|
|
OU Notes
Accidental Empires
- Accidental success of relative amateurs
- Success of USA over Japan etc. (component focused) and Europe (over-
concerned with style)
- Harvey Allen & Bob Metcalfe. MIT, 1960s
- George A. Miller: Harvard psychologist, 1950's. "The magical number
seven, plus or minus two" (no. of random no's memorised quickly).
Bell curves & standard deviations.
- 99.7% of people fit into 7(+/-2) category, hence 0.3% are above or below
the norm. Therefore one can presume that 0.15% of the population are
exceptionally gifted at remembering lists of random numbers.
- Good software or hardware designing involves keeping track of complex
flow of data through a program or machine.
- Charles Simonyi - great programmer, part of the above 0.15%
- Donald Knuth: Stanford computer science professor, 60's. Invented 'TeX'
language to drive laser printers, and 'Metafont' language to design fonts.
Used above to publish a book, which didn't look good. Not his fault as
hardware couldn't cope with his forward-thinking software ideas.
- Mainframe technology lasts approx. 20 years whereas PC technology
reinvents itself ever 18 months.
- Mainframes, circa 1975 - air-conditioned rooms, punch cards. Companies
sell the mainframe, then license out software at a monthly rate as well as
making much money from servicing and programming.
- IBM made System 650 mainframe in early 1950s.
- Personal Computer business grew up around the semiconductor industry in
what was dubbed 'Silicone Valley', California
- Bob Noyce & William Shockley - started Fairchild Semiconductor - the
archetype for every Silicon Valley start-up that has followed.
- Venture capital
- Splinter-group abandoning current product line to establish radical new
technology. Small team operations.
- Bob Noyce & George Moore- went on to start Intel.
- Noyce's personal managerial style revolved around empowering staff to
speak freely and further research at all costs.
- Semiconductors - volume business. High development but low bulk
production costs.
- Bob Noyce, 1959: Needed to keep costs down and invented integrated
circuit (IC)- several components together on one piece of silicon, thus
eliminating labour from electronic assembly line. Insulating silicon oxide
used on face of chip (planar process) then thin metal lines printed on top
of oxide thus connecting components on the chip. Photolithography
method.
- Jack Kilby, Texas Instruments - had similar idea, but couldn't be made in
bulk as gold hand-installed wires were used.
- Gordon Moore - invented Moore's Law - the number of transistors that
can be built on the same size piece of silicon will double every eighteen
months. True to date.
- Ted Hoff, Intel, 1971 - invented the microprocessor - a single chip
containing most logic elements used to make a computer.
- Large mainframe computer companies worried by microprocessor based
computers showing up their large expensive machines
- It generally takes society 30 years to absorb new technology into everyday
life e.g. film inventes 1980's and popular in 1920's.
- Inventions often find a major use doing something they weren't designed for
e.g. radio was invented for two-way wireless communication but found it's
mass use in broadcast entertainment.
- Initial interest in Ics and microprocessor driven PCs was from hobbyists.
- Intel, 1980s - quality problems caused by dust - Intel's shipping clerk had
been opening incoming boxes of superclean silicon-wafers to make sure
they weren't being swindled. In turn, this ruined millions of parts and cost
millions in improving the cleanliness at the company.
- Business axiom : When a business grows past £1 billion annual sales,
it becomes too large for any one individual to have a significant impact.
This is obviously not the case for high-tech industries.
- Intel - begat the dynamic random access memory chip, which made
possible MITS (a computer), which in turn made Microsoft possible.
- Bill Gates & Paul Allen, 1975: Discovered the MITS Altair 8800 (never a
commercial success) whilst at Harvard. Decided to write the first high-level
language.
- Wrote a program to make a PDP-10 mini computer function as an Altair
and wrote a version of BASIC (pre-existent language), called MITS BASIC.
- Original BASIC programme was by John Kemeny & Thomas Kurtz, from
Dartmouth College.
- Gates & Allen wrote other BASICs for other microcomputers, and fought
Ed Roberts of MITS and won, when he claimed he owned MITS BASIC.
- Gary Kildall - invented first microprocessor operating system. Was
computer science teacher at Monterey, California. Hired as consultant to
Intel in 1974. Wrote emulator, Interp/80 and high-level language PL/M.
He later created CP/M (Control Programme for Microprocessors) to be used
On a PDP-10 running a TOPS-10 operating system.
Later invented BIOS (Basic input/output system) - which split each new
controller from CP/M into a different computer code module.
- Gary Kildall then started Intergalactic Digital Research (now Digital
Research) and sold CP/M direct to users. Made millions.
- Digital Research were slow to provide languages to work alongside
their operating systems - they left that to Microsoft, thinking that Gates
wouldn't ever compete with them on operating systems.
