Spoiled by Perl

[afoken]

Some context: My initial problem of making sense of a pile of Excel VBA junk (see [OT] Finding similar program code) is largely solved, and I have started re-implementing the calculations in C#. C# is obviously not my first language, neither the second or third one. It's more somewhere around number 20. I've lost count. Ranting about a certain Hunchback-Gollum-Salvatore C derivate caused a misunderstanding and triggered a reply by jdporter:

wait - do you mean C#? Because imho C# is a great language.

Jdporter is not alone, I've already discussed a lot with my coworker, who really likes C#.

The re-implementation is my first C# project, after a tiny "hello world" command line program just to test the VS2019 installation. I have already helped to debug some C# software, but rarely on the code level, more on the algorithm level, and of course some rubber duck debugging like "a < 10 looks wrong if you want to check that a is larger than 10".

C# is no huge surprise. It is just another language derived from C and some other languages, with OOP, and with its own set of quirks, strengths, and weaknesses. C# is designed for OOP, and supports OOP much better than some other langauges. OOP is not an afterthought, it is clearly part of the original design. Over the years, some syntatic sugar was added to the existing syntax, allowing to write quite short, but still readable code.

But I miss some really trivial things I'm used to in Perl:

The <=> and cmp operators together with ||

In C#, many classes (e.g. List<T>) implement a Sort() method that expects the C# equivalent of a compare function. You end writing something like this for a simple compare:

someList.Sort((a, b) => a.Foo.CompareTo(b.Foo));

Well, CompareTo() is not pretty, but still acceptable. I would prefer the (invalid) C# example below, perhaps using operator overloading to call CompareTo() behind the scenes. Unfortunately, C# has no <=> operator that could be overloaded.

someList.Sort((a, b) => a.Foo <=> b.Foo);

Compare to Find:

var first = someList.Find(e => e.Foo >= bar);

Compare to sort in Perl:

@someList = sort { $a->Foo <=> $b->Foo } @someList;

Things get ugly if you want to sort by more than one property:

@someList = sort { $a->Foo <=> $b->Foo || $a->Bar cmp $b->Bar } @someL
+ist;

In C#, I ended up with that:

someList.Sort(
    (a, b) =>
    {
        int rv = a.Foo.CompareTo(b.Foo);
        if (rv == 0)
        {
            rv = a.Bar.CompareTo(b.Bar);
        }
        return rv;
    }
);

You can't use a.Foo.CompareTo(b.Foo) || a.Bar.CompareTo(b.Bar) in C#, because || is strictly boolean.

(Yes, I know that LINQ exists and has its own syntax plus extension methods to work with data in a way similar to SQL.)

redo

Restart a loop block, bypassing the conditional. See redo. It simply does not exist in C#. You have to use goto, which is at least ugly; or use an inner loop, which opens another can of worms (see below).

I rarely need redo, but some of the calculations need a redo. In VBA, it was simply implemented via goto jumping backwards over some hundred lines of code. My C# code also uses goto, for the lack of a better loop control.

next LABEL, last LABEL, redo LABEL

Again, those do not exist in C#. You can not get out of an inner loop AND start the next iteration of an outer loop, exit the outer loop, or restart an outer loop. Again, you have to use goto. If you need all three variants in a single set of two nested loops, you need three labels: One at the start of the outer loop to simulate redo, one at the end of the outer loop to simulate next, and one after the outer loop to simulate last. You could also break out of the inner loop and use a bunch of helper variables, which don't make the code more readable.

Again, I rarely need to control an outer loop from an inner loop, but some of the calculations need that, too. Again, VBA used goto and jumped wildly around, and my C# code also uses goto where I could not avoid it.

Alexander

[eyepopslikeamosquito]

An appreciation of C# is helped by remembering the turbulent political era during which it was born. At least, during the 1990s, I felt depressed and disgusted by the outrageously expensive legal battles erupting between big software corporations, such as Sun Microsystems, Microsoft, Apple and Oracle. This constant childish corporate bickering drove me away from the Java and C# desert ... into a refreshing oasis of more open technologies, such as Linux, C, C++, Python, Ruby and ... Perl! :)

In the late 1990s for example:

Sun filed this suit for copyright infringement, claiming that Microsoft had exceeded the scope of its license by creating an enhanced version of Java that was fully operable only on Microsoft's operating system...

