[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Strategic Orientation - Scope, Scale, or Means of Monopoly
John Robert Behrman <jbehrman@netropolis.net> wrote --
How MS got so big and how it stays there is more interesting
and actionable than billg's Tom Swift fantasies. Moreover,
its core monopoly is MS_DOS, which is to say whatever it has
worked on top or underneath of PC_DOS, namely Win98 and a
few real-mode drivers for video, sound, and ethernet cards.
I think it is best to make the case against MS narrowly.
That vastly simplifies, (i) proof, (ii) relief, (iii)
remedies or (iv) corrections, as Texans call punishment.
Proof -- evidence -- indication -- call it what you will -- that His
Imperial Highness manipulated the software and hardware industries from
the early eighties on fairly *screams* at us from the command sets of
nearly all the big-name programs sold on the general market.
You want narrow, I got narrow. My focus is so narrow that I miss a great
deal -- but a narrowly focused beam penetrates deeply, and given the fact
that on first examining Windows in 1986 -- when I had high hopes for GUIs,
trusted Microsoft implicitly, and didn't know Bill Gates from Joe Blow --
I saw that it wrested control of our machines from us and that it was
wrong, wrong, WRONG, maybe what I have to say is worth listening to.
(If the software is not indisputably our property, the hardware *is* --
and control of how *our* property responds to *our* actions must be
returned to *our* hands.)
Yes, it's the keyboards and keystrokes business. OK, maybe you'd just as
soon do without a keyboard; or 104 keys and counting is fine by you; or
you figure that everyone uses a mouse now so it doesn't matter. It
matters. If it didn't matter, the big players in this industry would not
have bent over backwards to keep us from understanding and using certain
kinds of keystrokes. By so doing -- by acting as though certain resources
simply weren't there -- they made both their own work and ours harder. If
it didn't matter, there wouldn't still be people beseeching His Imperial
Highness to provide a way, even a difficult way that makes us finish the
job ourselves, to make our keyboards work as some of us know they can.
Below, unedited, is a post to alt.folklore.computers in which I present my
observations. Some people might find it "technical" in places. It was
intended for a highly technical audience. An audience that is quick to
jump on false assertions. And none has jumped on anything in this yet.
But first, an excerpt from a March 29, 1994 column by _The Age_ technology
editor Charles Wright. I figure there's both exaggeration and truth here.
Even if there is only a tiny grain of truth, it fits in very well with my
observations:
....WordPerfect's founders, Bruce Bastian and Alan Ashton... didn't
want their rapidly growing company to be messed up by anything like
professional managers, so they brought in Bastian's brother-in-law,
Pete Petersen, as office manager.... Most of the senior staff seemed to
have been hired on the basis of personality, rather than qualifications
or experience. And because the owners didn't like giving the employees
stock, like that Bill Gates character up there in Redmond, Washington,
people like Petersen got to earn as much as $US5million a year.
<http://werple.mira.net.au/~jromney/TheAge/cw290394.htm>
And now, as promised, warts and all....
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
URL: news:6h5u0f$2nd@news.seed.net.tw
Date: 16 Apr 1998 21:42:39 GMT
Subject: Re: What does Micro$oft do Well ?
Newsgroup: alt.folklore.computers
This goes back a ways. No rest for the weary, etc., etc. Apologies.
Quoting me and responding, Kirk Israel <kisrael@cs.tufts.edu> posted in
<news:6fdr9q$fan$4@news3.tufts.edu> --
>: Then developers started pretending the Ctrl key wasn't there.
>
> Huh?
. . .
> What system are you using? I hardly ever use the mouse in Win95 [...]
I was probably logged onto a SunOS system when I read and downloaded the
message you quoted my response to. I may have transferred the message to
an MS-DOS box to write the response. I posted the response from a Linux
system. I'm required to use a W95 machine at work, but I use it for work
only. And that under protest.
The subject was keyboards and the Control key.
- The keyboard is the area, most often a rectangle, that has as its
boundaries the top-row numeral/symbol keys, the shift keys, and the
space bar. There are no keys outside that area, only buttons.
