SILICON SOAPWARE

wafting your way along the slipstreams of the Info Highway


from Bubbles = Tom Digby

= bubbles@well.sf.ca.us

= tgdigby@netcom.com


http://www.well.com/user/bubbles/



Issue 8

New Moon of August 25, 1995

Contents copyright 1997 by Thomas G. Digby, with a liberal definition of "fair use". In other words, feel free to quote excerpts elsewhere (with proper attribution), post the entire zine (verbatim, including this notice) on other boards that don't charge specifically for reading the zine, link my Web page, and so on, but if something from here forms a substantial part of something you make money from, it's only fair that I get a cut of the profits.

For more background info, details of how the mailing list works, etc., look at issue #Zero.

If you email me a reply or comment, please make clear whether or not it's for publication.


Now that Labor Day is getting close, here is a cartoon script for football season ...

Open in the back yard of a human couple's house, autumn colors on the leaves. Mice and birds are discussing what to do about the cat. Suddenly one of them notices that the tool shed is open. Inside we see a can of "Football Brown" paint, scissors, and a roll of duct tape.

Next morning as the early worms are out getting caught, we see one bird out in the open, seemingly oblivious to the cat's slow approach. But suddenly, just as the cat is about to spring, he's rushed from behind by a horde of mice and birds bearing lengths of duct tape. There's the cliche cloud of dust obscuring a fight, and when it clears there's this sort of cat-sized blob of duct tape that's wiggling a little and meowing faintly.

The mice and birds paint the faintly meowing cat-sized blob of duct tape Football Brown and carry it off in triumph to the stadium. A sign over the entrance tells us the big game is that afternoon.

We see a couple of scenes of football action, perhaps involving whichever player is carrying the "ball" falling on top of it as the other 21 players fall on top of him.

Then we hear the announcer: "He's going to punt, and there's the kick! It's over the fifty, the forty, the thirty... wait a minute, it looks like the tape's coming undone. And the cat's loose, running past the twenty, the ten, up into the stands at the five, both teams in hot pursuit. Drinks and hot dogs flying left and right as they race through section A-7, now A-6, now heading up to B-5. Now as you can see from this helicopter view they're out in the parking lot ..."

Back at the house the couple are sitting on the sofa, watching TV. The game is on with the husband watching and the wife none too happy with hubby's choice of program. Then there's a commercial and he goes to get another beer while she switches channels to a soap opera.

Now their cat, escaped from being a football, finally makes it home and runs into the house through the cat door. Since it's a cartoon the football players can distort themselves enough to squeeze through after the cat, one by one, SquoooootchPoit! SquootchPoit! SquootchPoitPoitPoit! for a total of 22 "Poit!"s. And maybe another "Poit!" or two for the officials.

Wife grabs the phone, starts to dial 911 to report two football teams chasing a cat round and round the living room, busting up furniture and everything. Just then the doorbell rings. Husband or wife opens it to reveal a TV camera crew. And with the camera crew is a manager holding a big bag of money. Wife hangs up the phone.

So our closing scene is a long shot of the house shaking violently to the sound of occasional crashing noises as the chase continues. We zoom in on the couple, now independently wealthy, sitting out on the patio. They can afford separate TV sets now, so the wife is watching a soap opera as the husband follows the cat chase on TV. Then we pan over to the bushes and zoom in on a mouse telling some birds, "Even if it isn't quite what we planned, at least it's keeping the cat occupied."

End of cartoon.

A little violent perhaps, but maybe absurd enough to keep the violence from seeming real enough to be objectionable.


"Where are you taking that shovel?"

"To the wagon."

"Why?"

"You told me to hitch up the mules."

"Then it's true what they've been saying?"

"What's that?"

"That you don't know your ass from a hole in the ground."


Have I Heard That Song Before?

Someone somewhere brought up the matter of looking up a tune you vaguely recall, but don't know the title of. Or suppose you're researching folk music, studying a tune and its ancestors and cousins. Or maybe you're a lawyer handling a copyright suit. You know one tune, and want to know what other tunes sound similar. What kinds of search tools exist for this?

I've seen books that list tunes by note values, but they suffer the same problem people who can't spell have with dictionaries sorted alphabetically. Get one note (or letter) wrong near the beginning and you're miles from your target. What you really need is some kind of a pattern-recognition algorithm that fuzzily looks for similarities.

