by Rod McKenzie, MPCUG President

                       Rough sledding

  Networks and network operating systems have been absorbing a fair
  amount of my time lately.  There's a lot there for a user to
  learn and understand.  New and faster adapters are being
  introduced to speed up the transfer of information between
  hardware, and hence, between users.  As the applications become
  more graphically oriented, and object oriented, the network
  traffic will continue to rise. I've tried to be as well informed
  as possible because the future seems to be filled with workgroup
  computing applications.  MPCUG relies on a network to run the
  BBS.  Below the PC Board software that we see on the screen,
  there's a network that keeps the three PCs working together.

  I've discovered that starting a network of ones own can be a
  major frustration. It starts with the hardware and goes straight
  through to the software. Everybody's to blame for this mess,
  everyone, including me.  It happens because we enjoy an open
  architecture with the PC-compatible systems.  That means an
  adapter, such as a mouse, a modem, a network card, a scanner, or
  a CD-ROM drive, can make use of almost any combination of base
  address, interrupt, and DMA channel that it wants, provided the
  software that drives it knows were to find it.  But here's the
  kicker.  It shouldn't use any of the above addresses, interrupts,
  or DMA's if they are already in use by something else.  There's
  no easy way to tell, and over the life of the hardware and the
  software there can be many changes. It's left up to the user to
  keep it all straight.

  Maybe I'm missing something, but why shouldn't there be one
  unique interrupt for each address, and leave it at that?  I guess
  that the original specifications for the PC bus were wide open,
  and the designers were busy getting around the limits set by the
  8-bit bus that was available for adapters and peripherals.  Who
  had ever thought of local bus systems back in 1982?

  Careful buying and trying

  The best that a user can do is look for new hardware that offers
  lots of alternatives for settings, and better still, can be set
  through software. Unfortunately this is also not the low cost
  alternative.  It's important to keep track of the hardware that's
  already installed and where it "hooks on" to the system.  This
  may mean keeping track of jumper settings and knowing what the
  command line switches are that belong in the CONFIG.SYS and
  AUTOEXEC.BAT or other batch files.  I strongly recommend that
  anyone who plans to make significant changes to their system
  print out copies of these two mentioned files.  Better still,
  make a bootable floppy which will bring up everything just as the
  C: drive does it.  That's extra insurance against the slim
  possibility that something will go wrong.

  I'm not really the superstitious type.  I don't hang a rabbit's
  foot from my hard drive or throw salt over my shoulder if I find
  lost clusters with the CHKDSK/F command.  It's true that I have
  been known to say a little silent prayer when I start loading
  Windows applications that I live long enough to finish the job.
  This networking stuff won't go away and I'll learn to live with
  it, but there's a certain stress involved with change.  Terms
  like frustration do cross my mind.

  Computing in the middle of the 1990s will start to be more user
  friendly, cutting down on the minutia that users need to know to
  keep the hardware and software working together.  Some people are
  devoted to the endeavor of computing so much that they will cope
  with whatever comes along.  Not everybody is that way and the
  industry needs to back off the complexity to assure that more
  people will see the benefits of using a computer, rather than the
  hassles that get in the way.  I'm optimistic that it will happen,
  through standardization and smarter hardware, too.