knudsen (12/29/82)
Two questions about Smalltalk: (1) Is their any hope of ever creating a version that would run on a reasonable home computer (8-bit micro, 32-64K memory)? OK, is it possible to make a "Tiny-Talk" version that would do the above, while preserving the advantages of OOP etc.? Background for above: (LISP has been discussed on net already): Alan Kay suggested years ago using ST in his "Dyna Book" computer (the Epson HX-20 is ALMOST there), which led me to believe that ST should not need "gold plated" (Dorado) ECL workstations to run it. Also a recent article said that ST objects have 10s-of-bytes overhead, not 100s like 42's. So why so big & slow? Given the calibre of PARC's software jocks, the non-existence of home-computer versions can only mean (a) absolutely can't be done (b) nobody's taken the time yet (c) Xerox doesn't want it done. Since Xerox is well into office machines, might ST at least fit into an 8086/8 or 68K based "personal"/business box? Why isn't Xerox selling it themselves for this? (2) Don't misconstrue the following as another example of "I don't need ST to do OOP." I want to know if the following is a step in the right direction from conventional to OO programming: I just wrote a C macro package to implement FIFO ring buffers of arbitrary (user-given at create time) number of ring segments and ring lengths. The resulting data structure may be accessed only thru certain function calls to add & remove data, test block signals, etc. In fact, I could imagine that no data structure exists, that I am simply talking to this queue "object" thru calls. Is this something like OOP? Or is it just good old David Parnas' data encapsulation? (CM-U grads, many of whom work at PARC, should remember him.) Please give me some reasonable answers to above. I'd like to know more about the language, philosophy, and CHANCES THAT I MIGHT EVER AFFORD IT AT HOME. Our Computer Club room is locked over the holidays, so I can't get in to borrow the BYTE mag issue on ST. And I'm keeping too open a mind to comment on the appropriateness of its cover (tho it might apply to some of the ST discussions on the net). mike k