Bill,

Thanks for your thoughtful reply.

I'm not suggesting that J be made available in different language  
versions. I simply lament the loss of the special characters. Indeed,  
one of the really good features of the special characters is that they  
are not language-specific. If well-designed they convey their meaning  
readily to readers of any nationality. Sort of like the international  
road signs.

I like your suggestion about "thinking in J". For me, that happens in  
two ways. In writing J, I start with the abstractions inside my head:  
"This'll be a rank 3 array that needs sorting. I need to do a rotation  
and apply a function to various cells etc". Then I have to translate  
that into J's symbols. It helps if that step goes smoothly (so that I  
do not have to iterate, rephrasing for errors of syntax or wrong  
symbol choice.) Having symbols that are evocative helps. And if it  
helps, then I tend to remember it. The second way I need to be able to  
work with J is reading it; both my old sentences and those of others.  
Several authors have taught that a good way to deepen one's  
understanding of the language is to read sentences written by others.  
I'm all for that. I'm always for a free ride and if others have a good  
solution to a problem, I'm happy to adopt it in my own work. Now, in  
order to read J, I have to parse the text before me. Perhaps I'm a dim- 
wit, but I find this really difficult, for the reasons I described  
earlier. If only there were something that would leap off the page,  
something to help break up the whole line into smaller pieces.  
Because, this step is difficult, the pleasure is diminished and I  
avoid reading much of other people's code.

So, I enjoy the visualization part, breaking a problem down into steps  
executable in J. That's what I consider to be "thinking in J". That  
part I don't forget. I just want some help translating thoughts into  
actions. If it were less difficult, I'd remember it better.

Mike Powell

On 2-Jun-09, at 6:11 PM, Bill Harris wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Mike Powell <[email protected]> writes:
>
>> Why not allow J users to use Unicode names?  The J language can keep
>> the reserved names it uses now. Just let the users experiment.  
>> Surely,
>> if they make use of this and run into problems transferring code or
>> commentary back and forth, that's their look out; not something  
>> that J
>> Software Inc has to address. After all, Unicode data is being
>> transferred back and forth all the time in other contexts. The issues
>> are not special to J. Perhaps, by letting the users experiment with
>> different skins, something interesting and useful will emerge.
>
> Mike,
>
> I'm all for freedom, but I have a serious caution here:
>
> Please think carefully, and then leave well enough alone.
>
> Why?
>
> First, I was involved years ago in early thinking at one company about
> localizing software products.  I ran into attempts others had made to
> "localize" things like Pascal, for example, by translating all the
> keywords into various languages.  That takes a standard language and
> makes it nonstandard.  That requires you translate each and every
> program when you move between platforms.  It was a disaster that was
> thankfully dropped early on.
>
>   Pascal is a good example for another reason.  Many vendors seemed to
>   want to augment their versions for commercial reasons,
>   unfortunately.  That meant they'd have a standard Pascal plus a few
>   features people would want to use that would lock them into that
>   vendor.  That messed up reusability.
>
> Even simpler things make trouble.  For example, the x. -> x conversion
> challenge still crops up occasionally, although I guess I have to  
> admit
> that the language is not worse off for the change. :-) If we make
> transliterations or translations or conversions such as that, we make
> code reuse very hard.  J is so concise, though, that code reuse is
> eminently practical, for I can learn phrases with immense utility  
> simply
> by reading and experimenting.  It would be a shame to lose that.
>
> It's been my experience with J that most things make great sense  
> after I
> grok them better.  I don't know what you're doing, but I encourage you
> not to think of J as a dialect of APL and translate mentally back and
> forth but simply to think directly in J.
>
> Bill
> - --
> Bill Harris                      http://facilitatedsystems.com/weblog/
> Facilitated Systems                              Everett, WA 98208 USA
> http://facilitatedsystems.com/                  phone: +1 425 374-1845
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.9 (GNU/Linux)
>
> iEYEARECAAYFAkolzdUACgkQ3J3HaQTDvd+tlACgg2mh8fXLdFFnvUUQWb1xz3Nv
> ijgAn1HyIG7BlFWF+3IdRCceiT6gDjLk
> =dDS5
> -----END PGP SIGNATURE-----
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to