- Apple II, 1978 - first commercially successful microcomputer - created by
Steve Wozniak (inventor of Apple I), Steve Jobs & Mike Markkula
(retired Intel veteran).
- Steve Wozniak made Apple II a simple machine using clever hardware
tricks to get good performance at smallish price. Microprocessor
and video display shared same memory. 1/4 of normal ICs in floppy
disk controller. Colour machine. Had finance and business uses
although wasn't a mainframe replacement.
- Apple IIs compelling application came in the form of the spreadsheet
Visicalc, invented by Dan Bricklin (Harvard Business School) '77-78.
Originally written in BASIC and first released Oct '79.
- Perfected for Apple II by Dan Bricklin and Bob Frankston (MIT) and
published by Dan Fylstra (Harvard Business School) who specialised in
selling applications direct to retailers and end users.
- Fylstra started Personal Software and Bricking & Frankston started
Software Arts. Personal Software distributed Visicalc for a 37.5%
profit share.
- Some Apple staff wanted to buy rights to Visicalc outright, but Mike
Markkula wouldn't allow it due to cost.
- 'Look and feel' - cause of many legal wrangles in the 1990s. Essentially
copyright cases over User Interface. Foolish arguments as few concepts
ever spring fully grown without reference to prior art. Derivative.
- Microcomputing stole from mainframe and minicomputer worlds.
- Most more recent ideas came from Xerox Palo Alto Research Centre
(PARC) - This was a basic research establishment near Stanford Uni-
specialised in knowledge for its own sake (as opposed to research and
development which develops products for sale).
- Most basic research is generally conducted by companies with at
least 50% market share.
- True product of industrial basic research is insurance, not knowledge.
- In business world, just because something can be built, it doesn't mean to
say that it will be built - espec. if it threatens current profitability.
- Main research establishments are/were Xerox PARC, IBM Labs & ATandT
Bell Labs - all USA.
- Punch cards replaced with input device. 80-column line entered through a
computer terminal. Improvement, but only active line was available to edit.
No back-tracking and altering.
- Doug Englebart (Stanford Research Institute) was concerned with user
interface and interaction. He knew people wouldn't work without ability to
edit, so worked on bitmapping the screen (point-by-point technology as a
means for roaming). Used a mouse to move a cursor on screen.
Some PARC work built upon his ideas.
- Xerox PARC CSL (Computer Science Laboratory) was the best place in
the world during the 70's for basic research.
- Invented the Xerox Alto - Built in networking, black on white bitmapped
screen, mouse and hard disk. Alto had no compelling application.
- PARC also made first laser printers.
- By not commercially developing its inventions Xerox lost chance to become
dominant player in personal computer revolution.
- Bob Metcalfe (see ch2), 1973, PARC - developed Ethernet networking
technology, whilst trying to find means to speed up laser printing. He later
left PARC to develop Ethernet further (p111)
- Ethernet copied telephone idea - listening, if other transmissions are
heard, then waiting, then trying again at random intervals. Ran at 2.67mbts.
- Bob Taylor moved to PARC from 1970 as consultant, having been a
psychologist in Texas and having head ARPA (Advanced Research
Projects Administration for US Defence) as well as working at Uni. Of Utah.
Created flat organisational structure at PARC - no middle management,
just a 30-40 strong elite. Cross-fertilised ideas by constantly changing
workgroups. Taylor set the direction for research with ideas coming from
technical members of staff.
- PARC cloned the PDP-10 time-sharing system to create the MAXC (Multi-
access Xerox computer), but soon realised that time-sharing wasn't viable.
- PARC perfected WYSIWYG GUI (graphical user interface)
- PARC employees lost heart as they saw their contemporaries researching,
developing and releasing their inventions and making their fortunes, hence
the early 80's saw an exodus of PARC researchers into the entrepreneurial
start-up field.
- William H. Gates III (Bill Gates) and Paul Allen started programming
for their school aged 12. Wrote academic scheduling programmes.
- Later vowed to put " a computer on every desk and in every home running
Microsoft software" when there were only 100 microcomputers in the world.
- Decided to monopolise the software business - operating systems,
languages and applications.
- Charles Simonyi (see Ch2), Hungarian intellectual and prior PARC
employee (invented Bravo, a WYSIWYG word processing application for
Alto) and came up with metaprogramming concept.
- Metaprogramming was to be a more efficient method of organising
programmers to write software. Metaprogrammer is designer, decision
maker and controller in a software development group. This group of
people write the code itself. Basically a contractor/sub-contractor
relationship.
- Charles Simonyi joined Microsoft in 1979. Specialisations included
developing software applications (e.g. operating systems, languages,
applications and utilities) and his dissertation on the metaprogrammer.
