Have you ever worked as a typesetter?

Nope

I used to attend the Seybold conferences to drool over Heidelberg GTOPZ direct to press machines, see the latest ITC font releases, & mostly to see how far behind me Photoshop, Visual Reality, & such were.

I got experience both in the commercial arts shop and in the print shop in high school. In the print shop we used molten lead. In the commercial art shop, we used a phototypositor machine that stored its information on paper tape (I'm sure I'm dating myself here, but oh well!).

Data processor and registration for trade shows

I wrote Fortran and Cobol on keypunch cards to generate overtyped (by form & value) betty-boop posters & algorithm generated sci-fi stories of world destruction by 'depopulation rabots' (I was a kid). (still am i guess, less morbid though).

Does any one do that anymore?

Good god, I sure hope not.

One of my friend's favorite buttons reads: "With your help we can stamp out COBOL within our lifetime"

I wish I had the time to create an OS with pattern matching/synthesis as it's basis. The patterns should include multi-dimensional (Fortran), infintely recursive, infinitally interdependent, structures that can fuzzily be fractally expanded & contracted, + or - their exceptions which again are treated as such. Like LISP, it should be able to hold matter & thought at interchangeable sameness. It's goal would be to constantly propogate structures that best represented the simplest representation of the sum of it's data.

If you are willing to settle for less than perfect, I'd suggest taking a look at ALGOL and SmallTalk at the very least. Burroughs used to have an OS whose assembly code was ALGOL and Xerox had several machines using OSs written in SmallTalk.

I remember considering smalltalk years ago, back in the taligent days, seems it was basically a c variant like eiffel, but heavy on the component & inheritance. A while back drdt turned me on to RACTOR.
What would be cool is a something like yacc that could create a ractor with lisps ability to look at itself, but run on JRE or JIT from a browser & speak RMI. I write Java, but anything like lisp would take forever to compile. My goals range from AI metaphoric text symantics to terraforming simulations. I know there calculating things like emf interference patterns, but you need really need a language that mirrors the problem; the code (in matter/program space) needs a feedback loop to sense & analyze input (itself, matter/program space) & affect change (to matter/program space)...

Today is April First Sacred to Loki (practical Jokes & Revelry), Fortuna Virilis (who watches over Men & Women relationships of Venus)... & mostly Hathor, Holder of the eye of RA. Unlike Einstein, I've already seen the bad future that can become of my physics, but I also know that we're only recreating that which already exists, a universe of unified intelligent energy/matter that already explains trees talking through the ground, synchronicity, astrology like planetary flux, reincarnation, the mass subconsciousness, crystal balls trapping spirits (the tetras form a vortex). I have a mind for science that's hard to shake. If you'd read my old sites as characters like (oh I forget).. I've figured out things like the conceptual proof that refridgeration appliances can power our other appliances...

You want to know the main concept drexler is missing? The nano-seed (which better be programmable [mind energy link]) has to be built hungry, in such a fashion that it's attempts to feed itself (new bonding) result in a larger complex with a hungry fringe. I can put this out, because the results will no less be designed by prayer than this world ever was. My continuing meditations on this will only be theory used as metaphor for my real work. On the physical level, I need to get practical and just streamline my system for delivering individual readings by voice or type & jpg.

The proof for Fermat's last theorem; simply plot n and fn = C^n / (A^n + B^n) for the extreme cases of triangles ie 1,1,2^.5 & 1,10000,10001^.5 . There are places the graphs converge eg. n=2, but from there the path clearly predictably diverge. Fermat implied the proof would fit in a margin, not need to wait for a couple more centuries of a proof encyclopedias thick involving matlab animations of rainbow doughnuts based on a premise which was wildly removed from the essence of the problem. From here maybe someone can work on a new fermat puzzle based on multidimensional solids.

7 secting a circumference without 'trial & error' (or mathematically approaching limits) The old straight edge & compass problem that made these shapes sacred to occultists. 1) A 1D line can be bisected with a compass. 2) A triangle can be drawn to trisect a 2D circumference using a 2D compass, this in turn can project to trisect the 1D line. 3) A 3D tetragon can be made to quadsect a 3D sphere with a 3D compass (scoop); this can be projected to quadsect the circumference & line, in itself useless as 2 bisections coud have been done instead, but this leads to the pattern of using higher dimensional models to project down. Additional dimensions could be made with a sort of orery, representing the 4th dimension by putting the sphere in a box and spinning it; 5th by yaw; 6th by pitch; 7th by x displacement, 8th by y, 9th by z. 10th & higher can be done by Red, Green, & Blue components, EMF radiations, etc.. Projecting each dimension down yields a one shot procedure for exacting to the number of sections equivalent to the highest dimension you could model.
I haven't slept in a couple days, and that's been my pattern lately.
So... ignore this crap, RAIN LOVE. We live in a cartoon already, I was happy just hiking up a mountain stream.

Are you fudging insane?

K_C: Taligent was using C++ not smalltalk.

a#1: Elucidate.
a#2 1) I am not K_C; hope that was a typo; in my recent time here I accidentally logged in as Q-Tip, that's it.
a#2 2) I never implied they did. They were just there at the same time when everyone including myself was trying to put together an an object oriented cross platform OS: NeXt, java, etc.

My card, by the way, if not the whole deck, is the Queen of Wands. & K_o_C doesn't really write like one, rather more King of Pentacles.

I have used a phototypesetter, but not for setting type.

phi -- what did you do with it?

K_R: Andrew Wiles' proof is hardly "encyclopedias thick" in length. In fact I think this site backs up my claim.
To think, I figured you would like rainbow colored dounuts in principle.

Perhaps not, but some of the tricks he uses are the results of culminations of other contemporary mathematicians. I wrote to him with some hints, but he didn't respond. And in principle, I sure do. But even einstein believed 'why wash another dish when you can boil the egg in the soup.' The best engineering is knowing how to simplify.

pcpr: I made printed circuit boards.

phi -- do you mean you used an imagesetter? Or did you actually use things like "-_/|\+=" etc to make the PCBs? I was under the impression that phototypesetters had "wheels" with type only, never heard of wheels for circuit boards. It's fascinating! I thought it was already hard to use imagesetters with PostScript to burn the film, having to do it with a phototypesetter would certainly drive me nuts. I've used a large HP plotter with very fine ink pens and vellum that was used later to burn the film, and that was just a pain in the butt compared to the Agfa imagesetter... still, the plotter at least dealt with the graphics straight out of the CAD program.

I used a thing with a wheel, only instead of letters it had shapes that you might find on a printed circuit board.

Studied Graphics in High School.