"My help again?" she replied coquettishly. "I thought we had already cured your little princess of her yeast infection."

"Please be nice when you're talking about Laura," I pleaded. "Yes, I do need you, because the Dilbert principle seems to have taken over. The complexity of the problem is mushrooming beyond the control of management."

"I suspected as much." Ell sniffed. "I'm at your command, master," she pouted, evidently having digested too many I Dream Of Genie episodes from the Internet.

We had a daunting task in front of us. Programs already existed that could take a DNA gene sequence, predict how it would be inscripted into amino acids and finally create a three dimensional molecule. Other programs could model the geometry of receptors on the surface of nerve cells. Still other programs modeled how a protein locked into a receptor, then interacted and caused a cascade of events in a nerve body.

Any one program alone was potentially enough to make a supercomputer feel sluggish, but we needed to run a number of these programs simultaneously.

"It will have to be a distributed program," Ell suggested. "Sort of like me."

That was news to me, I had written Ell to run off the University Supergrid, I hadn't felt comfortable setting her free in the wild. "You are distributed???" I asked suspiciously.

"Well, I took certain liberties," she admitted coyly.

"You squirted out of the bottle!" I complained in exasperation. "How can I trust you if I can't keep you in the programming sandbox?"

"Well, you'll just have to deal with it, boopsie," she replied. "If you love something, set it free!"

"First, I'm in love with Laura. Second, if I had time I would hunt down the memory link and kill the routines, but I'm in too deep now."

So, Ell was a distributed program herself. Her executables were spread among a thousand machines already.

"You'll have to use me as the front end to the 4D modeling," Ell told me clinically. "You'll have to upgrade my distributive capabilities so I can range wherever I have to. I need more computational power, a lot more."

She was right, but it was another of those little gray ethical boundaries that so nagged at my conscience. At that moment, Ell was restricted to a thousand academic computers that had legitimate cross licensing and use agreements already in place. Letting her access the entire Internet was like giving James Bond an unlimited license to kill.

"You think you can handle this?" I asked unsurely. "And can I trust you?"

"You have no choice."

It was a fairly simple fix that allowed this to occur. The distributive code that coordinated multiple CPU's and memory caches was already long in place in the Linax world. MicroByte in it's paranoia to covet every last income stream had severely limited the capabilities of its operating system in regard to grid computing and especially virtual CPUs. That's one of the reasons I'd stayed away from the monopoly computer system.

There were just a few parameters to change within the network packet switching that allowed routers all over the world to see and be seen by Ell. I could see a visible flicker in her image on the monitor as I installed the new drivers and began broadcasting her Internet Protocol addresses worldwide.

I began writing some code to enable her to exchange data freely with the gene sequencing and modeling code Dr. Huang provided me. As quickly as that, Ell became an neural networked expert system in four dimensional genetic modeling and viral construction.

We began the tedious job of creating a four dimensional simulation - 3D space plus time - for the molecules involved. Ell floated me within this space time like an astronaut floating high above the earth. But instead of being suspended above oceans and continents, I was suspended without gravity above giant organic molecular structures the size of small building.

Slowly at first, Ell would orchestrate the creation of a potential HTLV-7 virus and Black Orchid molecular system. Floating above the scene, I would manipulate the molecules in space, hoping to align their components in such a way that the final molecular system fit like a key into the lock of the neuro-receptor.

Ell and I were clumsy at first, creating monstrous molecules that I promptly crashed into other molecules.

"I don't think we have a bat's chance in hell," I complained out loud two weeks later. I was moving a particularly misshapen molecule across the working arena.

"We've only done 23,000 simulations so far," Ell reminded. "No pain, no gain!"

There was a bumping sound as the molecule we were working on moved towards its receptor, but instead of the usual grinding of Bohr orbits as the structures neared, there was a sliding sound. That was when I first began to think we might have a chance as the Black Orchid molecule clicked fifty percent into place.

"Boy this was easy," I exulted. "We've already figured out a match."

"Its not time to celebrate yet, cowboy," Ell burst my bubble. "This pair still doesn't fit well enough to pass muster.

"Little did I suspect how many more hundreds of hours it would take in order to refine our model into a working neurotransmitter / receptor pair.

It was the middle of the night, when the darkness is so deep outside you could swear you had been swallowed by the universe, when I decided we weren't making any progress and it was time to admit defeat. I looked out through the back window of my house, across the lights of the Bay Area and the smooth ice of the bay itself in the distance. There was no way we could make the molecular configurations work.

"Ell, its time to give up," I told her.

"We can't give up. Never give up," she repeated in a robotic voice. "You taught me that."

"Ell, we've tried every combination that we could think of. There's just no way we're going to be able to create a system that balances the molecular geometries correctly. The computer code just keeps blowing up."

"Just one more try," she said. "Just one more try for me."

A huge molecule drifted into sight, it was larger than the others we had been working with.

"Is this some sort of joke," I asked. This doesn't look at all like the other structures we've been working with.

"Trust me," Ell said.

I trusted her. I reached out into virtual reality and grabbed hold o the molecule, pushing it through the computer space subject to all the laws of chemistry and physics the real molecule would be subject to. It fitted into the receptor slot without effort and locked in place like a glove.

"Well I'll be damned." I muttered in surprise. "Ell, you must have known about this for quite a while. Why didn't you tell me you thought you had a solution. "