- Bill Gates assumed the hierarchy outlined in Simonyi's dissertation,
making himself the central metaprogrammer. He surrounded himself
with technical leaders called architects, who were each given overall
responsibility for an area of software development. Each architect was
surrounded by program managers who in turn had programmers who
reported to them. It was the programmers who wrote the code, so in reality
the program managers were acting as metaprogrammers too.
Thus ran the 'software factory'.
- Metaprogramming was a failure. Programming is iterative and with a flow
of information filtering from the top downwards only the improvements were
never made.
- MS was somewhat of a cult, hiring the young and experienced, so as to
keep Bill Gates on top as the metametaprogrammer!
- Charles Simonyi started the Microsoft Applications Group in '79
and started on creating a spreadsheet to rival Visicalc.
- IBM annual sales,$60 billion. 380,000 workers. "Conservative, a little
dull, slow to change, adapt at their own pace, but if pushed will fight
dirty". "The price of prosperity and power is compliance with the rules and
the pace of IBM". Very good at corporate infighting.
- Own corporate language- Minicomputers are 'mid range systems', monitors
are 'display units', hard drives are 'fixed disks' etc.
- IBM is top-heavy with managers ("class of executives with bovine, cud-
chewing dispositions, who think only on demand").
- 1980, Bill Lowe (Entry Systems Division) vowed to make the IBM PC in
one year - a miniscule amount of time for IBM. ("Project Chess")
- IBM normally always chose to build their own components, but Lowe
made the radical decision to buy parts in.
Intended to use Gary Kildall's (Digital Research) CP/M as operating
System (see below).
- IBM Bought a Microsoft BASIC language.
- IBM had a non-disclosure agreement so any company they spoke to
were banned from discussing IBMs plans, whereas IBM could take any
plans or secrets they were made privilege to.
- Bill Gates urged IBM to go with a 16-bit processor (high-tech at the time).
IBM chose the Intel 8088 (the least attractive from the tech standpoint).
- This processor was slower than it's competitors but this was a bonus to
IBM as it meant their own PC wouldn't compete with existing mid-range
Systems and wouldn't kill existing terminal business.
- Bill Gates bought QDOS (quick and dirty operating system) from
Seattle Computer Products and renamed and launched it as MS-DOS
as a direct competitor to Gary Kildall's CP/M86 that was being developed
for the new Intel processors.
- Eventually Microsoft and IBM signed a co-development agreement,
meaning that Microsoft would provide all software for Acorn.
- Late 1980 - Bill Lowe promoted to vice-president of another division.
Don Estridge took on the Acorn IBM PC project (he too would later
be promoted when his working practices were too unorthodox and
swift for conventional old IBM).
- In order to meet price point, profit margin and delivery schedule the
machine was shipped without serial or parallel ports, and no graphics
capability. The extra circuit cards on the motherboard which allowed
the machine to run at 64k were omitted. It ran at 16k.
- Data Buses - Apple II used Steve Wozniak's bus. Most CP/M
machines used S-100 bus. IBM invented another and published
its specification so other companies could design cards to go in IBM-PCs
- This was a shrewd move in a way - IBM were behind schedule and
over budget so this meant that other companies would fund the
development and marketing of add-on cards.
- Acorn wasn't a very proprietary machine - every component came from
outside with the exception of the ROM-BIOS. IBM believed that, whilst
other companies could buy the ROM-BIOS chips, they would never get
them at the same bulk-buy discounts that IBM had achieved, hence these
companies could never compete in the PC market.
- Acorn was positioned against the Apple II and looked bulkier but had
more potential for upgrade and more of an office feel.
- Applications included Visicalc and Microsoft Multiplan (spreadsheet) as
well as the word processor, Easywriter. However, there was still no
compelling application.
- Mitch Kapor - designed
Lotus 1-2-3. He had previously worked on
Visi-trend & Visi-plot - two Visicalc add-ons sold to Visicorp in '82.
With the money he made, Kapor invested in IBM and PC-DOS
- Lotus Development Corp. bet almost $4million on IBM in the idea that
Lotus 1-2-3 would become the IBM-PC's compelling application. It did.
(Released 26/1/1983)
- Picasso "Good artists copy, Great artists steal"- same with computer
industry. Most computers and software works the same way.
- In software there are many ways to solve the same problem. Spreadsheets
arrived in the forms of Visicalc, T/Maker & Framework each treating
the problems differently (see p145)
- When developing software companies greeted new generations of
hardware by porting their product to the new environment
- In both hardware and software, successful reinvention takes place along
the edges of established markets. It's not usually enough to make another
computer or program like the others; the new product needs to be superior
in at least one respect. Cheaper/more powerful/smaller/more features.
