This is the mail archive of the kawa@sourceware.org mailing list for the Kawa project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: kawa goals


On 10/14/2010 03:24 AM, Helmut Eller wrote:
* Per Bothner [2010-10-14 07:27] writes:

Actually, I was thinking about asking people for requests.
...

Out of those: iteration.

I agree this is important. I've been implementing the "Shootout" programs in Kawa, and having to use do or named-let is OK, but rather verbose and klunky.

It also is holding up my writing of the Kawa tutorial - I'd like to
use something more "list-comprehension-like".

At least something like Common Lisp's LOOP or
ITERATE.  Ideally something like SERIES[1] (with loop fusion for nested
series expressions) combined with a terse comprehension syntax for the
common case (i.e. map+filter+reduce).

I certainly much prefer series over loop/iterate. I have a rather radical design that I'll talk more about after I get 1.11 out the door.

In general, make Kawa more "friendly", e.g. add doc-strings, make it
easy to find the source file of functions, add an object inspector etc.

I think many of these are IDE issues, though I realize traditionally IDE's in Lisp have depended on eval and repl. Thus I don't think old-fashioned docstrings (in the sense of a run-time property) is really the right thing. Better an annotation (or a separate resource file), so the strings can be found either at compile time *or* at run-time.

[OTOH, to become "rich and famous" you should probably do something
related to parallelism and/or cluster/cloud computing :-)]

Part of the plan/goal for a "list-comprehension" is to optionally implement it like a "ParallelArray". More later ... -- --Per Bothner per@bothner.com http://per.bothner.com/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]