Hello Eric,
Thanks for this. As I was struggling with the use of stdin/stdout/stderr, I
waiting for this in J. I could not quite follow your statement on "work in
binary rather than text". What do you mean by this? I have seen the echo
interpreting ^D as a 'diamond' on my pc. Is this what you would like to see
differently?
1) I am little confused on the use of CRLF and ^Z in the jconsole. It looks
like the interpreter waits until ^Z to process the input buffer. Here, I
would like to see how the CRLF can be used to terminate the input buffer and
give the control back to the interpreter to decide whether the input is
complete or needs more from the user.
[This is a design point for writing console based interpreters]
2) I would like to see something like K's following features ...
.m.p is the prompt, e.g. ">"
.m.r waits for a response from the keyboard, then processes it,
e.g. .m.r:{\x} echoes what the user types.
[This facility would help in writing console based interpreters]
3) I am also a little confused on how stdout works. It seems to be repeating
the character. So,
stdout '>'
(output - NB. Why twice??)
4) Finally some supporting examples would help. Basically, I am looking for
superimposing an interpreter - written in J - on the top of J's console
interpreter. This would allow me to escape into J native and use my
interpreter seemlessly. This is more on the lines of an "internal Domain
Specific Language" (
http://www.martinfowler.com/bliki/DomainSpecificLanguage.html).
From the perspective of languages design and evolution, I see that support
for creation of internal/embedded DSL would
a) make the use of the general purpose language attractive
b) leverage the meta-programming constructs of the language (nouns, verbs,
adverbs, conjunctions, tacit programming) reflect in a domain specific way
leading to new usages.
My imagination is that with a powerful system like J as the base, allowing
for creation of internal DSLs that can "reuse" the operators of J would be
an interesting exploration. Since the internal DSL is superimposed, the
domains of interpretation are different and hence there should be no
conflict in the specific interpretation of the input. One can then think of
how the computation done in native J inter-operates with the computation in
the internal DSL. Well, I see a lot of potential here....let me know what
you think.
Probably this should be part of the JBeta mailing list. So, I am
cross-posting.
Regards,
Yuva
On 1/28/07, Eric Iverson <[EMAIL PROTECTED]> wrote:
The next release (602 beta) will have windows stdin/stdout/stderr work
in binary rather than text. They should always have been this way and I
consider this change to be a bug fix.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