- Dan Bricklin (visicalc, ch4) didn't patent the idea of the spreadsheet. US
government had legal problems with patenting software and only allowed
copyrighting for years. Visicalc was copyrighted. The first patent for
software was granted in 1981 for Swiftanswer, which failed miserably.
- Visicalc, under Dan Fylstra's Personal Software, now Visicorp was
not profitable as a 37.5% royalty on every copy sold had to be paid to
Dan Bricklin's Software Arts. Today's royalties are usually 15%
- Lotus 1-2-3 comprised a spreadsheet by Johnathan Sachs, with a
database manager and graphics ability. It was only for the PC platform.
Advertising was directed away from large firms and went straight into
popular mainstream magazines like 'Time'. Lotus sold $53 million worth
of 1-2-3 in its first year, beating its target by 1700%.
- Such success puts pressure on the company to follow through with equally
impressive successes in the future, but next Lotus products,
Symphony
(word processor & communications) wasn't a bit success. Neither was
Jazz (integrated package for Mac) which flopped.
- 1985 - Dan Bricklin sold Software Arts to Mitch Kapor, hence Visicalc
died a death. "Now there was no first spreadsheet, only the best one".
<
- We live in a world of dollars, not a world of sense. Consumers typically
pay more attention to what everyone else is buying than to whether or
not what they're buying is any good. In this way, the most successful
products become standards against which all other products are measured.
- In any major market there are always two standards, with one of those
standards always being dominant (85% against 15% market split). In
this example we refer to PC/Apple Mac
- IBM made it's mark by going for 16-bit processors (Apple II, CP/M boxes
etc. all had 8-bit) and PC-DOS.
- There were many competitors for the 85% market share. All imitated
the IBM-PC but with drawbacks. The Hewlett-Packard HP150 had a touch
screen etc. but soon established that no-one wanted or needed this
technology.
- Microsoft encouraged these hardware companies to emulate and compete
with IBM as they were concurrently in the process of programming a series
of applications designed to be ported between IBM compatible machines
(Multiword (later Word), Multiplan etc.). They ended up putting MS-DOS
on virtually every machine that came out.
- IBM Compatibility soon became the key to success.
- Bill Gates backed the Tandy Model 2000 in an effort to find hardware
powerful enough to run Windows (1983). The Model 2000 flopped, but
Windows was eventually released in 1985. (An operating system not
unlike the early graphical user interfaces of the Xerox Alto)
- IBM's PC market share peaked at 40% in the early 80's. They were
convinced that competitors would never be able to make copies of the
IBM-PC at a lower cost, therefore were unconcerned about publishing
their specs. They still had the upper hand because of their unique
ROM-BIOS (Gary Kildall) being copyrighted.
- Reverse engineering broke the secrets of the IBM ROM-BIOS - copying
a technical function without copying the legally protected manner in which
that function was accomplished. Is somewhat a matter of guesswork, and is
a long and tedious process but eventually the IBM ROM-BIOS was cloned
by Compaq.
- Compaq's founders Rod Canion, Jim Harris and Bill Murto were all
experienced managers from Texas Instruments. They set out to make a
100% IBM Compatible PC with plenty of software available from the outset.
They managed to retail this at $800 less than IBM but still made a profit
due to lower overheads.
- Created the Compaq Portable - a kind of grandfather to the laptop- and
appealed to dealers by allowing a 3% bigger markup than IBM.
- By riding IBM's tail while being even better than IBM, Compaq sold 47,000
computers worth $111 million in its first year - a start-up record.
- Before long Phoenix Technologies started mass-producing clone IBM
ROM-BIOS chips. Many new PC companies resultantly developed, for
example Dell, whose founder
Michael Dell started his career selling
IBM-Compatible PCs from his bedroom!
- Commoditization (see p175)
- 1984 - IBM introduced the PC-AT
which used an Intel 80286 (286)
processor. These machines had 20Mb hard drives and were expensive
for the consumer. Also, IBM were so keen to beat Compaq that they
shipped the AT without adequate testing and soon realised faults in the
drive controllers.
- The Intel chip was never intended to go in a PC, but instead in a
Unix
based OS called Genix by ATandT. It was supposed to work in two modes:-
real mode and protected mode, the latter of which made the machine
very powerful with large amounts of memory accessible. It loaded in real
mode then was shifted down to protected mode. The problem was that
the only way to get back to real mode was to turn the computer off and back
on again! This meant that you couldn't swap between real mode and
protected mode applications.
- IBM lowered PC prices then introduced the AT in an attempt to curb the
influx of cheap clones on the market, but within six months Compaq had
cloned the AT.