These relentless ongoing expensive Java legal battles with Sun drove Microsoft to pour millions into developing its own Java-killer, namely the C# Programming Language and .NET Framework (see also .NET and Mono).

Things got much worse for Java after Oracle Corporation acquired Sun Microsystems in 2010. Though Oracle didn't set the bar very high as the new custodian of Java, Microsoft did a much better job as custodian of the C# programming language (even embracing open source).

At least that's my take on why C# has steadily, relentlessly become a better technology than Java over the years. Oh, and hiring and keeping an accomplished language designer certainly helped -- by contrast, Gosling quit in disgust after Oracle acquired Sun.

I get a bit annoyed when people compare C# to C++ rather than Java. Of course, Microsoft themselves did that (presumably for both marketing and legal reasons) but to my eye C# is much closer to Java than C++.

At least that's my take on the history. Corrections welcome.

At work, we used Java in the early days, but gradually migrated to C# over the years. I don't think there's much Java left nowadays - though Groovy remains popular for DevOps work, notably with Jenkins.

Though I've mostly been successful at avoiding C#, it seems a pretty nice language to me - certainly it has plenty of fans at work (as noted at Five Ways to Reverse a String of Words (C#, Perl 5, Perl 6, Ruby, Haskell)).

I boldly predict C# will eventually overtake Java on tiobe. I remember the heady days when Java was top of this list ... hmmm, perhaps all the corporate bickering between Java and C# over the years helped propel Python to the top of tiobe today! :)

[eyepopslikeamosquito]

The Java/Microsoft fight of twenty years ago has left us today with what looks to me like an unfortunate legacy: extra work for language implementors ... or is it an extra challenge with the potential to improve the way languages are implemented?

AFAIK, Perl 5 has never seriously run on JVM or CLR. Though Raku has a JVM implementation, the CLR implementation (Niecza) seems to be abandoned. While Python 2 runs on both (Jython, IronPython), the Python 3 port to both these VMs appears to be unfinished - 13 years after the C version of Python 3 was released in 2008!

Note that I don't have any first hand experience here ... opinions/corrections welcome. :)

References

• Java Virtual Machine (wikipedia)
• List of JVM languages (wikipedia)
• Common Language Runtime (wikipedia)
• List of CLI languages (wikipedia)
• Common Language Infrastructure (wikipedia)
• Standard Libraries (CLI) (wikipedia)
• Raku Compilers
• CPython (wikipedia)
• Jython (wikipedia)
• IronPython (wikipedia)
• History of Python (wikipedia)

[Jenda]

ActiveState used to have tools that allowed you to integrate Perl and CLR based stull. The Perl itself was running using its own engine, but you could call one from the other. It makes more sense to do it that way anyway. CLR was not designed for dynamic languages.

[LanX]

Thanks for the informative overview ...

> claiming that Microsoft had exceeded the scope of its license by creating an enhanced version of Java that was fully operable only on Microsoft's operating system

So same strategy like with JScript / IE to kill/steal the main project. Just that the story here led to ECMAScript.

[karlgoethebier]

And sometimes good things simply vanish. I sadly remember this stuff which was my favorite waste of time as well as my cashcow for some years.

[jdporter]

Well, yes. It may be a great language, but that's in comparison to C++, Java, and other 3rd-gen languages - of which it is one itself. But it clearly ain't no Perl!

I appreciate your very thoughtful writeup. Thank you!

[holli]

I know the feeling. It's how people who learned Raku look at Perl: A good idea with sugar added later, but still leaves a lot to be desired.

In Raku, if you were to sort on the result of a method you would just say

@somelist.sort( *.Foo );

This will even use a schwartzian transform if the optimizer decides it is worth it.

