RE: CLIP's test of virtue -- and memory immunization ?

[Brian.Henderson@xxxxxxxx]

From: Robert Holmgren
Sent: Friday, December 10, 2004 12:13 AM

>> The file I was pasting to is 7.4 kb. Another document of about 1.5 MB was open...

>1.5Mb! That's huge. That's encyclopedia size - a couple thousand full-text pages. XyWrite will go bananas at about 2.3 - 2.6Mb (sum size of all open files in various XyWrite >windows), and I can only speculate that Editor may get flakier as it approaches that number (never done any tests).

In my world, a 1.5Mb ascii file is fairly small...3 to 5Mb is probably about average. I've found
that the maximum size that XyWrite (3 and 4) can open is 7.4Mb (you CAN open an even bigger file,
maybe up to 8Mb, by using Windows' "send to" function - I've always been very curious as
to why that trick works).

Note: Although XyWrite CAN open larger files, Robert's numbers of 2.3 - 2.6Mb might just represent
the limits under which Xy will function relatively well...see below. (By the way Robert, where DO
those numbers come from? Are they documented somewhere, or were they "handed down" from
the Creator(s)? I've seem them here before and always wondered.)

Mostly, extra large files behave well...but after ten years and several thousand jumbo (to Xy)
documents, I've found that there can be (but not often enough to force me to use another
program...like Perl) problems with big files. The problems, though, are always inconsistent. One of
the things I do most often in files is run complex search and replace macros...where something is
found, the cursor is moved in a particular pattern, then some sort of action is performed when the
cursor gets where it's supposed to be. A particular set of actions can be performed a thousand times
in a single file and, every once in awhile, XyWrite misses a beat...a backward search will land the
cursor 1 character off, or 5 cursor-right commands only get executed four times. That sort of thing.
It's always inconsistent though. I can start over and re-run the macro on the same file and it won't
happen again, or it will happen in a different place. In fifty files, something like this might
happen on!
 ce...in another twenty, maybe twice. These things never seem to have a pattern.

The most consistent problem I've found with huge files is with Search (SE, CI, CH). The bigger the
doc, the more likely that a search will fail in some way. If you think of the file as being
physically real, like a continuous sheet of paper, then the failures seem to relate to
"distance". A hundred targets will be found within the first third of the file and if
target 101 is near the end of the file, Xy might not find it (this, oddly, doesn't seem to be
related to Xy's technique of splitting some files into pieces). I've many times hit what seems to be
the end of a search (not found), moved to the end of the file and searched backwards for the same
thing, and have found more that Xy didn't see. One very curious related quirk is the difference
between CI and CH. In really big files, CI is much more likely to "fail" than CH. It just
seems to "poop out" at some point and its search/replaces stop being performed. The
curious part is that CH will (mostly) NOT fail in the same ci!
 rcumstance. It's almost as if (and I know some of you will hate the way I describe it) XyWrite,
being forced to actually (visually) "go to" each target is then "physically"
closer to the next target, and so makes it easier to "see". It really does seem like Xy,
when commanded to use CI, and never "moving", can't "see" to the end of some
files. After experiencing this over and over it's the only way I can describe what seems to be
happening. (And out of habit now, I NEVER use CI...even on small files.)

I'd have to say, a consistently repeatable problem like Si's is probably not size related.

BTW...this disquisition (or is it "disquistion") isn't really about Si's problem. I just
thought I'd take this opportunity to describe some aspects of XyWrite that, even though most of you
(maybe even none of you) are not likely to run across, nevertheless might find interesting (OK
OK...I admit the real reason is my feeling "left out", because the way I use XyWrite is so
different from most everybody and I don't feel I can contribute very often).

-Brian H.

Brian@xxxxxxxx
http://www.XyWrite.com