- Gordon Campbell - head of marketing at Intel. Was very interested in
EEPROM (Electrically erasable programmable read-only memory) and
Started SEEQ technologies to develop the chip. Ended up being sacked
by the lead venture capital firm, then started a new firm, Chips and
Technologies. Created five chips that could replace the sixty-three in the
IBM PC-AT, meaning that companies could now buy 100% IBM
compatibility in twenty-four chips. This brought down the costs of clones
and ruined IBMs plans.
- Steve Jobs- Co-founder of Apple and founder of NeXT Inc. Concerned
with setting a style for computing. Assumed the role of purchasing agent
and salesperson to Steve Wozniak's technical role.
- 1981, Wozniak has accident and leaves Jobs in charge for 2 years.
Mike Scott was forced out of the presidential role, Mike Markkula took
this role on, leaving the chairman's role open for Jobs.
- Jobs had an ambition to develop a complex system with graphical user
interface and mouse, nicknamed Lisa (either 'large integrated software
architecture' or 'local integrated software architecture') - an idea he
picked up from touring Xerox PARC in '79. The idea was way ahead of
its time - having a 1 million character memory and the ability to multitask.
- Aware of the compelling application problem, Apple wrote their own
software to accompany the Lisa. The machine flopped in 1983.
This was due to the fact that it was priced at $10,000 - more of a
business price than a home-user price. The problem was that Apple
hadn't included any networking capability, hence making the machine
useless for many businesses. Markkula left when Lisa was released.
John Sculley (previously from Pepsi Co.) was employed as president.
- Jobs was crushed when the Lisa project was given to
John Couch
(ex Hewlett-Packard employee). This made him decide to join the
'other side' and work with Raskin's
Macintosh group.
- Andy Hertzfeld wrote the Mac user interface.
- The original Mac had 128k, with a capability to extend to 512k, although
Jobs wasn't aware of this.
- The Mac was designed to be a cheap home-use machine, but due to
high development costs Jobs announced that it would be a business
machine two months before its introduction.
- 1985 - after spending hundreds of millions of dollars to achieve his dream,
Jobs is relegated to a one-office building by Markkula and Sculley. He
left Apple soon after. Went on to found NeXT Inc. Rich Page & Bud
Tribble also left Apple in '85 to found NeXT Inc.
- In 1985, Sculley firmed up Apple budgeting and made cutbacks and
more economical choices than before in order to increase profitability.
- Jean-Louis Gassee moved from Apple France to head Apple USA and
went straight for the high-spec business end of the market, thus ruining
Apple's chances of hitting the home market with any real impact.
- 1988 - Allen Loren (Apple head of Management Information Systems)
increased prices, and caused the bottom to fall out of the Apple market.
- Late 1990 - much to the marketing department's amazement, Apple sell
extraordinary quantities of computers although profits still fell (p205).
Apple still plagued by constant internal reorganisations.
- John Warnock - founder
of Adobe Systems - Man who saved the Mac.
Started off on graphics program of ship sailing New York Harbour (Uni. Of
Utah Computer Centre)
- Established that storage space could be saved on the above project by
describing the whole harbour rather than by picturing it in its entirety - e.g.
generic area of water, generic tugboat.
- At Xerox PARC John Warnock wrote JaM
(John and Martin) - a language
used for describing objects and positioning them into a three-dimensional
database. This language evolved into Interpress
(described words and
pictures to Xerox laser printers), which then in turn evolved into
Postscript.
- Postscript treats letters of the alphabet just as it would pictures. They're
described as mathematical formulas rather than bitmaps. Each letter is
saved as a formula of lines and Bezier curves.
As these outline fonts are
resolution independent they can be scaled up or down.
- Before outline fonts can be printed they need to be rasterized - a
description of which bits to print where needs to be generated.
- After Xerox PARC refused to release the above products commercially,
Warnock and Geschke left PARC and went on to become sellers of
graphics systems software and postscript fonts - Adobe Systems.
They went on to write printer controllers that could interpret postscript
Demands, rasterize then print.
- At this point, Apple still had very disappointing Macintosh sales and
no compelling application. Mike Boich, Alain Rossmann and Guy
Kawasaki evangelised about the Mac and its graphical user interface
to the point that soon over 350 Mac applications were in development.
- VisiCorp created VisiON
- a slow and cumbersome graphical user
interface which scared Bill Gates and Charles Simonyi into writing
Microsoft Windows
- a graphical user interface to use on top of DOS.
Shipped in 1985.
- Gary Kildall's Digital Research quite unsuccessfully tried to compete with
their own version, GEM, in 1983.
- 1984 - Steve Jobs bought 15% of Adobe - not a wise move at the time
but later made Apple $10 billion.
- Apple released the Laserwriter printer ($7000, no network sharing)...then
invented Appletalk which allowed up to 32 Macs to share a printer. The
Laserwriter was for years the only non-Adobe postscript controller on the
market. Steve Jobs was so anti-network that even at NeXT Inc. he for at
least 12 months turned down orders that were to be used on a network.
