You know - this argument gets rolled out every time operator
overloading comes up, but I posit this simple counter argument:

In Java, you have to write .add, .multiply etc. methods (e.g.
BigDecimal) because you can't overload the operators. This makes the
code a heck of a lot messier but doesn't bring any extra safety -
seriously what's the difference between overloading + with something
inappropriate and overloading .plus or .add with something that
removes or subtracts, especially because you can't use the operators
on your own classes anyway.

A badly named method is a liability whether it's symbolic or
alphanumeric in nature. It all speaks to poor software engineering
practices. Anything else is a smokescreen.

This is why Scala does not have "operator overloading", it simply
allows symbolic method names. 5 + 6 and 5 plus 6 can be equivalent,
both mean something to the developer using the object, both could be
abused.

Dick

On Aug 7, 3:32 pm, Reinier Zwitserloot <[email protected]> wrote:
> How's that a case for simplicity for scala?
>
> In java, "5 + 2" means just what you think it means, intuitively. If
> you want to know more still you'll have delve into the extensive JLS
> which confirms your suspicions.
>
> In scala you need to delve into the libraries, and you really have no
> idea what it could possibly do - every object can field its own
> definition of '+'. This isn't simple anymore; the drive to libraryize
> all complexity means that most seemingly atomic library operations are
> in fact not the lowest layer, but they build on a lower layer still,
> and in languages like scala, that lowest of layers is not all that
> natural.
>
> I continue to assert that claiming scala is simpler because the JLS is
> longer than the scala equivalent is the stupidest thing I've ever
> heard.
>
> On Aug 6, 10:59 am, Kevin Wright <[email protected]> wrote:
>
>
>
> > The idea that we can establish some sort of formal complexity measurement
> > for documentation is... interesting.
> > Although I think there's more joy to be had in measuring EBNF, or the size
> > of some other parser grammar considered complete for the language.
>
> > I'd also like to briefly explore one of the differences between the two
> > language specs... Consider the `+` operator.
>
> > In the JLS, this is covered in chapter 15, "expressions"
> > 15.18 - additive operators
> > 15.18.1 - string concatenation
> > 15.18.1.1 - string conversion
> > 15.18.1.2 - optimization of string concatenation
> > 15.18.1.3 - examples of string concatenation
> > 15.18.2 - additive operators for numeric types
> > Spanning pages 496-501
>
> > It's a bit different in Scala, which doesn't have operators as quite the
> > same way.  They're just methods in infix/operator notation.
>
> > Everything in Scala is also an object (no primitives).  Int, Float, String
> > etc. are still optimised in the compiler, and will often use primitives in
> > the generated bytecode, but within Scala code they are objects.  So `+`
> > becomes a method on the String/Int/Float/etc. object.  The Scala spec
> > doesn't list API methods any more than the JLS does.
>
> > What the spec *does* have is section 6.12.3, outlining how methods used in
> > the infix position have a precedence determined by the first character of
> > the operator name.  One beauty of thie approach is that it effectively gives
> > you operator overloading, allowing things like 
> > this:http://www.scala-lang.org/api/current/scala/math/BigDecimal.html
>
> > This is sufficient to be able to duplicate Java's behaviour, by building up
> > to it on the basis of a broader (and simpler) concept.
>
> > So "x" + 2  May look the same as Java, but it's actually achieved via the
> > `+` method on a String instance, and not a dedicated special-case operator
> > with 3 sections in the language spec.
>
> > On 6 August 2010 00:38, JodaStephen <[email protected]> wrote:
>
> > > Kevin Wright is fond of repeating:
> > > Java (3rd Edition): 649 pages, 7932 KB
> > > Scala (current in trunk): 191 pages, 1312 KB
> > > therefore Scala is less complex.
>
> > > But has anyone actually analysed the specs in detail?
>
> > > In code coverage terms, how many distinct "code paths" are there in
> > > each spec. That is surely a far better measure than number of pages.
>
> > > Volunteers for counting?!!
>
> > > Stephen
>
> > > --
> > > You received this message because you are subscribed to the Google Groups
> > > "The Java Posse" group.
> > > To post to this group, send email to [email protected].
> > > To unsubscribe from this group, send email to
> > > [email protected]<javaposse%2bunsubscr...@googlegroups
> > >  .com>
> > > .
> > > For more options, visit this group at
> > >http://groups.google.com/group/javaposse?hl=en.
>
> > --
> > Kevin Wright
>
> > mail/google talk: [email protected]
> > wave: [email protected]
> > skype: kev.lee.wright
> > twitter: @thecoda

-- 
You received this message because you are subscribed to the Google Groups "The 
Java Posse" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/javaposse?hl=en.

Reply via email to