INTERNAL AND EXTERNAL SOURCES OF TECHNOLOGY The Lab That Ran Away from Xerox Bro ustat On a golden hillside in sight of Stanford University nes- Hes Xerox's Palo Alto Research Center, a mecca for ti fed researchers—and an embarrassment. For the $150 million it has lavished on PARC in 14 years, Xerox has reaped far less than it expected, Yet upstart gompanies have turned the ideas born there into a crop of promising products. Confides George Pake, Xerox's scholarly search vice president: “My friends tease me by calling PARC a national resource. Not thatthe center has been utterly barren of benefits for Xerox. The company’s prowess in designing custom ships, to be used in future copiers, comtes largely from PARC. So do its promising capabilities in computer aided design and artificial intelligence, PARC did most ‘of the research for Xerox’s laser printers, now a $250- million-a-year business growing at 45 peteent annually and expected to (urn a profit in 1984. But Xerox hasn't cashed in on PARC'S exciting arch on computerized office systems, which was the center's original reason for being. According to Stan: ford J. Garret, a security analyst who follows Xerox for Paine Webber. the company’s office systems business Reprinted with pcr rom the SepenerS. 198. isan ne magarine, 1983 hy Tish. tn. Al ih rescved lost @ horrific $120 million last year and will probably drop $80 million in 1983. “Xerox has got a lot out of PARC." says Garrett, “but not nearly as inuch as it could have or should have.” Why has Xerox had trouble translating first-rate re search into money-making products? Partly hecause the Process takes time at any large company —often close to decade. Sheer size slows decision making. and the need to concentrate on existing businesses impairs manage- ment’s ability to move deftly into small, fast-changing ‘markets. This i a special problem for Xerox, still over= ‘whelmingly a one-product company whose copiers counted for three-quarters of last year’s $8.5 billion in revenues and almost all the S1.2 billion in operating profits Serious organizational flaws, acknowledged by high Xerox executives, havealso provedahandicap. PARC had weak ties to the test of Xerox, and the rest of Xerox hhad no channel for marketing products based on the re searchers’ efforts. The company has revamped office equipment marketing five times in the last six years. “Xerox has cteaked, twisted, and groaned trying to find ‘out how to use PARC’s work,” says an insider. While ‘Xerox has groaned, disgruntled researchers have left in frustration, These Xeroids as they showered PARC concepis—for designing personal computers, office equipment, and other products—on ‘competing companies PARC’s influence outside the walls of Xerox is an ironic tribute t0 the ambitious vision of the man who Founded the center in 1969. C. Peter MeColough, then672 pyar mance: enacTmENT oF TECHNOLOGY s News's president, charged PARC wit Iimnligy Xerox weeded 1 be formation’ in the ulfice. The new center, in a miuteuly el providing the gant three-story building whose rock-sinen atria Foster meditation, quickly lured many of the nation’s lean computer scientists, offering whal an alumous calls blank cheek and LO years without corporate interference Roughly half of PARC's money went for research in ompater science and hall for research in the physical seietees. Most a the glamour radiated from the eomputer crew. Members were notorious For long hair and beans ud for working imes shweless and hitless, They held rueous weekly meetings. in the tbean-bay room.” where people tossed around blue-sky concepts while reclining on huge pelletfilted hassocks, ARC's hotshots were nos just playing at bei Before long, computer scientists eecognized PARC as the ading source of research on how people interact with ‘computers The handsuft poliey at Nerox’s headquarters in Stam: fic, Connecticut, proved a double-edged sword, PARC researchers used their freedom 10 explore concepts for personal computing that have since swept the industry All sorts of computers, including some from Apple and IBM. now offer “hit-mapped” displays, which PARC ‘championed 10 years ago, Such displays link each ofthe thousands af dots on a video sereen to a bit of informs rion stored in computer memory. thus allowing the com puter to change each dot and create very fine-grained im