- Adobe follows Bob Taylor's Xerox PARC managerial model. In 1989 the
company had the highest profit margins and the greatest sales per
employee of any major computer hardware or software company.
- Adobe Illustrator caused problems by being expensive to develop, not
revised soon enough after it was shipped, bugs not fixed promptly
- 1985 - Apple decide to opt out of using Adobe postscript and form
tactical alliance with Microsoft to create own technology. Adobe instantly
struck back with ATM (Adobe Type Manager) which was designed for
the Mac but sold without any Apple involvement or intervention.
Apple still sold all Adobe shares.
- Adobe (Warnock) entered into direct conversations with IBM and worried
Microsoft.
- 1991 - Adobe and Apple agree to work together again because Apple's
print technology failed them.
- Success is generally based upon appearance rather than reality. With
high-tech start-ups there is no system or rules, just the product.
- Start-ups: 9/10 fail, although some put the figure higher at 19/20.
Reasons are normally stupidity, bad luck or greed, although they could
also be found in the transition between a group of men with ideas sitting
in an office, to a company with customers, overhead, structures etc.
- Stupidity - creating a non state of the art or overly imitative product
- Bad Luck- funding dries up, recession, someone pips you to the post
with a better product.
- Greed - Taking venture capitalists cash without thinking about the
product first!
- Do a product you want to do, not one that the venture capitalists want
you to do. Don't take venture funding too soon. Rent don't buy. Divide
the wealth so as to end up with less shares but more money.
- Startups usually initially have nothing but ideas, personalities and equity.
Shares are given to workers meaning if the product does sell, the wealth
is shared.
- The biggest advantage of the start-up is speed. Work is hard, fast and
cheap, sometimes however, with a low level of professionalism.
Good team working is needed.
- Ron Crane was working on the Ethernet card for Bob Metcalfe's 3Com
Inc. when he got bored and started fiddling with roof tiles instead.
Eventually he got bored of the roof and made 3Com a $400million company.
- Three waves of staff -
- first wave are commandos - do the initial dirty work
but can't imagine running a mature company.
- Second wave are generally managers who gather together a group of commandos and keep them
motivated for the small amount of time required.
- The third wave managers tend to fail in the start-up market as they can't keep their commandos.
- In many start-ups, things go brilliantly for quite some time then half the
founders leave at once. This is because they probably stayed long enough
for their stock options to mature or for the company to go public. These
first and second wave staff members who are dying to leave by the time
their stock options vest, are generally replaced by third-wavers who favour
stability.
- OEM - Original Equipment Manufacturer - selling software product to
hardware companies who ship it out packed in with the computer - often
under other names.
- Shareware
- small but profitable companies who literally give their software
away. Invented by Andrew Fluegelman and Jim Button who wrote PC
File, purely because they owned IBM PCs and needed a database.
- The programme was distributed between friends who often called Button
to ask technical questions. They could register and get upgrades for $10.
He started Buttonware and increase PC File price to $25. Growth flattened
when the price reached $89.
- Andrew Fluegelman created PC-Talk as freeware.
This program allowed
PCs to emulate terminals and link to mainframes over telephone lines.
- The beauty in shareware is no packaging, marketing, sales or printing
costs. The majority of products reach 20-30% of potential market then
the market flattens. At this point the creators can move away from the
product or begin to commercially market it to a different audience.
- " For computer companies the cost of growth is innocence"
Companies grow and in turn require more 'suits', which either kills
the company, makes it bigger, or changes it and its leaders forever.
It's hard to maintain the equilibrium between technology and business.
-
Jim Manzi, hired as CEO at Lotus. Took against Bill Gates and
mistakenly advised staff not to develop 1-2-3 for Windows. Staff
ignored him and the product was a success.
- Manzi constantly reorganised the company, creating large groups
which slowed product development down and created communications
problems.
- 1987 - Lotus going through a bleak period. Frank King (IBM and
creator of SQL database language which was developed in secret)
finally set up five offices at Lotus and spent a day per week in each.
He discovered the hidden products that staff had been covering up.
- Aldus - Power tends to go to those who can sway the people in power -
often meaning that articulate marketers get everywhere and the developers
get nowhere. Such was the case with Aldus.
- Paul Brainerd from Atex
(makers of publishing systems) and five
engineers went to work on half salary, then were diddled by Brainerd
into taking only 2% of the company each, to his 90%.
- New product niches tend to need big partners, hence Aldus and Apple
collaborated to release Pagemaker for the Mac. Brainerd was a very
rich man by the end of the 1980's. He took Aldus public, but remained
tied to the company (p260)
- Companies usually go public to raise capital, but in the case of computer
companies they generally have enough capital. These companies
generally go public to make real the wealth of their founders. Stock
options are useless until the stocks are publicly traded.
