"Precisely," replied Vigal. "Now let me show you the programs with a direct relationship to 9971.7-programs that will explicitly rely on the information from your work." The neural programmer pulled back the focus and let a whole new level of the blueprint slide into the room. A disparate group of perhaps twenty beads began to blink. "Now, if we add in the components of the system that rely on information from those programs ..." The focus pulled back even farther. Hundreds of beads were now flashing in perfect synchronization. "And so on and so on," concluded Vigal, waving a jittery hand at the exponential explosion of blinking beads.
Natch could feel the impatience swelling within him. He began drumming his fingers on the side table. "All right, I get it. There's a lot riding on this program."
"Any neural program."
"Fine. But I don't see what you're worried about. My code meets all the standards. It produces consistent results."
"I have no doubt of that," replied Vigal somberly. "If I ever had any doubts about your programming abilities, Natch, this has certainly dispelled them. It's not you I worry about-it's the companies working on all these other programs whose skills I question." He flipped the back of his hand at the pinpricks of light that had replicated throughout the room like a cancer.
"Why should I care?" said Natch through gritted teeth. "That's their problem."
"But Natch, you must understand ... you're not working with skin moisturizers or, or, breath fresheners. This is neural software. One major discrepancy between any two of these programs could cause a massive brain hemorrhage. And that is an unacceptable result. That's why we have to work slowly and with careful coordination."
"That's why we have Dr. Plugenpatch standards-"
"The standards are only one small part of the procedure," said Vigal with a tinge of sadness in his voice. "For process' preservation, Natch, why do you think I'm always going off to speak at all these conferences? It's so those of us in neural programming can keep on top of what the others are doing. Certainly, it's costly and time-consuming, but it's also effective. We have a higher standard to live up to here in the memecorp sector."
Natch crossed his arms in front of his chest and sulked. He tried to look at anything in the room but Serr Vigal. "I don't see how you can make a profit doing all that," he said. "I mean, who's going to wait three years for all of you to coordinate an upgrade to some obscure piece of cranial nerve software?"
Vigal shut off the diagram with a snap of his fingers, revealing the cheap carpeting and second-hand furniture of Natch's apartment once more. "You're beginning to understand life in a memecorp," said the neural programmer quietly. "We can't make a profit. We can't just rely on Primo's ratings or the whims of the marketplace to test our products for us. Because, as you say, people do not have the time or the inclination to pay attention to most neural software. If we didn't get outside funding, I would have to close up shop and send everybody home."
"Then why do it at all?"
"Because I enjoy it," said Vigal, "and because someone has to."
There was a long pause. Natch could feel 9971.7 mocking him from its berth in MindSpace. "So what do you want me to do? Abandon the whole thing?"
"I want you to do exactly what you're doing," replied Vigal, "but slow down. Natch, you've created an excellent long-term plan for the future of 9971.7. Tomorrow, you can begin by rolling back the changes you've put in place, and then we can start to enact that plan one piece at a time."
Natch had no response. He had expected Vigal to point him to some hidden flaw in the program architecture that only a wise and seasoned programmer could see. Instead, his guardian had only confirmed that the problem did not lay with Natch; the problem was the memecorp system itself.
Silently, Natch cursed the day he had ever signed on to an apprenticeship with Serr Vigal, and wondered if he would make it through the rest of his apprenticeship without going completely insane.
A few weeks later, Natch abandoned his grandiose plan for the neural software. He hadn't lost confidence in his abilities; on the contrary, he was more certain than ever that he could bring the program to a higher plain of functionality. But Natch had decided to leave Vigal's employ in eight months, when his contract ended. Until then, he would get much more experience trying his hand at a variety of projects than tinkering with just one.
Serr Vigal took the news coolly but with a tinge of disappointment. "Where will you go?" he asked his protege. "The fiefcorps?"
"That's not the place for me right now," said Natch, leaving the obvious unspoken: he still suffered from the taint of the Shortest Initiation. Hiring someone with Natch's notoriety could provoke a boycott from the creeds and the L-PRACGs, or a backlash from the drudges. "I've decided to set myself up on the Data Sea as a ROD coder," the youth continued. "Like Horvil."