ages. Apple's now. easy-to-use Lisa flaunts a display that ‘can be divided into "windows" For view es ‘of work at once. as well .a pointing device, or “mouse for giving commands, PARC did the lion's share of work ‘on both ideas at all hows — sean severut pi But Neror’s loose management also encouriged Pal Not outssand-bolts product development, By the mid 197s. the center was hard at work on the AU, ane. pensive machine with some of the attributes ofa personal ‘computer, which Was supposed to serve as a tolype. Alto and its software becarne so popular inside Neros. whete PARC installed a couple of thousand atthe stems, that de researchers began to see to overstep its charter, whieh was to do research, uuch pro- them es commercial p they slaved like distillers of moonshine whiskey to op the Ato For the market duets, Out of top mnagement’s Sen oun. which vas ehompioning st rival mechine market as Nerow’ S140. nal workstation, Uae a pen TRATEGY.-DEVELOPING A FIRATS INNOVATIVE CAPABILITIES ally relies on is oxen processing power and memor Star worked well only when linked with oth ae equip (See "Nerox Noms toward the Olicear ne Future" Fovnme, May 18, 1981) PARC rhs non took om he development go but also dominated Xerox unit set upto text mar seach polyps. Ths group gt oer I Atoninne inthe White House bah bouses of Congress. and she ccnpanies snd universities. Unwilling to support ra tnachines, Xerox gillained the Altos in FO ie dated the whole test.anarketing group Veterans ofthat group have bcen the eet evangetigy ‘fPARC technolo John Ellenby,one ofthe nity agers, later Founded Grid System. His Compass com Pier approximates seme prescient PARC convepts ng used inthe to. It's portable, uss a bitsmapped dip an easily hooks up with remote commpters. Nt S8,000 0 $12,000, the Compass sounds too costly t0 be popular hut Grid expects revenues of mote than 828 milion iy 1982. first fll ear of operations: in August Grid ei it was onthe verge of prota Anothertnanager of the test marketing unit, Ben Weg. bret, had previously been onc of PARC's brightest eh sical talents. Convergent Technoloies of Santa Clas California, founded in 1979 to make workstations Picked off Weubret and to colleagues 10 design sol ware. Cowvergent’s won processing. program shuns Some of its origins inthe form ofa “piece able,” a type of sotvare developed at PARC. Hallows computers with fel sna memories to prccess fon documents edocs this by storing only the changes mide when editing. stong with the original version, stead of the orignal Dts a foll-length edited version, as other programs do Consenicnces like that have helped Convergent land contacts that could puedce some $450 milion in sles mpanies tht ave developed their Charles Sinionyi, whe defected from Hungary styles himseli "the messenger RNA of the PARC ins” He worked atthe center for seven years, mostly on Bravo. a texting sofwite program forthe Ato that never er Xerox to big computer ¢ reached the market. “We werent supposed todo pr Tike than. 'o Brava started out as a sub- terfue, Bul when people at Nerox sav it, they wanted 10 uuseit inside the company. Bravo was why people used Al ios just as VisiCate was the reason people bought the ‘pple I «chi product's market potential Simonyi expected some brilliant executive Phot wasn't domb. he ays, “but it was naive fo assume stich a person would coe front Net Simoni founda ssarmer welecnneSECTION ONE: INTERNAL AND EXTERHIAL SOURCES OF TECHNOLOGY 673 Micmsolt Corporation, bused in Bellevue, Washington, sic fn up $50 million in sales of personal-compater solvate inthe year ended last June, big chunk of this ears sales. Which should approach $100 million, will Come Hon Microsoft Word, a streamlined version of Brine. Lisa is the unkindest cut of all, In December 1979, were Jobs. then Apple's vice chairman, visited PARC with sone colleagues to poke around, They saw Small talks mining tools."Their eyes bugged out recalls Laserenc who helped develop Smalltalk They inslerstood it significance better than anyone else who bi sisted.” Seven months later, Jobs hired Tester, having decided fo use many Smalltalk features inthe Lis The Lisa had (© be priced at $10,000, two to four times Jobs’ earlier estimates. But it seems to be taking ‘fl Apple claims to have shipped as many Lisas in July the fist month they were available, as Xerox