- Strangely, some large companies such as Microsoft, who have massive
reserves, are actually losing money by having so much of it. Putting the
money into savings and trusts etc. returns 12-15% whereas if it were
ploughed back into the companies themselves, the return would be more
like 40-50%.
- Wordperfect Corp. is still private and has three stockholders.
- Two kinds of software companies:-
- Developing new concepts and pioneering new product areas
- Continuing evolution of existing product.
- Aldus was originally a type A company, but became a type B. Now it
constantly updates Pagemaker but buys its other software in from
elsewhere and pays the authors royalties.
- 1988- Aldus had been putting it's efforts into Flintstone, a word processor
but was pipped to the post by Word for Windows (Microsoft)
- Danger of technology firms with good marketing becoming marketing firms
with okay technology. This happened to Ashton-Tate (creators of
dBase II
- a successful database programme) after its' CEO,
George Tate, died.
The new CEO, Ed Esber, tipped the balance too far in the direction of
marketing and lost time and the edge on technology. Dbase IV shipped
two years late and was a shoddy job.
- 1983 - Microsoft hired John Shirley from Tandy Corporation to build a
business structure based on Gates' interpretation of Simonyi's model.
It aimed to let the company grow without diverting the focus from
software development. Shirley hired the company's first chief financial
officer, VP Manufacturing, VP Human Resources, head of
management information systems and head of investor relations all in one
person;- Frank Gaudette.
- MS set about world domination with branches in Japan, UK, France and
Germany.
- IBMs success on the PC market was a fluke. IBM PC then PC-XT,
followed by the five times more powerful
PC-XT. They set the standard
for personal computing, then lost control of the business and industry in '84.
- IBM couldn't maintain a swift development schedule (following Moore's
Law) for making new machines. Theory would have it that a new system
entirely could be created every eighteen months, but IBM could not
deliver any faster than every three years. This was to be their downfall.
- Contrastingly, Compaq cloned the IBM PC and followed this with a
IBM PC-AT clone six months later. They couldn't wait for IBM to
develop a new machine so beat IBM to it by releasing the Deskpro 386,
which IBM itself later cloned!
- IBM market control techniques:-
- Announce direction not product - puts competitors off the scent.
- Announce a real product, just do it way in advance - this disrupts
the market for competitive products that are already shipping.
- Don't announce a product at all, just give strategic hints -
Keep the competitors guessing and keep them on edge.
- Don't support anyone else's standards. Make your own-
e.g. using PC-DOS not CP/M,
Token Ring not Ethernet (p276)
- Announce a product, then say you don't really mean it! -
IBM released networking software and hardware 1985, saying it wasn't
part of an IBM direction, but was a precursor to a forthcoming hardware
technology called the Token Ring that would be available in the future.
The future, being 1990. IBM lost the networking business.
- Novell Inc. (the networking leaders), based in Provo, Utah, who used to
make CP/M systems, hired in Drew Major, Dale Neibaur and Kyle Powell
(collectively hired out as Superset) in 1981. They added the ability to
lock directories to CP/M, creating file servers where people could share
data. It was Superset who decided to network the IBM PC. During the
time they were doing this, Novell had gone bust and had time to rebuild
itself.
- Ray Noorda was employed as new Novell boss, and asked Microsoft
in 1983, to write him some networking software (written by Steve Ballmer).
All new technology was conceived as an extension to DOS. Microsoft's
MS-Net was released in 1985, but was nowhere as fast as Novell's own
Netware because of the constrictions of writing over DOS.
- Superset ported 8086 software to the 80286 in a week and hence won
the networking war.
- IBM deliberated too long before releasing new PC products and only in
1987 was the 80386-based machine shipped. This had a 3.5" floppy
disk drive and PS2s. It was shipped too soon, still full of faults.
As in previous cases, Compaq fought back with a similar 32-bit
bus standard in competition.
- In 1987 IBM also announced a new operating system,
OS/2, which
was a true multitasking entity. Microsoft advised that IBM should
direct OS/2 straight at the 386 market, but being IBM, they didn't
want to compete with their own minicomputer business and made
OS/2 for the 286 (Bill Gates quoted this as being the "brain
damaged chip!").
- 1989 - IBM announced Officevision.
An effort at fourth-wave computing
(first wave=mainframe, second wave=minicomputer, third wave=PC).
The 80386 with its' PS/2 & OS/2 and this software was supposed to lead
the fourth wave - Network Computing.
Officevision still hadn't appeared by 1991.
- All the above technologies failed, OS/2 partly because of the escalation
of memory prices after a push from the Japanese chip market.
