Björn wrote:
> I agree that it is important to document well
> how to communicate with files etc.
I agree that hints and pointers to library functions should be interspersed in
our Vocabulary pages, e.g.
== ;. (cut) ==
Cut does such and such...
== uses ==
Cut is often used to parse text:
<;._1 '/please/parse/me!'
(more examples).
Though of course, like many primitives, ;. is type-agnostic and can be
used to parse any time of array, not just text.
== notes ==
1. Many standard J utilities take advantage of these features, and
provide convenient, pre-packaged parsing functions for many
common cases. In fact, most J applications rely on these
standard
utilities rather than building a parser from scratch with cut.
For
example, we could've written the above as:
require'strings'
'/' cut '/please/parse/me!'
(more examples)
See the libraries strings <link>, csv <link>, etc <link>. You
may
load these libraries in your profile (link), so the utilities
are always
available.
2. Advanced parsing can be achieved with the dyad ;: (FSM) <link>
(and its monadic form <link> has a pre-packaged parse for J code).
We could do similar things for the documentation of 1!: (e.g. "Most J
applications rely on the standard definitions in the files and dir libraries
<links>, rather using the 1!: foreigns directly).
But this is about the greatest level of detail appropriate to a reference work.
For more in depth treatments of the libraries, or "how tos" for various common
programming problems (e.g. reading files, parsing text, building GUIs), we
could refer to existing tutorials and labs, or write supplementary,
non-reference documents where the existing material is thin.
But let's keep the Dictionary to its purpose: definitions.
-Dan
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
