You are reminding me of a variety of issues. I remember, when I was first learning APL, that I was frustrated with how inefficient APL was at doing what I might call streaming IO, or what I might call wart-for-wart emulation of ad-hoc code.
I also remember how kx.com's languages deal with this issue - console output is all serialized (string values are always quoted) and is directly interpretable as input by the language (which makes setting up distributed systems a breeze, at the cost of a bit of excessive decoration if someone tries to use the console for formatted I/O - but the language provides other I/O mechanisms). I also am remembering some of the decoration tools and devices I have used in other contexts, to study the data and understand how a system goes together. Anyways, there's always tradeoffs (and I often wish for a 5!: representation which could be used in parenthesis in a string being passed to ". where the current linear representation assumes something different). Still, I liked your presentation, and it seems like a promising (and possibly maddening) avenue of approach. Thanks, -- Raul On Mon, Feb 3, 2014 at 10:28 AM, robert therriault <bobtherria...@mac.com> wrote: > As a bit of a response to the whole question of what we see and what we get, > I put together some html and css that allows a view of J results on JHS that > gives type information implicitly in the display. > > I think the blog post and demo video gives the flavour of the journey. > > http://bobtherriault.wordpress.com/2014/02/02/using-html-and-css-to-view-types-in-the-jhs-platform-of-j/ > > Cheers, bob > > On Jan 20, 2014, at 3:01 AM, Joe Bogner <joebog...@gmail.com> wrote: > >> Makes perfect sense now. The atom vs list distinction wasn't clicking >> earlier. I had become so used to working with arrays that everything >> became an array and I had completely forgotten about scalars. The >> dictionary entry on nouns also covers it well, >> http://www.jsoftware.com/help/dictionary/dicta.htm. >> >> >> On Sun, Jan 19, 2014 at 11:18 PM, Henry Rich <henryhr...@nc.rr.com> wrote: >>> The primitives, namely ;: u;.n u\ u\. u/. produce lists even when there is >>> only one item in the partition. Very regular. >> >> >> I understand better why it would do that now. As it partitions a list, >> it is likely simpler and performs better to create a list for each >> partition instead of determining whether there's only one item in the >> partition. I think of it as a splitting a char[] array into other >> char[] arrays instead of char for single and char[] otherwise. >> >> >>> There's just something special about a single character >>> or a single number: they are atoms. >> >> This and the dictionary entry explains why $ 'a' or $ (<'abc') returns >> blank - since each are atoms. >> >> Thanks again >> ---------------------------------------------------------------------- >> For information about J forums see http://www.jsoftware.com/forums.htm > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