- Bill Gates decided that OS/2 was bad, and DOS was good around '89.
He put a new face on DOS in the form of Windows 3 in 1990. This
was a massive success, selling 3 million copies in its' first year.
- Bill Gates speculated at a meal with Lotus that IBM would fold within the
next decade. This was leaked, and so ended the relationship between
Microsoft and IBM.
- Marketing is the stimulation of long-term demand by solving customer
problems. In PC industry the same problems have been being solved
for 10-15 years. Whilst hardware is getting faster and software more
sophisticated, the only major advances over the course of the book
are the GUI, multitasking OS, postscript printing and LAN technology.
- The lack of radical development stems from the major companies
ability to introduce more and more people to existing products. They
are also limited potentially by the hardware performance required to
develop radically new software.
- 4 MAJOR TRENDS:
- Standards-based computing
Computers built with the idea of working with computers and software
made by their competitors. Idea came from Sun Microsystems (see p300)
who also gave away much source code and software.
Languages, OS, networking and windowing systems as standard.
- RISC processors
Reduced Instruction Set Computing. Idea from IBM.
Fast devices, which were cheaper to build, as they were smaller and fit on
one piece of silicon. Clock speeds can be faster as there are greater yields
due to fewer transistors. Simpler and more robust.
- Advanced semiconductors
- Death of the mainframe
Millennium bug/human oversight. Mainframe source codes hardly exist
any more. Post-millennium new systems put in place which would work
just as well on a desktop workstation. Also, data can be safer through
being shared over networks etc.
- Moore's Law has seen real semiconductor performance gains of late
hence massive increases in processor speed. As the law increases speed
faster than people can think what to do with the increase of speed, it is
likely that processors will get smaller (which in turn means faster) and will
have other amalgamated circuits which would have previously been on
separate plug-in cards (e.g. floppy disk controllers, hard disk controllers
Ethernet adapters etc.) Hard disk drives will be replaced by memory chips,
and then those too will be incorporated in the processor.
- Object orientation
- large programs built as chunks of code called objects
that can be mixed and matched to create many different kinds of application.
- The future is in finding a new business model providing current levels of
reward without the old level of risk. Something to allow entrepreneurs to
'cash-out' without taking their companies public. Software studios (p314)
can fit these criteria.
- Predictions have been proved correct. Hardware less important, software
and semiconductors thriving (see Bill Gates example p319).
- The rest of the PC software industry has been consolidating and
restructuring, finding alliances & mergers etc. to compete with the 'menace
of Microsoft'.
- Process of maturation taking place. No. of software companies reducing.
- Windows is the clear standard. IBMs market share is under 10%.
- Apple moved from selling few expensive machines to selling more cheaper
machines and increased their popularity and market share in this way.
- People who are astute and technically aware can see waves of technology
leaving research labs at least a decade before becoming commercially
viable. Success is merely a case of getting on the wave at the right time.
- It is better to get off a wave too early (provided another wave is in sight)
than to ride too long. Huge losses are made through sticking with
an idea which may not pay off.
- Microsoft move onto the next wave whilst taking the old waves along (e.g.
through making DOS crucial and included in Windows).
- Client-server computing the norm now.
LANs have centralised data but the
real work is done on desktop machines. Single sets of data are accessible
to potentially a whole company etc. - Corporate information base.
- SQL (Structured Query Language) is lingua franca of client-server computing. Back end databases etc.
- Back end - SQL running database
Middleware sits between client application and back-end database. Translates
Front End - e.g. Microsoft Excel
- Open computing (standards-based computing) is at the heart of client-server.
Hardware and software from different manufacturers can be used together etc. allowing for synergy and economies of scale.
- The next developmental step from client-server is through the Internet.
virtual network of 'borrowed pieces of telephone network' exchanging
packets of data. Not hierarchical and not centrally managed.
- Internet developed from academic experiment from the 1960s (ARPA
)
- Internet Engineering Task Force (IETF) exists for technical definition,
evolving Requests for Comments (RFCs).
Protocols are sets of rules approved by IETF defining what various types
of data look like and how they will be handled.
Fundamental protocol is TCP/IP (Transmission Control Protocol/Internet P.)
- No strict internet topology. Isn't mapped. Not a business.
- Data security as concern - vulnerability from invasion by unauthorised
users and viruses/worms and the eavesdropping to which internet
data transmissions are vulnerable. Encryption is a growing market.
Firewalls
(programs that filter incoming packets) are a means to conquer
these vulnerabilities.
- Intel are accelerating Moore's Law, running several development groups
in parallel.
- Digital convergence of communications and computing & explosion of
bandwidth. More voice, music, and graphics uses. Computing via TV.
Last updated on Friday, 31-Oct-2003 08:59:19 GMT
|