- Buttons and pointing devices make fine adjuncts to a computer system,
especially for children, casual users, artists, and the handicapped. For
able-bodied professionals, the keyboard is the primary input and control
device. You can have a full-function general-purpose computer system
without a pointing device; you cannot have one without a keyboard.
(To anyone who thinks that using the keyboard as your primary input and
control device inevitably means typing things like :s/^.*$/\U&/ all day,
or remembering what Alt-Shift-X does in ten different programs, the
kindest thing I can say is that you have been seriously misled.)
- You cannot have a full-function digital keyboard without a Control key.
For the sake of the many people who *must* use the keyboard for the bulk
of their work, the Control key must be located where it can be pressed
without bending a wrist or removing either hand from the middle row of
alphabetic keys.
- In combination with 32 of the character keys, the Control key produces,
directly or indirectly, 32 single-octet values defined as non-printing
control codes in every public encoding standard on the books. These are
the only command codes generated and recognized by all systems and
transmissible via all media. The standards bind four of these codes to
keyboard functions: Backspace, Tab, Carriage Return, and Escape. The
remaining 28 control codes are available for application designers (but
*not* operating environment designers) and users -- especially users
whose work is keyboard-intensive -- to assign as they see fit.
In the remainder of this article I will bow to conventional usage and
use the term _keyboard_ loosely. Where necessary for clarity, I will
refer to the (usually rectangular) area described above as the "main
block."
- The 28 assignable main-block Control-key combinations, together with the
character keystrokes and a semi-modal style of operation such as has
been the norm for mass-market commercial software since the late 1970s,
are sufficient for building a user interface, graphical or not, as easy
to learn as any mouse-based UI, yet far more efficient and ergonomical,
and completely platform-independent. (Oops; think I used a dirty word.)
(Need I add that such an interface would in no way rule out or interfere
with mouse-and-button-based operation?)
- Environments that must intercept certain control codes for serial-line
handshaking and other system functions are rare and easily dealt with.
Environments that intercept those codes for purely historical reasons
are less rare but even more easily dealt with. Any environment that
disables or ignores control codes originating at the keyboard, or limits
their use to "shortcuts," constitutes an attempt to undermine the most
basic standards of computing and prevent users from learning techniques
that they can easily transfer to standards-based environments.
Now, about developers and the Control key.... I take it your "Huh?" means
you really don't see it at all. That means I'll have to tell it as I see
it. All of it. Going back to 1982. I'll try to be brief. I said *try*....
Got an Apple II by accident. Typed translations in WordStar. Liked it.
Glanced at the technical manuals, couldn't understand 'em, put 'em away.
Got dragged to a computer club meeting. Stayed for the beer and the social
life. Saw that others could print italics in WordStar while I couldn't.
Got curious. Dragged out the manuals. Learned hex. Studied an ASCII table.
Realized that on micros and minis, main-block Control-key combinations are
as universal as character keystrokes. Saw that with decent key bindings
and on-screen help I should never have to move my hands from the main key
block, reach into a corner for a Backspace key, or look down at the
keyboard on any microcomputer in existence. Fell to floor. Wept with joy.
Became a technical writer. Still using CP/M, wrote manual for early
MS/PC DOS program. Noticed that Control-M produced an eighth-note symbol
instead of a Carriage Return code; found that *all* the control-range
keystrokes produced IBM-only graphic characters (you know, smiley faces
and boxes and triangles and stuff). Spoke to lead developer. Was told,
"Hmmm, that won't do. We'll fix it." It was fixed.
Was assigned to write manual for company's first Mac program. Found that
the Mac's keyboard (this was in the 512K days) had *no* Control key and
could not produce the full seven-bit ASCII range (some 32-bit system,
huh?). It had a "Command" key in the bottom left corner, and the Mac OS
used Command plus X, C, and V as "shortcuts" for Cut, Copy, and Paste.
Remember that.
Was assigned to write manual for company's first Windows program. Set up
Windows 1.0; checked out the six bundled applications. *Not one* used a
*single* main-block Control-key combination for *any* function *at all*.
The "shortcuts" for Cut, Copy, and Paste were Del, F2, and Ins. Remember
that.
