light weight types
in [Java Help]
|
Prev: Putting passwords in a properties file?
Next: Interview
From: Roedy Green on 3 Oct 2009 14:24 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 3 Oct 2009 15:04 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 3 Oct 2009 15:54 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 3 Oct 2009 16:12 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 3 Oct 2009 17:14
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". :) |