[LanX]

coincidentally I procrastinated this Sunday implementing syntactic sugar for exactly this in Perl5.

There are some edge-cases tho, which keep we wondering ...

In Raku:

• how do you decide its a cmp or <=> ?
• what if you want to apply another ordering? Like Czech sorting?
• how do you sort by multiple criteria? Method chaining?

> This will even use a schwartzian transform if the optimizer decides it is worth it.

an optimizer detecting a good case for a Guttman-Rosler Transform transform would be more impressive tho. ;-)

[holli]

how do you decide its a cmp or <=> ?

The dispatcher does that by finding the appropriate multi-sub that implements the actual comparison.

# cmp. Generic, "smart" three-way comparator.

multi sub infix:<cmp>(Any,       Any)
multi sub infix:<cmp>(Real:D,    Real:D)
multi sub infix:<cmp>(Str:D,     Str:D)
multi sub infix:<cmp>(Version:D, Version:D)

# Compares strings with string semantics, numbers with number semantic
+s, Pair objects first by key and then by value etc.
# if $a eqv $b, then $a cmp $b always returns Order::Same. Otherwise O
+rder::Less or Order::More.

what if you want to apply another ordering? Like Czech sorting?

Rakus Str class is fully Unicode aware and by default sorts characters on the numeric value of the codepoint.

dd 'a ä b'.comb.sort;
(" ", " ", "a", "b", "ä").Seq;

If you wanted to implement your own behaviour you could just add another multi

subset CzechString of String;
multi sub infix:<cmp>(CzechString $a, CzechString $b) {
  # sort logic here
}

Or you can simply pass your comparator to sort

sub czech-sort(String $a, String $b) {
  # sort logic here
}

@czech-words.sort( &czech-sort );

how do you sort by multiple criteria? Method chaining?

You can adress that in the comparator like in Perl, or make the comparator return a list of things.

dd ([1,'z'], [2,'a'], ['1', 'a']).sort( *.[0,1] );
(["1", "a"], [1, "z"], [2, "a"]).Seq

dd ([1,'z'], [2,'a'], ['1', 'a']).sort( { $^a[0] <=> $^b[0] || $^a[1] 
+cmp $^b[1] });

[Jenda]

var sorted = somelist.OrderBy(x => x.Foo());

But yeah, the woke butterfly that killed Perl wins on character count. The twenty years of waiting "paid off".

[holli]

The twenty years of waiting "paid off".

Yes. Indeed it has. Check out this Raku code of mine which is a solution for this years day 14. Any experienced programmer should be able to make sense of, and mostly understand it after spending 2 hours with the docs. I doubt the same is true for any perl implementation of similar brevity.

my ( \o, \r ) =
    .head.comb.List,
    .tail.comb( /\w/ ).map( -> $a, $b, $c { "$a$b" => [ "$a$c", "$c$b"
+ ] } ).Hash
        with cache $*IN.slurp.split: "\n\n";

sub day14( \n )
{
    my %d = o.rotor( 2 => -1 )».join.Bag;
       %d = [(+)] %d.map({ r.{.key} »=>» .value }) for ^n;

    .Bag.values.minmax.elems - 1
        with ( o.tail => 1, |%d ).map: {
            .key.comb[0] => .value
        };
}

say day14 10|40;

[ikegami]

Nice. Complete version:

var sorted = someList.OrderBy( _ => _.Foo ).ThenBy( _ => _.Bar );

I think the above is produced by the following:

var sorted =
   from item in someList
   orderby item.Foo, item.Bar
   select item;

[ikegami]

If it was a stable sort, you could use

someList.Sort( (a, b) => a.Bar.CompareTo(b.Bar) );
someList.Sort( (a, b) => a.Foo.CompareTo(b.Foo) );

But it's not. Another nice thing about Perl. (use sort qw( stable );, which is currently the default.) JavaScript also has a stable sort (at least since version 10 or EcmaScript 2019).

That said, I like C#. It's quite good at doing what Perl does well.