From: Roedy Green on
On Sat, 03 Oct 2009 11:12:10 -0700, Roedy Green
<see_website(a)mindprod.com.invalid> wrote, quoted or indirectly quoted
someone who said :

>
>I first developed a lust for this light-weight type

In scientific/engineering, you can't afford the overhead of a
full-blown Java type. You want something that prevents mismatch
errors, but that does everything at compile time, much like generics.
For that, you are willing to give up a lot e.g. aggregates,
inheritance. You want the same code generated that if you had used
primitives instead of types.

Perhaps another way to look at this is how could you implement a final
Java type containing but one primitive in a highly efficient way?
Perhaps it could be autounboxed/boxed.


--
Roedy Green Canadian Mind Products
http://mindprod.com

When you can�t find a bug, you are probably looking in the wrong place. When you can�t find your glasses, you don�t keep scanning the same spot because you are convinced that is where you left them.
~ Roedy
From: Kenneth P. Turvey on
On Sat, 03 Oct 2009 11:24:39 -0700, Roedy Green wrote:

> Perhaps another way to look at this is how could you implement a final
> Java type containing but one primitive in a highly efficient way?
> Perhaps it could be autounboxed/boxed.

This I understand...

Have you ever looked at the Common Lisp type system? You could declare a
variable to be an integer between 7 and 23, but not 12. The compiler was
expected to use this information for the purposes of optimization.

I could see the use of something like this in Java, but I am concerned
about adding another complex syntax to the language.

--
Kenneth P. Turvey <evoturvey(a)gmail.com>
From: Roedy Green on
On 03 Oct 2009 19:04:24 GMT, "Kenneth P. Turvey" <evoturvey(a)gmail.com>
wrote, quoted or indirectly quoted someone who said :

>I could see the use of something like this in Java, but I am concerned
>about adding another complex syntax to the language.

Java is obviously not the final solution. How fast do we evolve, and
do we do it by modifying Java or starting over.

I think the next big round of progress will come when we abandon the
notion that computer languages are conceptualised as linear ASCII text
files.

I think it is time to think of programs as directed graphs, displayed
with filters to show you only what is of immediate interest.

The problems we have are being overwhelmed by verbose detail and
inscrutably complex syntax (i.e. generics). We have to think about
hiding what is not relevant for now, as a fundamental part of
"language" design.

how can we add features for one group of programmers without baffling
those who don't need it?

Just as class libraries have forced programmers to give up doing
HashMaps "their way", this same standardisation process needs to
continue, so that AI/class libraries take over more and more of the
fussy details, in particular screen layouts, data entry and data
validation.

See http://mindprod.com/project/scid.html

Another metaphor is the Google Earth map where you can zoom in to
various levels of detail, or zoom out to see the whole thing.
--
Roedy Green Canadian Mind Products
http://mindprod.com

When you can�t find a bug, you are probably looking in the wrong place. When you can�t find your glasses, you don�t keep scanning the same spot because you are convinced that is where you left them.
~ Roedy
From: Lew on
Roedy Green wrote:
> The problems we have are being overwhelmed by verbose detail and
> inscrutably complex syntax (i.e. generics). We have to think about

Generics is not inscrutable. With a little study, as little as reading the
short free chapter on generics from Bloch's /Effective Java/, it's utterly
scrutable.

Not a good example for what is otherwise a decent point.

--
Lew
From: Dave Searles on
Roedy Green wrote:
> On 03 Oct 2009 19:04:24 GMT, "Kenneth P. Turvey" <evoturvey(a)gmail.com>
> wrote, quoted or indirectly quoted someone who said :
>
>> I could see the use of something like this in Java, but I am concerned
>> about adding another complex syntax to the language.
>
> Java is obviously not the final solution. How fast do we evolve, and
> do we do it by modifying Java or starting over.
>
> I think the next big round of progress will come when we abandon the
> notion that computer languages are conceptualised as linear ASCII text
> files.

To some extent, that's already happened. It's called "Lisp". :)
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
Prev: Putting passwords in a properties file?
Next: Interview