"Are you sure, Natch? ROD coding can be extremely-"
Natch cut the discussion short. "Yes, I'm sure," he said.
Vigal did not begrudge his young apprentice's decision. In fact, he seemed to forget all about it over the next few months. Natch quietly upgraded a number of optic nerve programs to the new Plugenpatch specs, working at the glacial pace his master had requested. When Natch multied to Vigal's office to collect his end-of-contract bonus and say goodbye, the neural programmer responded with a surprised "Oh!" and gave him a feeble hug. Natch could see a host of worries fluttering through his guardian's head, but Vigal had evidently decided to keep them to himself.
Thus ended Natch's brief career in the memecorp sector.
RODs were Routines On Demand, bio/logic programs that catered to the indulgently rich. There were no contracts, no guarantees, no fringe benefits. ROD coders simply scouted the Data Sea for a spec they could engineer quickly, rushed to build it before someone else did, and then launched it in hopes that their patron would like it enough to pay. The applications for RODs ranged from the frivolous (Fab-a-lous Nails 15, now with programmable cuticles") to the lascivious (Tit-o-rama 8, the total breast bounce controller") to the purely ridiculous (DiscSpeak 3c, "for the true connoisseur of ancient recorded-sound emulation-now with pop, hiss and warble!").
An expedient programmer could typically launch two or three RODs per week. Horvil had maintained that pace for over a year before moving on to the more specialized field of bio/logic engineering.
"Let's estimate three sales a week," said Natch to his old friend one night as they strode around London discussing career options. He summoned a virtual calculator in the brisk spring air and began plugging in numbers. "So we'll multiply three by the average asking price for an ROD, subtract out equipment and overhead ..."
Horvil sliced his hand through Natch's holographic calculator. "Wait a minute, Natch," he said, shaking his head. "It's not that easy."
"What do you mean?"
"I said I launched two or three RODs a week. I didn't say anybody bought 'em. You never sell all the RODs you launch. Sometimes you get beaten to the punch by another programmer. Or your patron changes his mind and pulls the spec from the Data Sea just for the fuck of it. Or you're sabotaged by the competition. There are assholes out there that post fake ROD requests, you know."
Natch frowned and adjusted his numbers downwards. "Well, it's still not that bad-you've just got to factor in recurring revenue. You know, maintenance fees, subscription fees, upgrade fees. That's how the fiefcorps make a profit, right?"
The stocky engineer chuckled, pleased to be in the know for once. "Nope, can't count on any of that either. These are RODs, Natch-you have to build 'em so quickly that they won't withstand any kind of heavy use. The shelf life for a ROD is about twelve to fifteen weeks, and that's only if you've done your homework. After that, the buyers just get bored and move on to the next new thing."
Natch began to wonder if he had made a rash choice in leaving Vigal's employ. Memecorp work might be torpid and dull, but at least it provided a sense of stability. But once Natch stepped out on his own, nothing was guaranteed. A year from now, he could be trol
ling the old cities and living in trashed-up skyscrapers like the diss, like his mother had done.
But certainly Natch could make a decent living in the ROD coding game if Horvil could. He had a tremendous respect for his hivemate's intellect, but Horvil's business sense was a little skewed. People tended to fall into the ROD world because they couldn't hack it in the real bio/logics market. But Natch never doubted that he had the skills and the pedigree to make it to the top ranks of the fiefcorps. He was not on his way down; he was on his way up.
All Natch needed to do was persevere, produce quality work, and establish a reputation. Eventually, the ill wind that drifted around him would dissipate and the brand on his forehead would fade; the Shortest Initiation would be permanently tossed into the dustbin of Yesterday's News. Then the channelers and capitalmen and fiefcorp masters who patrolled the Data Sea for fresh talent would find him, and he would resume his rightful place in the bio/logics world.
