set them by
function calls.

CC

 

[Joost Diepenmaat]

&batch_a;
&batch_b;
&batch_c;

Erm, you *do* know what that does, right? In most circumstances you want
to use

batch_a();
batch_b();
batch_c();

instead.

See perlfaq7

 

[Ela]

Fourth, if possible, place your common code (our variables and

functions) in a PM and import the backage. This way, you can separate
the interface and the functionality, and this will also allow you to
refactor your code if and when you have the time to do so. You don't
need to create classes and objects to do this.

Fifth, as a strength building exercise, you might attempt to place ALL
your code, every line of it, in user defined functions and execute the
program by the judicious call of these functions. Of course, you would
still need to declare your lexical variables, but you set them by
function calls.

CC

I don't quite understand the last 2. Maybe I need to refer to some Advanced
Perl Programming? Anyway I guess your first 3 suggestions should be
sufficient to solve the headache-causing clumsy coding problem. Thank you
and all others' help!

 

[ccc31807]

Erm, you *do* know what that does, right?

I know exactly what calling a function using & does. As I understand
it, it uses your programs underscore varibles.

In most circumstances you want
to use

batch_a();
batch_b();
batch_c();

I'm sorry if what I posted was confusing. Yes, you are correct, and I
certainly did not want to mislead anyone. However, I certainly wasn't
misled when I wrote what I did and didn't think that anyone else would
be, either. In any case, you can call a function clearly and
ambiguously using &. And I just prefer to do so. TIMTOWTDI.

CC

 

[ccc31807]

I don't quite understand the last 2.

1. Please note that I may have misled you by suggesting that you call
your functions using &. While you can certainly do that, ommitting the
& and using the () will probably suit you better, at least until you
get a handle on your personal preference and scripting needs.

2. I strongly suggest you understand the last point. What I mean is
that the body of your script contain only comments, variable
declarations, and function calls. Your functionality is abstracted to
your subroutines. This promotes clarity and modularity, and can help
tremendously with debugging. It separates your logic from your
behavior so you can think about your program logic without considering
implementation. When you write your functions, you don't have to think
about your logic.

I frequently find myself writing the logic in pseudo code as comments,
writing subroutines that implement peices of the logic, and adding the
call to the main script using the pseudo comments as documentation.
This isn't a good way to write big applications, but it is a good way
to develop small to moderate size scripts.

CC

 

[RedGrittyBrick]

_
RedGrittyBrick wrote:
::

::
:: Edager W. Dijkstra. Go To Statement Considered Harmful. 1968.
::
:: In nearly 30 years of programming I have completely avoided use of GOTO
:: in a professional context. All I have seen continues to convince me that
:: Dijkstra was right on this matter. YMMV.

Dijkstra also wrote:

"Please don't fall into the trap of believing that I am terribly
dogmatic about [the goto statement]. I have the uncomfortable
feeling that others are making a religion out of it, as if the
conceptual problems of programming could be solved by a simple
trick, by a simple form of coding disciple!"

Edsger W. Dijkstra, personal communication, 1973.
Quoted by Donald E. Knuth, "Structured Programming with the go to
Statement", 1974.

What Knuth and Dijkstra are saying is that it's not so much the use of
"goto" that should be avoided, the real horror is unstructured programming.
But most people only know the title of Dijkstra's paper, and not the
content [1], let alone know what others have to say about it.

Guilty as charged. Perhaps I am unlucky in that almost all the examples
of GOTO I have had to wrestle with, have been in examples of "spaghetti
programming". It is certainly possible to write spaghetti code without
using GOTO, but use of GOTO seems to encourage this trend in many
programmers.

I'm not afraid to confess I do use goto. I prefer:

again:
my $result = do_something;
goto again if it_failed;

over

my $result;
my $success = 0;
until ($success) {
$result = do_something;
$success = 1 unless it_failed;
}

OK. I don't prefer the former, but I see your point.

However, in this case do_something is three hundred lines, I hope you'd
consider putting it into a subroutine do_something() - or perhaps
re-factoring into several subroutines. The OP is looking for an
"alternative strategy" that doesn't involve subroutines. The OP appears
to be hoping that using GOTO will avoid the perceived difficulty of
reorganising code into well-formed subroutines or objects and methods.

I infer the OP prefers

...
<stuff>
...
<perhaps hundreds of lines of intervening code>
...
<similar stuff>
...

to be transformed into something like

...
my second_time = 0;
stuff:
<stuff probably containing many IFs / GOTOs for dissimilar bits>
if (second_time) goto continuation
...
<perhaps hundreds of lines of intervening code>
...
second_time = 1;
goto stuff:
continuation:
...

instead of

...
stuff("foo", ...);
...
<perhaps hundreds of lines of intervening code>
...
stuff("bar", ...);
...

sub stuff {
<various stuff>
}

I still feel that re-factoring unwieldy and repetitive code into
subroutines (or objects and methods) is a more useful skill to learn
than how to sprinkle GOTO statements into that unwieldy code.

 

[John Bokma]

In nearly 30 years of programming I have completely avoided use of
GOTO in a professional context. All I have seen continues to convince
me that Dijkstra was right on this matter. YMMV.

So you avoid last, next, etc. and returns in the middle of subs as well?
To me those are all forms of GOTO.

 

[xhoster]

So you avoid last, next, etc. and returns in the middle of subs as well?
To me those are all forms of GOTO.

Where do you draw the line? All looping and branching are also forms of
GOTO, if that is the way one wants to look at it.

Xho

--
-------------------- http://NewsReader.Com/ --------------------
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate
this fact.

