[Top] [All Lists]

Re: [ontolog-forum] How not to write specifications

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Wed, 02 Jul 2008 11:46:22 -0400
Message-id: <486BA2CE.4070001@xxxxxxxxxxx>
Ed,    (01)

Let me start with the following point, because I completely agree
with you that security is of the utmost importance.  However, the
requirements of the RIAA are diametrically opposed to the security
that we agree is essential:    (02)

 > But I don't think it was the RIAA that created the security
 > issues that spawned Vista.  Think who is liable for everything
 > but the first $50 on credit card frauds, and who is deeply
 > concerned about penetration of databases of private information
 > that was acquired by law.    (03)

The major problem is that the RIAA is asking for an impossibility:
distribute their product widely, but don't let people copy it.
Inserting the forward and backward pointers to keep track of every
I/O access slows down the machine, and doubles the size of already
bloated software.    (04)

Microsoft could have produced a more secure version of XP in 2004.
Instead, they rewrote Vista from scratch in order to keep track
of every copy of the latest song by some pop tart.    (05)

 > The point I was making is that trying to support all of it, along
 > with upward compatibility with earlier underdesigns, and bad
 > ideas like "integrating" the browser into the operating system,
 > had a much bigger impact than the degree of "modularity" in the
 > software design.    (06)

The decision to move the GUI support into the kernel made every
point you mention more difficult to implement and maintain.    (07)

 > But unlike Microsoft, [Apple] didn't have 18 other companies making
 > new and wonderful display hardware and graphics accelerators and
 > enhanced sound systems that Dell and Sony and IBM and HP and ...
 > elected to plug into their hardware platforms.    (08)

The hardware and software technology for virtualizing I/O devices
and even CPUs was developed by IBM in 1966.  The CP/67 and VM/370
operating systems were rock-solid systems that could accommodate
any kind of real device and emulate any kind of virtual device.
But I'll admit that there were three major problems:    (09)

  1. Intel technology in the 1990s wasn't as advanced as IBM
     technology in the 1960s.    (010)

  2. The timing requirements of high-speed video put a premium on
     minimizing software layers between the device drivers and
     the application programs.    (011)

  3. Microsoft inherited a bunch of bad design decisions made by
     IBM Boca Raton (a notoriously retarded IBM lab), by the guy
     who designed QDOS (Quick-and-Dirty OS), and by themselves.    (012)

But the decision to destroy the hard interface between the GUI
and the kernel was a self-inflicted disaster that was totally
unnecessary.  For years, the gamers who wanted to squeeze every
nanosecond out of their machines continued to boot into DOS to
get low-level control of the devices.  And professionals got
Unix-based graphics workstations from SGI.  Windows NT 4.0
couldn't compete in those markets, and it was foolish to try.    (013)

 > And NT was not a part of the Windows 95 or Windows 97 or Windows 98
 > or Windows 2000 products.    (014)

Windows 2000 was based on NT; you may be thinking of Windows ME,
which was a badly done revision of W98.  In any case, the W9x
products were GUIs on top of a version of DOS, but with the old
Windows 3.11 features that broke the clean interface between
DOS and the GUI.    (015)

As I said in my previous note, the IBM-MSFT collaboration produced
a solid foundation for a clean 32-bit OS/2 with a clean GUI called
Presentation Manager.  IBM demonstrated that it was sufficient to
support concurrent operation of the old Windows 3.1 plus an object
oriented GUI that includes functionality beyond what Vista supports.    (016)

Microsoft could have adopted a similar strategy in the mid 1990s
while providing *better* upward compatibility and a smoother
migration path than the dual path of W9x and NT->2000->XP.    (017)

 > MVS was every bit as ugly and ungainly as Vista, and for many
 > of the same reasons.    (018)

I was never a fan of MVS, and developing it was a terrible
technical decision made for political reasons.  But the later
rewrites of MVS produced a solid, secure, and efficient system.
And as I said before, the IBM internal reasons were very different,
both technically and politically, from the Microsoft reasons.    (019)

John    (020)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (021)

<Prev in Thread] Current Thread [Next in Thread>