Selective blindness works on me too, Joey. As I read your reply the suspicion grew on me that you were accusing me of being selectively blind... :-)
In fact I was; and having pressed "Send" I wondered if I hadn't over-expanded on a simple misunderstanding caused by my woolly phraseology, and that a 1-line restatement of what I'd said mightn't have sufficed. I'm glad you got more out of it than that. Anthony Camacho said once in public that what the APL world needed was a well-documented account by someone who'd used APL to make a million. Well, I've had the privilege of working on (at least) one such product. It didn't make *me* a million, but my bosses all knew I loved programming so much I'd do it for nothing. I think I could still restore the history of the Kunzle Planning System (KPS), the brainchild of George Kunzle. IBM knowingly let him take it with him when he took early retirement from its UK HQ at Cosham where he'd worked as a planner -- one of the people who calculate how much profit IBM plans to make in the coming year. His simple but significant tool became Adaytum Planning (AP), then Cognos Planning, when Guy Haddleton sold out to Cognos in 2003 for -- what was it -- $5M? Then IBM recently bought Cognos -- and effectively purchased back at great cost a system they'd nearly let slip into landfill somewhere, as you put it. Throughout its career nobody apart from developers seemed to have any inkling it was written in APL. Why? Because we took care to conceal the fact. APL is still viewed as a sordid perversion in IT circles. Ian On Mon, May 25, 2009 at 11:02 PM, Joey K Tuttle <[email protected]> wrote: > Ian, > > I'm sorry - but you demonstrated a phenomena that is of interest to > me. As I tried to think of what I would call it, I came up with > "selective blindness". A google of that showed it was an appropriate > choice - e.g. > > http://sixsentences.blogspot.com/2008/03/selective-blindness.html > > The way it manifested in my case (this time) was that I read your > post as saying that you wouldn't be doing anything more with j and > that the portability had sealed your decision. How's that for not > reading carefully?!? > > Anyway, I'm very glad I asked you to explain, because your response rates a - > > WOW!! > > and I think it could be a big part of an answer to Devon McCormick on > what to say about J at BAPL09 (Devon, comments?) > > I imagine that you know my origins were in APLSV as well, and I agree > with all the points you make below. Certainly I was aware of Mohansic > ASDD and many other uses inside IBM. This bings up an interesting > topic which we might start a new thread with as well. > > Recently at the CHM ( http://computerhistory.org ) I was part of a > chat about how few documented cases of APL usage they had access too. > Almost everyone in the APL/J/K community has a bag full of ancedotal > success stories. But it seems not enough of them have been credibly > recorded and donated to the museum. > > I think APL and related systems have contributed a LOT to computing > in the past 40 years - but while many people have "heard of it" and > there are strong opinions about it, documents are hard to find and > many, such as important systems inside IBM are probably in a landfill > somewhere.... > > Anyway, thank you for answering my question! > > - joey > > > > At 22:25 +0100 2009/05/25, Ian Clark wrote: >>On Mon, May 25, 2009 at 6:51 PM, Joey K Tuttle <[email protected]> wrote: >>> ... I am curious what you mean when you say, "[I] >>> probably will never write another original APL application. J's >>> portability between Windows, Mac and PDA alone sees to that for me." >>> Can you explain to me what you mean by that?? >> >>I'll try to explain, Joey, without taking too long I hope... >> >>Ive written a lot of APL in my life. It started in 1973 with APLSV in >>collaboration with IBM ASDD Mohansic (remember it?) and in recent >>years has entailed originating and maintaining code in major >>commercial products written in Dyalog APL and latterly in APL+Win -- >>and a lot of experience in porting code between the two. Code written >>by people who didn't appreciate []ML<-3, or that the []AV's are >>devastatingly different, and that the two sets don't overlap. Code >>that needs frequent execution of B[a;b;c;d;e...]<-1, where B is a >>massive Boolean and you have (a b c d e ...) as a vector, and don't >>want to blow up the ws because the engine converts Boolean to 32-bit >>under the cover to implement an indexed assignment... I could go on >>and on. >> >>This was back in the last century, and things may have gotten better >>with APL since... but I doubt it's that better. >> >>You could say I'm heavily invested in APL, from two of its leading >>vendors, and not initially inclined to try new-fangled dialects. At >>first J gave me the queazy feeling I had when, as an Algol programmer, >>I first saw APL back in 1973. >> >>Well, after less than 10 hours flying time with J, I was able to see >>where an awful lot of nasty code could have been avoided. Code which >>mopped up most of my work hours, not to mention the weekends preceding >>each version release of my product. >> >>Now I'm retired, but I still haven't given up developing products. >>There's TABULA (Vector 20.4) I'm tarting up for the eco market. I'm >>prototyping it in APL and was delighted when Dyalog shipped a version >>running under Windows Mobile, because that I see as its main platform. >>For the customer, I hasten to add -- not for me as developer, because >>I don't do ships in bottles. I plan to develop on the Mac from now on >>-- but Dyalog hasn't got a Mac version. Windows had made me my living >>for the last 15 years, because it sure makes work for the working man >>to do. But now that I'm working for myself, not on someone else's >>time, I don't have to be bothered with its sheer nonsense. >> >>Also I don't have a spare couple of grand to keep up with the latest >>releases of Dyalog APL and APL+Win -- but that's not the key issue >>because I could always find a customer to buy me the products I need. >>No, the key issue for me is that I've written J code in Windows, >>including GUI code, transferred the files to the Mac, also to my HP >>iPAQ, an easy matter because they're ASCII txt files and it's just a >>case of moving the dongle... and the app works First Time. >> >>Now I have a lot of nostalgia for APL. It's a rare expertise -- not an >>easy thing to walk away from. I still have APL code out there to >>maintain for clients, so I won't say I'll never write another line of >>APL. But for an original (ie not a heritage) application, J is going >>to be my tool of choice. >> >>Why? For 100 good reasons... >>1. code portability, especially with GUI. >>...The other 99 don't matter. >> >>On the downside, I have to set-to and learn J in my grey hairs, which >>I'm finding truly mind-bending. But on balance, life's too short to go >>on battling with APL, under Windows. >> >>Ian Clark, aka J Rabbit. >> >>BTW who's seen >>http://en.wikipedia.org/wiki/Criticism_of_the_APL_programming_language >>? >>It would be better called "The Case Against APL" because like a good >>counsel for the prosecution, the author carefully avoids saying >>anything which could be construed in its favour. But it covers the >>ground. >>---------------------------------------------------------------------- >>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
