Discussion of large systems
wizard at voyager.net
Wed Nov 1 13:20:30 CST 2006
On Wed, 2006-11-01 at 11:12 -0700, Richard wrote:
> In article <1162398718.6425.53.camel at linux.site>,
> Warren Wolfe <wizard at voyager.net> writes:
> > On Tue, 2006-10-31 at 19:54 -0700, Richard wrote:
> > > Does anyone have enough tape drives that they can do the
> > > multi-drive merge/sort tape algorithms described in Knuth's
> > > Art of Computer Programming?
> > That would be six, I believe. But, why? Just to say one did it?
> No, because having all those tape drives visibly spinning makes for
> a familiar sight when it comes to old iron. People are *used* to "old
> computers" looking like a large bank of spinning tape drives.
Oh, okay. To me, "it looks cool" is a perfectly acceptable reason
to do it. But, if the aim was to sort some massive quantity of data, a
bit of technical assistance with the algorithm might be in order. As I
believe I mentioned before, laziness plays a factor in my decisions,
although I prefer the term "time management" for reasons which should be
obvious. I certainly am going to ASK if you want help with sorting data
before I offer you assistance.
But, don't think I don't approve of just wanting one's collection to
look good. For further effect, you might wish to mount (taking care to
allow for safety) some roman candles, to create that Irwin Allen effect,
should there ever be an infinite loop in one of your programs. The same
people who know that "old computers" consist of spinning tapes know that
programming errors start fires and cause explosions.
> It boggles my mind why people think I am trying to perform "the
> world's most efficient and timely sort" when I propose such a stunt.
> If I wanted to do things efficiently and in a timely fashion, I
> wouldn't use old hardware in the first place!
Yep, that's a given, and I never thought you were trying to set any
records in your effort. It could be, though, that you were actually
intending to sort data that is on tape. It has happened -- I know, I
For myself, I like to compare algorithms. There was a WONDERFUL DOS
program out there back in the early IBM PC days that took random data
(or data pre-sorted in various ways) and sorted it using various
algorithms, from bubble sort to heap sort to quicksort. The cool thing
was that all the retrieval and display code was identical, so one could
literally WATCH the data being sorted, and the time it took was affected
ONLY by the efficiency of the sorting algorithm. VERY instructional.
This program was called, unimaginatively, SORTDEMO. As computers got
faster, it became pointless, as ALL the sorts were over about the same
time they started, so one could no longer watch the data being
re-arranged. I just checked, and I still have that program. Maybe I'll
set up an old PC just to run it... it's sort of like a cyber-lava-lamp.
Warren E. Wolfe
wizard at voyager.net
More information about the cctalk