 

[John Bokma]

Where do you draw the line?

That's a good summary.

 

[ccc31807]

Where do you draw the line? All looping and branching are also forms of
GOTO, if that is the way one wants to look at it.

If you want to ~look~ at it, you can draw flow charts and make sure
that each structure, sequential, looping, or selection, has only one
entry point and one exit point. Alternatively, you can construct a
control flow graph and make sure each structure has one start and one
end node.

All the next, last, redo, and return statements do is exit a loop
according to the value of a variable. True, they are forms of a goto
instruction, but I think we have the idea that goto implies goint to
any arbitrary point in a program while the next, etc., just break out
of the inmost loop (except for labeled loops, but even then they don't
transfer control arbitrarily.)

I think that the point is writing clear, concise, and even beautiful
code, avoiding excessive subjectivism and ideosyncratic expressions,
and following consistent stylistic conventions. This will never go out
of style.

I expect that in about twenty years, someone will write an article
that Java will be considered harmful. This won't be the fault of Java,
but of the habit of forcing all code to conform to OO conventions.
This won't be Java's fault, but the fault of all those who conform
excessively to the One True Way, and discover much later that it's
just another way.

CC

 

[Ben Morrow]

Who the hell uses goto today???

Depends on the language. While I have never used goto in Perl, it is
common in C to see a function structured like the one below: see, for
instance, much of the perl source. This is a consequence of having to
de-allocate everything manually in C. Of course, one has to be careful:
it is all to easy for such a function to degenerate into an
incomprehensible mess.

Ben

int
do_something(...)
{
Foo *foo;
Bar *bar;
int rv;

foo = allocate_Foo();
if (!foo) {
rv = ENOFOO;
goto out;
}

bar = allocate_Bar();
if (!bar) {
rv = ENOBAR;
goto out_foo;
}

rv = try_something(foo, bar);
if (rv < 0)
goto out_bar;

do_something_else();

out_bar:
free_Bar(bar);

out_foo:
free_Foo(foo);

out:
return rv;
}

 

[szr]

Depends on the language. While I have never used goto in Perl

Just a sort of a semi-tangent question, isn't using next and last with
labels (to, say, break out of a parent loop from an inner loop in a set
of nested loops) implicitly using `goto` ? If you think about it, the
syntax is the same:

Loop1: while (...) {
Loop2: until (...) {
if (some_condition) { last Loop1; }
elsif (some_other_cond_a) { next Loop1; }
elsif (some_other_cond_b) { redo Loop1; }
}
}

I'm sure most everyone has used this sort of construct at some point or
another to break from an outer loop from within an inner one.

Correct me if I'm wrong, but are not last/next/redo with labels basic
`goto` statements that enact a last, next, or redo on the loop
associated with the label?

Does this constitute a necessary evil, or more rather, demonstrates an
instance where `goto` (in some form) has a place?

 

[Joost Diepenmaat]

Correct me if I'm wrong, but are not last/next/redo with labels basic
`goto` statements that enact a last, next, or redo on the loop
associated with the label?

Not exactly. Since you can only use them to go up the call stack another
way to think of them is as special exceptions with built-in handlers.

For real fun, see common lisp's condition/restart system.

In any case everything that alters the program flow can ultimately be
seen as a (conditional) goto. Even such basic operators as && and if()

Does this constitute a necessary evil, or more rather, demonstrates an
instance where `goto` (in some form) has a place?

It just demonstrates a useful pattern for program flow, which IMO is
better handled by having explicit operators with predefined behaviour
than by goto. The main problem with goto for me is that it's so
unrestricted it *can* lead to code that's very hard to understand and
bugs that are hard to analyze, not that it's inherently evil or anything
like that.

Perl has enough useful flow control operations and other features that
it's rare to run into situations where goto() is the cleanest / clearest
/ obvious choice.

 

[szr]

Not exactly. Since you can only use them to go up the call stack
another way to think of them is as special exceptions with built-in
handlers.

Yeah but they seem to use the goto-label system in a way still (when
using loops with labels.. see two paragraphs down.)

For real fun, see common lisp's condition/restart system.

Will do.

In any case everything that alters the program flow can ultimately be
seen as a (conditional) goto. Even such basic operators as && and if()

True. Though my point was a somewhat about the fact that they use the
same labels that goto uses for it's jump destination so next/last/redo +
label seemed like a goto with an additional instruction that acts on the
loop belonging to that label's context.

It just demonstrates a useful pattern for program flow, which IMO is
better handled by having explicit operators with predefined behaviour
than by goto.
Agreed.

The main problem with goto for me is that it's so unrestricted it
*can*
lead to code that's very hard to understand and bugs that are hard to
analyze, not that it's inherently evil or anything like that.

True. I've just always seen it something that does have it's place like
anything else, and must be handled with much are when used (though I do
not really explicitly use goto in my programs, just never found the
need, to be perfecly honest

Perl has enough useful flow control operations and other features that
it's rare to run into situations where goto() is the cleanest /
clearest / obvious choice.

And this is probably /why/ I've never really needed to use goto in Perl
(though I also never needed to use it and c/c++ and the likes.)

 

[John Bokma]

One could put forward the argument: Those are syntactic sugar to cover
those cases where the "no goto" rule deserves to be excepted.

Instead of rules I prefer to ask myself the following questions:

1) can I understand my own code next month without effort (can I after
one year)
2) can a peer do the same?
3) can someone who has some programming experience read the code without
wasting too much time.

I always get a good feeling if a customer contacts me that a
modification has been made to my code and that it was easy to do so.

Making up rules and then consider syntactic sugar exceptions sounds odd to
me.