Natch's end-of-contract bonus was enough to keep him afloat for a month or two. Vigal offered to buy his young protege a set of bio/logic programming bars as a parting gift, but Natch declined and bought the bars himself. He did not want to feel beholden to his guardian for anything. It was not an idle investment; the extended function sets on the new bars would enable him to take coding shortcuts and thus program faster.
Now that Natch had emerged from Vigal's shadow, the city of Omaha held no appeal to him. He hopped around the globe looking for a place to settle, and finally found an apartment in Angelos that suited his tastes. The place made Horvil's spare room look like a palace by comparison. Still, it had everything he would need to start a ROD business. There was a bed to sleep in, space to hold a decent-sized workbench, and proximity to downtown Angelos, where the public multi facilities were abundant and cheap.
The next morning, he got to work.
Natch decided to begin with a field he was familiar with, so he chose optics. He skimmed the Data Sea and found a request for an eye transformation program that looked like it might be a good place to start. Bio/logics had made setting one's eye color as easy as editing a database entry, but the woman who had posted this request wanted something more. Vellux of Beijing wanted her eye color to sync with the colors of nearby flowers. In a room of violets, she wanted violet eyes; in a field of ivy, she wanted green.
It seemed like a simple enough programming task. A morning spent nosing through the Dr. Plugenpatch archives helped him better his understanding of the optical programming interface. Natch fetched the OCHRE specs from Dr. Plugenpatch, projected them onto his workbench in MindSpace, and started planning his strategy.
Natch found plenty of machines floating around the eyeball that he could harness to accomplish his task. Thanks to the OCHREs, he could query the iris and determine the color of its pigment; he could also query the retina and parse the colors in the user's line of sight. But a number of vexing questions remained. How would the program identify flowers in the retinal image? How would it distinguish between petals, stems and leaves? How would the program funnel the millions of shades of yellows and reds and purples into a narrow palette of 16 colors? What if Vellux was looking at seven different flowershow would the program rank the order of importance of these flowers and assign an appropriate eye color?
The longer Natch struggled to unravel these tangled questions, the more questions arose to ensnare him. Normally, it took hours for the body to process color changes through the personal preferences database. Unless this woman Vellux planned to stand still for long periods of time, he would need to find an alternate solution. Luckily, he found a number of sub-routines on the Data Sea that would do the job quicker. Natch chose one called Weagel's Eye Wizard, which had received excellent ratings from Primo's a few months back. But the program required access to a batch of proteins for building the pigmentation ... which could only be done by requesting resources from another OCHRE nearby in the choroid ... but the OCHRE in the choroid needed to register its supply requirements with the brainstem.
It took Natch most of the day and into the night to come up with a satisfactory blueprint for the project. At six in the morning, he sat back and took stock of his progress. The holographic model floating above his workbench looked like a mutated grasshopper, but Natch knew he could not afford to trifle with aesthetics this time around. As he was examining his handiwork, the building interrupted him to slide a bowl of hot oatmeal onto the kitchen counter. When was the last time I ate? Natch asked himself. He could not remember.
But food would have to wait. Natch created a new instance of the MindSpace bubble and placed his model inside it for future reference. Then he grabbed two of the bio/logic programming bars out of his satchel. The new bars were light enough to wave in MindSpace for hours, yet solid enough to withstand thousands of accidental bashes against a workbench. Natch took a deep breath and attacked the empty MindSpace bubble with zest.
Morning became afternoon; afternoon became evening.
The young entrepreneur sculpted his code quickly, using virtual blocks of logic as his marble, and programming bars as his hammer and chisel. Gradually the mass came to resemble the mutated grasshopper of Natch's diagram. He had been working for thirty-six hours straight when he finally laid down the programming bars. The bowl of oatmeal had disappeared, and Natch couldn't remember if he had eaten it or if the building had simply whisked it away untouched.
Natch stepped into the hallway, which was lined on both sides with tulips. He fired up EyeMorph 1.0 and was pleased to discover that everything worked as designed. His eyes quickly slid from their natural blue to a mottled shade of purple. Natch retreated to his living room and tested his handiwork against a number of floral images on the viewscreen. So far, so good.