has shipped Siar. oF 8010s, in 19 months of availability, The Star Wy concepts used in the Lisa, has prog which emborlies oem ill tarred. The influential Seybold Report on Prom fessions! Computing calls it “a jack-of-alltrades which des owe really well” Sales suffered initially because some othe Stas software was late in coming to market Ofice equipment analysts have started referring 16 PARC style systems as “Lisalike,” not “Starlike.” Apple's nextcomputer, Macintosh, scheduled to ripen into acom- mercial product by the end of this year, could further items Apple with PARC’s ideas. The engineering miat- ‘ager for Macintosh came trom PARC, where his last big project was a personal computer From shis, Xerox might appear to bave muted the » make it big in personal computers with PARC's Some Xeroidls are sure the company could have becn an early winner if only it had launched a less expensive Alto in the late 1970s, Unlike the Star, the Alto ‘say on “open!” computer, easy for outsiders to program, Indspndently written software has helped touch off the esl computer explosion, so the dissidents have. a ‘outsiders can't write chines Pint, Because the Star is “close Progra For it To mourn the Alto, though, isto blame unfairly those sho billed i. Xerox was out to produce office equipment ind no olive equipment supplier, including IBM. fore S16 thut personnal computers would compete with theit us. Iwas inconceivable that the cost of computer temery would decline 31 percent a year. as it has forthe last. years, or that today’s microcomputers ould be “poset as yesterday's mainfzame computers, Xetox Pil ws ilk concentrated not on freestanding. personal computers but on clusters of workstations that shate the tse of computer hardware, That say, customers could spread high hardware costs across many workers. And suppliers could dettay the costs of theit prized sales fotces with big-ticket orders Besides, Xerox had, and still has, ulterior motives in the office. Competition in the copier market keeps grow ing, and the company’s chief aim has been to protect copier installations by strengthenis i its control of large lucrative accounts. Companies that can sell comple oflice systems—-workstations. with reliable software. printers, and data-stora ‘work—have a stronger lock on their customers than do suppliers of stand-alone equipment. Thus, the Stat, which works well only when hooked up with other Xerox sear seemed to fit the company’s strategy better than fie standing Tittle computers would. ‘The complete-system approach, moreover, was more compatible with Xeroxsexpunsive ways of thinking than the alternative of making piecemeal improvements on an individual machine like the Alto, Big companies often ccan't make the modest efforts needed to probe ei markets. “I's a problem when you're getting your feet wet in a new business,” says Jack Goldman, formerly Xerox's research chief. “Ina large company, every prod ct must be @ home un (justify the costs of marketing and development. ‘That has been especially true at Xerox, which owes its existence to Xerography. one of the longest homers o1 record. Top management “followed the big-bank strat- s one veteran. “They wanted to build absolutely the best office system instead of taking things bit by bit. ATPARC, the company’s urge to build the best at the ex pense of the merely better, like an Alto, had its own name: biggerism, Biggerism could pay off in some ways, to be sure Xerox has big hopes for Ethernet, a PARC-invented net- work that uses a cable and transtiting devices to connect different types of office equipment, By souping up the performance of PARC'S original version of Ethernet ‘Xerox drastically raised theeost oFhooking up. toas much «as $5,000 per comnection. That move discouraged sales and deterred other equipinent makers trom adapting their Imuichines to talk through Ethernet. But now, improved chip technology has sliced the cost of connecting by about Wwo-thirds. Over 70 office equipment makers are using Ethernet or plan to. including Apple. The temporary set- back helped keep Ethernet from becoming tie industry standard, but 7 is. standard. The only 0 likely to sot a standard is IBM.) e devices, all Linked into a ne + company674. parr nce: Enncrwent oF TeoMMoLooy sth Xerox still thinks PARC’S work can produce some big bits. No one is more convinced