One possibility I've thought of is to represent the melody as a graph of pitch vs time. As you look at this with coarser and coarser resolution, the curves for different melodies begin to look alike. If at first nothing in the database matches the graph of your unknown tune, you make the images blurrier and look again. Sooner or later something will match.

If you want a measure of similarity between different tunes, you see which pairs of melodies become indistinguishable at what resolution. The finer the resolution at which they look alike, the more similar they are.

To generalize beyond images of graphs: Apply some lossy transformation to a set of musical themes. Then apply some other lossy transformation, or perhaps the same one again. As you repeat this, certain themes become indistinguishable from one another, until ultimately you have all the themes degenerating to the same primordial item. By noting when various distinctions get lost, you construct a similarity tree.

To search, you might first transform the presented sample to something near the primordial root. Then you find the most similar major branch. Then search that major branch for the most similar minor branch, and so on, down to the leaves.

There are many possible transformations, and I have no idea which, if any, would work. Research is needed. Perhaps we should study how people mis-remember melodies, then get ideas for transformations from that?

Once we have some promising methods, we'll need a large database of tunes for them to work with. That involves finding existing online databases, data entry for material not already online, and possible copyright questions. But these are bridges others know how to cross. I'd say the main hurdle is coming up with some way to do the comparisons.

Is this feasible? Would there be a market for it? I suspect so, but I don't know how to go about proving it.


Some car won't start. Doors open, and a bunch of cartoon characters pile out. They open the hood, look at the engine, but don't know the first thing about fixing cars. Onlookers gather, but none offer to help.

One of the characters looks at the dead car, at the unhelpful crowd, back at the car. He starts shouting about how the whole thing makes him hopping mad.

Another joins in. Now several are raving incoherently, waving their arms around and actually starting to hop up and down in their frenzy. Soon they're all jumping up and down as the cacophony grows to a climax. It's mostly incoherent babble, but now and then we can make out the words "Hopping Mad."

Suddenly the car's engine roars into life. Anger forgotten, the characters pile in and drive off. Just before they leave, however, one spectator manages to ask what's going on.

The driver replies: "Haven't you ever seen a jump start before?"


Patenting Pi

One controversial development in the field of molecular biology is the patenting of genes. As I understand it, somebody will discover a gene in some organism or something, figure out what the DNA code is, and then patent it in case it turns out to be useful. There have been lawsuits, and there has been talk of changing the patent laws to eliminate this kind of thing on the grounds that nothing is really being invented. I'm not a lawyer, and have no idea how this is going to turn out. But last I heard, people were still applying for patents on what is essentially found information.

Now suppose the gene patents are upheld? That opens up all sorts of possibilities. For example, why not patent parts of pi?

Crank up your computer, calculate some digits of pi nobody else has gotten to yet, then patent your results. Anybody who wants to use pi beyond the previously calculated precision will have to pay you royalties. And if the same sequence is later found to turn up somewhere in other irrational numbers like e or various square roots, that's infringement, and you can have your lawyers go after users of those numbers for royalties also.

In fact, I may go into business soliciting investors for this. I even have a theme song, which I can play at seminars to get the investors I'll be trying to recruit stirred up and in the proper mood. (I may have to pay song royalties to Winnie the Pooh, but that's a minor expense compared to what I expect to rake in if the scheme works.)

The "Patenting Pi" theme song: 

     Patenting, patenting, patenting pi.
     I have a scheme whose potential is high.
     Ask me my business and I will reply
     Patenting, patenting, patenting pi.


     Patenting, patenting, patenting pi.
     Find some new digits, we can if we try.
     Ask me my business and I will reply
     Patenting, patenting, patenting pi.


     Patenting, patenting, patenting pi.
     The offer is open, so come in and buy. 
     Ask me my business and I will reply
     Patenting, patenting, patenting pi.

We could be seeing the start of a whole new industry here.  Math will
never be the same again.  


                                -- END --

About the Author

(40k GIF)

This page was created by Tom Digby and is copyrighted with a fairly liberal "fair use" policy.

Email = bubbles@well.sf.ca.us

Home Page = http://www.well.com/user/bubbles/