Started using MS-DOS. Checked out the command interpreter and utilities.
*Not one* recognized *any* main-block Control-key combination other than
the few inherited from the CP/M command-line interpreter. This remained
the case for ten years.
Checked out MS/PC DOS manuals. Lots of hex notation -- but *not one*
hexadecimal ASCII table like the one that had been my Rosetta Stone for
understanding computer basics. Character tables were numbered in decimal
and showed IBM-only graphics (you know, smiley faces and boxes and
triangles and stuff) for the codes in the control range. No chance of an
interested beginner learning hex from those tables or finding out that
main-block Control-key commands derive from the same basic standards as
character keystrokes. ("But each Control-key command does something
different in each different program." Character keys don't? Try the
sequence Alt, E, A in WordPad and Paint. So much for consistent UIs.)
First x86 machine died; got new one. Typed `dir', tried to hit Control-M,
got `dirM'. Exercised maximum self-control. Found real keyboard, swapped.
LONG NOTE: Windows came out in August 1985. The "enhanced" PC AT
keyboard with Control in the bottom row appeared in April 1986. Its
Control key was in exactly the same place as the Mac's Command key. At
some point after that -- exactly when, I don't know -- Microsoft
started using Control-X, C, and V for Cut, Copy, and Paste in Windows.
I'm not saying IBM moved the Control key to let Microsoft copy Mac
keystrokes. I wouldn't discount the possibility, either. I *do* believe
the Control key was moved to make keyboard-only operation uncomfortable
and inefficient. I've met a *few* keyboarders who prefer a bottom-row
Control key; *not one* of them likes it where IBM put it in 1986.
Some people say IBM wanted to make the keyboard more typewriter-like.
By adding a second Alt and a second Control key? By using `Enter' where
all typewriters and most other computer keyboards have `Return'? People
had squawked about departures from the Selectric design in 1981; the
changes they'd asked for had been made on the original PC AT in 1984.
IBM released another system with Control in the bottom row in early
1986: the RT PC. A Unix machine. Does anyone think Unix folk were
clamoring for more typewriter-like keyboards? I think IBM was "testing
the waters" with the RT, and the market accepted it because (1) Control
isn't that important in vi or X, and (2) Emacs folk were already used
to a bottom-row modifier key in the form of Meta.
Back to the PC AT. IBM didn't just move the key. They changed the
interface circuitry and the scan codes as well. Was this necessary?
Joseph H. Allen writes in the PC Hardware FAQ --
....The keyboard didn't have to be redesigned -- there
were enough extra scan-codes for the AT's 101 key keyboard
and the repeat mechanism could simply have been moved to
the BIOS. But no, they had to redesign everything. Sigh.
No, it can't possibly be the case that someone wanted to make sure old
keyboards would be unusable with new machines and the old layout would
slowly but surely die. It just worked out that way by chance. Right.
Around this time the Mac got a Control key. In the home row, too -- but
not for long. Apple soon moved it, and they went IBM one better. From
<http://bigbang.stanford.edu/~parkb/computers/mac/capslock.html>:
Switching CapsLock and Control... seems to be a recurring question
on comp.sys.mac.system.... Apple does not provide a software
solution.... The first depression of the CapsLock key generates
only the KeyDown event... even without a mechanical lock.
...and from Apple Desktop Bus Q&As, Technote HW 505, "PowerBook Control
and Caps Lock keys can't be remapped," October 1990:
With the... PowerBook keyboards with nonmechanical Caps Lock
keys... no keyup sequence... is sent from the keyboard, which
makes it impossible to write an ADB handler that remaps the keys.
Also, there's no way to change the behavior of the keyboard
inside or outside the driver, because the keymap handling is
hardwired to the Power Manager chip.
Now *that's* what I call *thorough*. (End of long note.)
Was required to switch to WordPerfect. Opened up the 1000-page manual.