than John Shoeh, a re rmarkably hard-boiled former PARC researcher whe be time the company’s office systems chief fast October His inst priority isto expand the number of Xerox prod= communicate over Ethernet (20 do now, in- cluding laser printers and facsimile machines). Making a winner out of the Star will take mote elfort, Because the technology is old and the system tries to do so many things, the workstation seems expensive and inept in many functions, especially compared te Lisa Shocl wants t0 bring out a fess costly version of the $15,000 Star, whieh be s movement to narrow the Lisa's potential market share The other clas, i his view, will he IBM's personal co puter armed with a ListTike set of programs written by VisiCorp, Priced at some $7,000, thae system won't com pete directly with the Star bur willbe far cheaper than the Lisa. It will also tap into Ethernet —thanks to a helpin, hhand from Xerox. Says Shock: oing t0 be a between the lower priced Star and commodity {ype computers that run better software, IPH be a tough place to compete. The company’s support of PARC has never wavered This year’s budget of $35 million oF so sill set a record ut changes have taken place. Last March, Xerox ap. pointed a new director of PARC, William Spencer. A vet- tan oF two decades at Bell Labs, Spencer admires AT&T's ability co transfer technology out of the lab by attaching satellite labs to major manufacturing. plants PARC'S main shortcoming, thas been a tack ‘oF management attention, We started things that didn’ match what was going on in other parts of Xerox.” Spencer is trying to produce a better fit by meeting a couple of times a year at PARC with Xerox's division managers, some of whom havent visited for years. Every three Weeks or se he breakfasts with Shoch. snd they've sarted a joint hiring pr spend their fnst year or so at PARC. then join the office There's sae: he feels, systems group, Time is oa Spencer's side. Having taken its Humps in the ofice systems business. Xerox has a better fie on What kinds of products make sense. While Shoch’s divie sion stil steugetes to discover a successful way of selling office systems, PARC, h ed much of the tech nology MeColough sought, is stepping up its work on a now frontier: vers-huge-seale integrated circuits used sverything fron di copier breakdowns 10 ‘connecting personal compurers with mainframes, “The Foundation for our futwe will be the next generation of chips says Spencer. who originally ca setup a fine for makina them shales part of our Wok now ‘When a company wns tsk it isin a ne hay ness sod ase OF tecology is neces Bn hardly sufvient. Widhot clear anderstaing ni’ porate strategy and presse from hingt) mgenn, troup, eventhebenttecogistscan get otothand ne icky par isto strike balance hetcen econ creativity and yeting your money's worth, Ne 1 PARC wy Du Pont Kevlar® Aramid Industrial Fiber Clayton Christensen As Howard W, ral managerof Du Ponts Tex tile Fibers Department, heacked toward his Wilmington office on the morning of February 27, 1974, he knew that in a fow hours he would soon fice some stiff questions from the company’s Executive Committee. Entering the lobby of the Nemours Building he glanced at the huge John McCoy mural from the 1939 New York World’ Fair, which proclaimed in image and word Da Pont’ motto, Better Things for Better Living... Through Chemisty.” Less than at week earlier, he tad submitted to the X-Committee, as it was known in Du Pont parlance, the first of a three:part appropriation request for capital t0 build a S0-mitlion pounds/year plant for the manufacture ‘of Kevlar® Aramid fiber, his department's newest devel opment—a miracle fiber of extraordinarily complex chemistry that was pound-fTor-pound five times stronger than steel. Kevlar truly was, Swank believed. a better Savane: Profs Chston yr bis ce a the bass incon lier eet o nfl in of 3 adnate sition. Ws go valent of linger eae. "Du Pont Kevlar Ari lst er iy. 15 writer by Professor David. Hel Matin Bie ello st the supervision af Paves Richa! S. Revenblonn, Captnisht looky the Presiden an lk of Harn! Callere oem copies w poyest permission to reprodce mute. all ESI SS 708, ‘rite Marsan Business Sele! Pb. Bosom, EA O26 opr gedp ar atl et, No pt th pubic ny BE phobic, sord neta sytem, toed in spent a ing evening. ur ierwise-—sihou the rms of Hara Basins 5