Nah, this ain't workin'
It's no way to do it
Just puttin' numbers on a CRT
This ain't workin'
There's a better way to do it
Show me my data in full three-dee
I'm using a spreadsheet to sort out options, and to compare
the costs of different combinations of services from different
providers. This is all well and good -- it's a more suitable
use of the tool than what I usually do with it (like most people:
use spreadsheets as quick & dirty databases, one table per
sheet, using vlookup() and hlookup() to create relations[1] if
the database is more than just a flat file).
The problem is, this is the second problem in two weeks for
which I want a three-dimensional spreadsheet. I want to see a
transparent cube, with values in cells addressed by row+column+plane.
I'll settle for slicing my data up and linking chunks with
references and lookups, since that's what I've got, but I want
that three-dimensional representation. Intuitive. Convenient.
(Actually, for one problem I really want to see a tesseract[2].)
Instead, I'm sitting here trying to figure out the most reasonable
way to slice things up to be laid out on one or more two-dimensional
grids.
Y'know, I've said before that a lot of representations of
VR in fiction strike me as
silly -- nifty, dramatically effective, but silly -- because they'd
use resource-intensive renderings for things for which we already
have very efficient interfaces. So I mostly see VR as something
for games (and other entertainment), and for specialized
applications. It makes sense for a biochemist to play with
modelling software that lets her "walk around" and "grab" virtual
molecules. Depending on what they're making, it could be useful
for mechanical engineers as well. The "everything you see is a
metaphor for a chunk of code or a real-world interface"
approach from some cyberpunk fiction strikes me as ... well,
difficult to get right (in the sense of being more useful than
a more straightforward but less interesting -- in a novel/movie
context -- interface, unless you're tweaking complex information
streams to take advantage of human pattern-recognition abilities[3],
but even then it seems as though it'd be more useful for alerts
and general status reporting than for detailed information and
hands-on interaction[4]. So I've had trouble believing in worlds
where full-sensory VR is the way all tech-savvy characters interact
with computers. But as of this morning, I'm ready for consumer-priced
VR just for the sake of a three- or four-dimensional user interface
for Gnumeric or Excel.
I'm ready for my VR multi-dimensional spreadsheet now. As soon
as I can get ahold of the goggles and the data-glove, and find out
where to download the demo software... But I wanna try it out for a
while before I decide whether to get a jack implanted in my neck for
a neural hookup.
[1] Yeah, I'd prefer having a real query language and
greater efficiency (enough vlookups and Excel gets pretty slow),
but when I'm creating a database for myself I'm usually impatient
to start using it, and misusing a spreadsheet means being able to
take advantage of certain data-entry shortcuts without having to
design custom input forms. Obtaining better front-end tools should
probably go on my to-do list, I suppose. I use real database tools
when I'm solving someone else's problems (because then I'm
much more concerned with doing it right than just getting the first
bunch of answers).
[2] It's just more intuitive to me, to represent data
in the number of dimensions the data have, as much as can be
conveniently rendered. I hated
Karnaugh maps in my Computer Architecture class until I
realized that the three-variable map was a 2x2x2 cube "unfolded"
by putting a hinge in the middle of one face, and the four-variable
Karnaugh map was a similarly unfolded tesseract. When I drew a
cube on the board to solve a problem in class, the professor asked,
"Okay, wise guy, what do you do with four variables," so I showed
him: I drew a tesseract (okay, a two-dimensional representation
of one common three-dimensional projection of a tesseract, if'n y'wanna
get all technical at me), I'd been using my own notation to map
six-variable symbol tables onto a tesseract for homework problems,
because I never worked out drawing pictures of more than four
dimensions in any useable way[5]. The instructor found me annoying.
[3] I must remember to dig up the InfoWorld
column in which someone suggested an "environmental sounds"
approach to data-center status monitoring. I know from
experience that subtle changes in one's acoustic environment
can be a very effective way of receiving status information
about one's equipment: I remember being able to hear a
paper jam in the Xerox 1075 copier and usually know which
section it was in before the machine had detected it, while my
attention was focussed on something else entirely. I remember
being able to tell when someone had logged in on the Xenix box
I administered, from the sound of the hard disk. Now imagine
engineering that sort of thing intentionally, so that data
center staff know "the web server is experiencing unusually
heavy load" because of the rough equivalent of the sound of
twigs snapping, or "the WAN has started having latency problems"
because "the birds stopped chirping". This is where I can see
VR interfaces being especially powerful, but this doesn't
require VR to implement. I've been meaning to write a
separate entry about it for a while, and may get around to it yet.
[4] And yes, geek (and SF fan) that I am, whenever I
make this point my brain spawns a background process trying to
come up with more ways in which it would be useful.
Let's just say that I've simplified my argument for the sake
of brevity, and we can discuss when and how I see a metaphorical
VR UI being more useful than decorative another time. And that
my thoughts on the matter are far from finished.
[5] In middle school I was told that humans could not
visualize four-dimensional objects (obviously, by someone who
hadn't seen enough three-dimensional projections of tesseracts
to find one that "folded up in his brain" the right way to make
the idea click). In high school I was told, "of course we can
visualize four dimensions with training, but not five." Several
friends took this as a personal challenge. One claimed to be
able to picture six-dimensional geometric solids. I managed
to visualize a few simple shapes in five but nothing complex
-- just enough to convince myself that there had to be other
people who could do geometry in more dimensions than that
"visually" in their heads, if "mere ordinary me" could reach
the edge of five with a struggle and hold four dimensions in
my head long enough to sort out diagramattically whether I was
doing polar<-&gr;rectangular coordinate conversions correctly.
Not that I find four-dimensional visualization easy
-- I do have to slow down, and keep reinforcing bits of it that
try to slip away -- but it's "there" enough to be useful to me,
and four a four-dimensional representation of four-dimensional
data to be a good way for me to spot the patterns and relationships
in those data. (Note that drawing four-dimensional
diagrams to show somene else what I'm thinking remains quite a
problem.) How to build four-dimensional representations
into a VR interface is something I haven't thought about yet.
Oh bother. The footnotes took over again, didn't they? And
I just sprayed for those last week.