Under heading `Keys to Know' in chapter 1, section 1, found only the
subheadings `Backspace', `Cancel', `Delete', `Help', and `Num Lock'. Next
heading: `Function Keys'. No more keys to know, folks! Checked the index;
found *two* main-block Control-key combinations listed (a later edition
lists four). Set to tapping; found --
Ctrl-@ a b c d e f g h i j k l m n o p q r s t u v w x y z [ \ ] ^ _
. a . c d e f g . . . . . . n o p q r s t u . . . . . . \ ] . .
. . b . . . . . . i j . . m . . . . . . . . . . . . . . . . . _
@ . . . . . . . h . . k l . . . . . . . . . v w x y z [ . . ^ .
Inserted into text but treated as meaningless ...: 16; documented: 0
Inserted into text and actually used ............: 5; documented: 2
Treated as command keystrokes ...................: 11; documented: 3
Of the three documented main-block Control-key command keystrokes, two had
the same function, and one -- in the top row, but not impossible to hit by
accident -- *canceled* any custom keystroke definitions.
(Speaking of custom keystroke definitions, if you just *looked* at the
keystroke definition screen and hit the key to leave it you were asked,
"Save changes?" This did not happen in any other command screen.)
WordPerfect was created for Data General minis and adapted for IBM
PC/XT/AT, IBM PCjr, Apple IIgs, Amiga, Atari ST, Macintosh, Unix, and VMS
systems. Among the keyboards used with these systems, some have Alt, some
have Command and Option, some have Meta; some have Home and End, some have
Help, Do, Find, and Select; some have no function keys, some have
twenty.... Different names, positions, and codes.... Need I name the only
modifier key these keyboards all have in common besides Shift? Isn't it
just a tiny bit strange that the designers of a word processor -- a
program in which people *type* all day -- would avoid using a universal
main-block key, avoid even *mentioning* it, and put obstacles in the way
of anyone trying to use it?
Got on the company's LAN, tried the cc:Mail message editor. Found that the
control-range keystrokes *all* produced IBM-only graphic characters (you
know, smiley faces and boxes and triangles and stuff). Something that a
software project manager once agreed without question needed to be *fixed*
had become the *norm* for mass-market software.
Better wrap this up. Not long ago I asked some old-timers which company
started this business of using only buttons (platform-specific and
non-main-block keys, if you must) for program control. One said,
"WordPerfect started it!" Another said, "No, it was Lotus!" "No,
WordPerfect!" "No, Lotus!" Well, whoever started it, these two biggies of
the application software market in the second half of the eighties worked
very hard at keeping our hands off the main block (and, I should add, at
making on-screen keystroke help a pain to access). Is this because that's
what "the market" wanted? "The market" was new to computers, and had no
idea what was possible. (I recently told another technical writer that
software can bind any keystroke to any function at all. He was astounded.)
And the other biggie? I recently found a description of Microsoft Word for
DOS keystrokes on the Web. *No* mention of *any* main-block Control-key
combinations at all. Surprised?
....We were also fascinated by dedicated word processors from Wang....
That's why, when it came time to design the keyboard for the IBM PC,
we put the funny Wang character set into the machine -- you know,
smiley faces and boxes and triangles and stuff. -- Bill Gates
<http://www.pathfinder.com/fortune/magazine/1995/951002/cover.html>
Dan Strychalski dski at cameonet, cameo, com, tw
Apologies for the anti-spam devices and non-threading newsreader.
-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
By the way, don't expect people from WPCorp who know the score to come
forward....
....WordPerfect's founders, Bruce Bastian and Alan Ashton, are about
to pick up $US600million each -- and this only a year or so after they
took out $US180million in profits -- from a software product that we
have always regarded as, how shall we put this, imperfect....
<http://werple.mira.net.au/~jromney/TheAge/cw290394.htm>
...Insiders at Novell... are pulling the trigger in a big way.... The
biggest seller was Director Alan Ashton (former WordPerfect owner), who
got out of 795,000 shares [of 1,393,243 unloaded by eight executives in
January and March 1995, when the stock was in the $17-21 range, after
peaking at about $35 in 1993 and dropping to about $14 in the summer of
1994 -- DS] <http://www.cda.com/investnet/periscope/950503.html>
As far as I'm concerned, they don't have to. Computers *do* talk.
Dan Strychalski dski@cameonet.cameo.com.tw