Raul wrote: >In the case of the recent example (the "verb phrase" thing), >the dictionary provided a clear and unambiguous
Sure, I agree. Perhaps I should clarify mi position. Although I have only been programming exclusively in J for many years, often several days and occasionally even weeks pass without doing any coding. As a result, I still consult the dictionary regularly. I really appreciate its examples and clear and concise descriptions but, from my vantage point, its conciseness goes too far sometimes. >In the case of fork, in other languages when the specification >leaves something unaddressed, the handling of that issue >is up to the implementation. The problem is that the implementation might convey the wrong message; for example, it allowed for a very long time to pass an adverb or a conjunction as an argument to another adverb (and in a meaningful and useful way in my opinion, although perhaps dangerously). Henry wrote: >What I want to see is something like the Oxford Annotated Bible - the >full text, with running footnotes, so that it is easy to see text and >commentary together. Not only will it be easier to read, it will be >easier to modify, and therefore more likely to be added to by novices. Yes! The Annotated Dictionary would by my favorite reference source for the core language. Dan wrote: >PS: If I ever finish the PrimitivePrimitives project, it could be a basis for >an alternative, highly redundant >Vocabulary. And I would not be too surprised >if some of the entries made it into the examples sections in the >true Vocab. >After all, they would be J code, not redundant (competitive to the normative) >English, and the >examples are already (purposefully) redundant. I have argued for defining J in J as well but it could become complicated,as Roger pointed out in http://www.jsoftware.com/pipermail/general/2001-November/008262.html , and extremely difficult to faithfully redefine some primitives as the current somewhat deficient emulation of the decommissioned foreign trace illustrates and "There would still be required a description in English that is complete in itself" anyway (which could also be annotated regardless whether it is clearly complete or not;). From: Henry Rich <henryhr...@nc.rr.com> To: Programming forum <programming@jsoftware.com> Sent: Sat, January 16, 2010 12:36:52 PM Subject: Re: [Jprogramming] The Ambiguous Dictionary >> it would be fabulous if the Wiki pages initially contained the >> full text from the Dictionary. > > We could also link to the definitive DoJ entry, without having to reprise it > (or keep it up to date), and still allow users to rely on Dictionary2 with > complete confidence. Of course we can quote and comment upon relevant > parts directly in the page. What we should produce is not a guide to the Dictionary, but a complete Dictionary with commentary. A user shouldn't have to hop between the guide and the DoJ to be sure that they haven't missed something. Start with the Dictionary text, and intersperse commentary: then you will have created The Definitive Reference For J. We should aim no lower. A document that is commentary only will not be worth reading for a while, because it won't have much. What it does have will be incomplete and will require the user to keep two documents open at once. What I want to see is something like the Oxford Annotated Bible - the full text, with running footnotes, so that it is easy to see text and commentary together. Not only will it be easier to read, it will be easier to modify, and therefore more likely to be added to by novices. I wouldn't worry about keeping our pages up to date - changes to the DoJ aren't exactly coming fast and furious. Henry Rich ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm From: Raul Miller <rauldmil...@gmail.com> To: Programming forum <programming@jsoftware.com> Sent: Thu, January 14, 2010 4:59:17 PM Subject: Re: [Jprogramming] The Ambiguous Dictionary On Thu, Jan 14, 2010 at 3:58 PM, Jose Mario Quintana <josemarioquint...@2bestsystems.com> wrote: > Should or should not, in principle, the dictionary be a > clear and unambiguous reference source for the core > language? In any case, where is (or was) the level of > redundancy regarding the fork’s order-of-execution? In the case of the recent example (the "verb phrase" thing), the dictionary provided a clear and unambiguous description of the parsing rules -- just not in the sentence in question. In the case of fork, in other languages when the specification leaves something unaddressed, the handling of that issue is up to the implementation. But J has always had a canonical "most recent version", and "implementation dependent but people only use one version" behavior can be relied on without having to be a part of the standard. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm (<"1) ________________________________ From: Dan Bron <j...@bron.us> To: J Programming <programming@jsoftware.com> Sent: Thu, January 14, 2010 2:19:32 PM Subject: Re: [Jprogramming] The Ambiguous Dictionary Without taking a stance on the issue (I'm not sure where I stand), I will note there are types of writing and writers who try to eliminate redundancy to the extent possible, and this is seen as a virtue. For example, mathematical proofs (IIRC), and for that matter, programming. There are also (different) types of writing and communicating where redundancy is intentional and typical. One could argue dictionaries fall into this category (where a definition is often described with multiple synonymous terms and some examples). Critical communiques are also often redundant (but not always re-expressive, which warrants attention). Then again, most of us ask for more redundancy in the DoJ when we're learning the language, and find it fine as it stands once we already know what it means (aka are already fluent in the language). Ken would have been the first to (emphatically) tell you that he did not intend the Dictionary to be used as a tutorial, but rather a reference, which is in accord with the previous paragraph. He wrote other material to be used as tutorials (and pioneered the use of those interactive tutorials, the Labs). And several others have supplemented this external (to the DoJ) learning material. -Dan PS: If I ever finish the PrimitivePrimitives project, it could be a basis for an alternative, highly redundant Vocabulary. And I would not be too surprised if some of the entries made it into the examples sections in the true Vocab. After all, they would be J code, not redundant (competitive to the normative) English, and the examples are already (purposefully) redundant. The project itself would be redundant to the Dictionary :) Please excuse typos; composed on a handheld device. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