Rest is coming soon, he promised himself. I just need Plugenpatch approval, so I can launch the program on the Data Sea, and then I'll sleep.
Natch approached the Plugenpatch process with more than a little trepidation. As he had discovered while apprenticing for Vigal, a successful test was no guarantee of approval. Fiefcorp programmers could not hope to cover all the combinatorial possibilities of a fully functioning OCHRE system. No, only large entities like Dr. Plugenpatch and Primo's had the facilities to do that. Natch swallowed his fear, packaged up his work, and routed the program to Dr. Plugenpatch's automated verification system.
Eight minutes later, as Natch sat on his sofa sucking down a fizzy bottle of ChaiQuoke, EyeMorph returned from the verification system peppered with rejection notices.
Mindful of the time, Natch tore through the Plugenpatch recommendations. He realized to his chagrin that he had left a loophole that might allow excess protein buildup in the choroid. Any decent OCHRE system would be able to deal with such an anomaly as a matter of course, but Dr. Plugenpatch's standards were rigid and uncompromising. The catchphrase from a thousand Creed Conscientious advertisements rose unbidden in his head: Always preserve your bodily computing resources! Natch sourly picked up the programming bars again and began reweaving connection strands.
The next rejection took eleven minutes for the system to process.
The rejection after that took sixteen minutes of analysis.
Natch decided to abandon subtlety and just finish the wretched program. He suspected that someone had already beaten him to the Data Sea while he was here fumbling with Dr. Plugenpatch rejections, but he couldn't just abandon the project now. Natch furiously patched up the remaining holes, disabled a few features that seemed problematic, and fed it into the Plugenpatch system.
Twenty minutes later, the verdict was clear: success!
Natch hastily bundled the program together, slapped on the standard fore and aft tables that the Data Sea required for its cataloging agents, and launched. He called up the ROD optics listings on his viewscreen so he could see the results with his own eyes. The evidence on the new releases board glared at him in small black letters:
EYEMORPH
Version: 1.0
Programmer: N
atch
Yet he felt no sense of triumph. EyeMorph 1.0 may have slipped past the gates of Dr. Plugenpatch, but Natch knew the program was still riddled with inconsistencies-the kind of inconsistencies that Primo's would certainly notice when they dredged the Data Sea for their bimonthly summation of the ROD coding world. Not only that, but because Natch had used Weagel's Eye Wizard to perform some of the heavy lifting, part of his profits would be swallowed up by licensing fees. He would be lucky to break even on the project.
Natch was shambling towards the bed for a long-overdue slumber when a message arrived.
You gave it your all I hope you had fun 'Cause you got your ass kicked
By CAPTAIN BOLBUND.
Horvil was at a loss to explain Natch's failure. He wriggled his head free of the blankets and stared drowsily at the ROD listings scrolling up and down his bedroom viewscreen.
"How does he program so fast?" griped Natch from across the room, where he was wearing tracks into the carpet. "Who is this guy? `Captain Bolbund'? He beat me by an entire day on EyeMorph, Horv! What's he doing that I'm not?"
The engineer flopped around to face the wall. "Maybe it was a fluke," he said. "Maybe he just got lucky. It happens."
"It's not a fluke. This Bolbund has beaten me four times in a row now. "
"Four times? How the fuck d'you run into the same guy four times in a row in this business? That's no accident."
Natch shook his head. "Of course it isn't an accident. I keep taking him on, and he keeps massacring me. Even worse, he always sends me this awful poetry when he wins." The young entrepreneur forwarded some of Captain Bolbund's doggerel to Horvil.
Horvil read silently for a minute. "This is terrible," he mumbled. "Ten thousand spell checking programs out there, and this asshole still spells slaughter with a `w'." He sat up in bed, stretched, and shot Natch a worried look. "Listen, Natch, I don't think you get it. When you're a ROD coder, you gotta keep moving or you'll get in a rut. Didn't your mother ever tell you that you win some and you lose some, but life goes on?"
Infoquake Page 13
