I guess I'm not really asking for the Dic definition of $: to be changed, much less its implementation. I'm just wondering what NuVoc should say about it.
NuVoc can afford to be informal, quirky and non-doctrinal. It can jettison attempts at a watertight phraseology in favour of one or two examples. It can adopt an arms-length attitude to the product and its formal documentation -- if that's helpful to novices. In fact I think novices should have the last word here. I notice Randy (=ArrayMac) has quite independently posted his discovery of $: and how to use it in an explicit definition. Which reassures me other newbies scratch their heads over $: and it's not just me doing my usual trick of seeing problems where nobody else does. It's not unreasonable for the novice to turn to the system locales like z, jijs and j, for examples of how the experts use J. Now, in locale 'j' I see: printfiles=: 3 : 0 '' printfiles y : ... which thereby offers itself as "the way it's done" -- because when I try to improve on it by: printfiles=: 3 : 0 '' $: y : ... ...that simply does not work. That's something that NuVoc should address, if only to bow to the fact and reassure the novice s/he isn't just having a bad morning. As it happens, there's one single instance of $: being used in the whole j-locale, and it shows the "real" way to do it (...everyone agree?). Which happens to be the way Randy has just posted: loadp_j_=: ''&$: : (4 : 0) ... The message is clear -- and it cuts the Gordian knot of trying to delineate what $: is referring-to when used inside an ambivalent explicit definition... Don't do it. Use a dyadic definition. I'm thinking aloud here. Not trying to formulate doctrine but just anticipate what NuVoc ought to say. Perhaps what I ought to have done is go ahead and draft the $: entry for NuVoc and let people poke it about. It's easier to see what's wrong with something once it exists, than conjecture in advance what it should look like. On Mon, May 9, 2011 at 5:35 PM, Raul Miller <[email protected]> wrote: > On Mon, May 9, 2011 at 12:19 PM, William Tanksley, Jr > <[email protected]> wrote: >> Raul Miller <[email protected]> wrote: >>> A qualified rephrasing of the dictionary definition, which I believe >>> would be accurate would be: >>> $: denotes the longest verb that contains it at the time when the >>> verb is first used. >>> Of course, this delegates the issue to the concept of "first used" but >>> that can be broken down fairly simply: >> >> Does "first used" mean "first assignment or display"? > > Both of those seem like uses. Wouldn't you think so? > > But the definition of $: does not really matter for display, and > "first assignment" was already covered by my breakdown of what first > use would mean: > >>> 1. When the verb is presented with a noun argument or arguments, or >>> 2. When the verb is given a name. > > -- > Raul > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
