Hi Don, Firstly just want to say that I'm impressed with the thought and effort that you've obviously put in to this. I think this issue appears a number of times in the forum archive, but as far as I'm aware you are the first to do more than just talk about it! I apologise in advance if my comments below seem to be rather negative but I hope, like me, you prefer negative feedback to none!
Re: ideas for symbols & notation. I find the use of boxes around symbols (eg. for i and F quite ugly and they tend to break my perception of the sentence as a whole and make me want to focus on that particular symbol as being self-contained rather than as interacting with the surrounding symbols. I can see the recognition benefits of using the square root symbol, but I don't like the proposed symbol for square and the disadvantage is that the consistency of the symbols used in J is lost (*: %: +: -:) If I remember that *: is square then it follows that %: is square root - same goes for double & halve. You suggest that a verb followed by an adverb should be underlined to show that a new verb is formed- I'm not sure that that concept scales very well when multi-symbol verbs or adverbs/conjunctions are involved. Should the following tacit verb followed by an adverb all be underlined? If so should +/ and the other underline be nested? (+/ % #)\ I think that the use of display methods such as: http://www.jsoftware.com/jwiki/OlegKobchenko/Map%20Display may be of more use here. In my earlier post, I suggested that it would be good to separate the ideas regarding symbols from your ideas on an adverb alternative to fork. Part of my reason for this is that my impression from your paper (rightly or wrongly) is that you have not really understood the tacit form and the relationship between the tacit form, the explicit form and sentences entered in the session. Entered in the session this is an explicit sentence that contains a tacit bit: <. 0.5 + (+/ % #) C Assigning that same sentence to a name says that the whole sentence should be interpreted as tacit. It would be the same as entering this into the session: (<. 0.5 + (+/ % #)) C Which is the same as: (<. 0.5 + +/ % #) C These are both evaluated as two forks and a hook. To get the desired answer you could enter the following sequence of 3 forks in the session. ([: <. 0.5 + +/ % #) C Note also that the tacit J code you give for calculating standard deviation does not give the correct answer. In terms of portability and security I'd prefer .pdf over .doc any day! I suspect Raul was suggesting the content would be better in the wiki. Ric > -----Original Message----- > From: [email protected] [mailto:general- > [email protected]] On Behalf Of Don Watson > Sent: Friday, 13 March 2009 10:03 > To: General forum > Subject: [Jgeneral] J with APL > > Everyone seems to be talking about different issues than the one I was > addressing on this thread. I would greatly appreciate help from anyone > who > is willing to read the paper and commented on it. > > In my proposal, nothing changes for J programmers. Keying is > done in identical ASCII, storage is done in identical ASCII, > transmission > between systems is done in identical ASCII, mailing is done in > identical > ASCII and every user of J continues using ASCII in exactly the same way > without knowing that anything has happened. The change is transparent > and > causes none of the difficulties that are being discussed. > > The change is solely intended to assist those who are learning > Mathematics > and very casual users of J. It allows them, in effect, to say to the > system: > "When you come to write my J language statements to the screen or to > the > printer, I want you to look up every two character primitive in which > the > second character is a "." or a ":", in a table and use the special > symbol > you find there instead - that's all. This is easy to do. > > There is a second component to the proposal to do with adding a second > kind > of fork. Again, it is an addition that no J programmer need know > anything about or ever use. It is proposed that for this fork any J > expression is permitted on either branch of the fork and to the left of > the > fork - not just a verb. This makes it lot easier to parallel > mathematical > formulae in tacit programming for those learning to apply mathematical > formulae. > > The handwritten script I posted was a mess and the pdf form was not > popular. > I have created the primitive symbols in Paint and added them into a > word > document. This revised paper has now replaced the previous one at: > > http://bcompanion.com/Compromisedoc.doc > > Thanks > > Don Watson > > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
