prbuycomputer

How To Buy A Computer
By Al Giovetti

In order to make an intelligent decision when buying a computer, the
consumer needs to know a little about the hardware components that make up
the computer device.  The same is true when you buy a stereo system, car or
any other complex device.  The more you know about it, the better.  If you
are buying a car and you don’t know the difference between bias ply versus
radial-ply tires and steel versus rayon belts, you will not have a chance
when discussing the price of the vehicle.

A computer has a few basic components: microprocessor, read-only memory
(ROM) chipsets, random-access memory (RAM), hard disk drive (HDD), floppy
disk drive (FDD), video card, sound or audio card, network interface card
(NIC), hard disk controller card, modem card, compact disk read-only memory
(CD-ROM), compact disk read-write drive (CD-RW), keyboard, mouse, and
digital versatile disk (DVD).  Some of the components that are optional when
buying a computer are the monitor and printer.  You will need a monitor and
printer to make the computer work, but the monitor and printer from your old
computer may work equally well with the new one if you purchased good,
sturdy models of these.

The performance of a computer is one of those things that is rarely more
than the sum of its parts.  The manufacturer of a computer may be important
only in how they stand behind the computer with their warranties when
something breaks.  If you choose to have a manufacturer custom configure
your computer, the use of brand names is very important.

Software manufacturers who make a special effort to make sure that their
software works with these hardware products recognize brand name components.
Also, when there are problems with software and hardware compatibility there
is a greater chance that a fix, patch or work-around will be made for these
brand name products.

  I highly recommend constructing your computer with brand name products to
avoid compatibility problems with future software purchases.  Some of the
better companies include Hewlett-Packard (printers), Sony (monitors), 3-Com
(NICs), U.S. Robotics (modems), ATI and 3Dfx (video cards), Sound Blaster
(audio cards), Intel (chipsets and microprocessors), Microsoft (keyboard and
mouse), and Seagate (HDDs).

The microprocessor, or central processing unit (CPU), is the brain of the
computer. And over the years has been thought of as the computer’s most
important component.  The brain determines how much information can be
processed simultaneously and how fast the information is processed.  The
microprocessor interprets instructions gathered from input devices, computes
the results, and transmits those results to output devices.

Input devices are those devices that input information and instructions into
the computer from sources outside the computer.  Typical examples of input
devices include the mouse and computer keyboard.  There are highly
specialized input devices that are often not seen on most computers, such as
the touch screen monitor that picks up your intentions when you point at an
object on the screen.

The microphone and camera are two input devices that are seeing more usage
on computers these days with the expansion of computer telephony.  You can
transmit pictures and sound over computer connections to almost any location
on our planet with very little communications delay.  Many grandparents are
using these devices to keep tabs on their children and grandchildren in
remote locations.  Young people are finding new uses for dating online with
these input devices.

Output devices are those devices that communicate to computer users the
information that comes out of a computer, which is often called output.
Output devices include monitors and printers that display or print out the
information so that it can be read.

Input-output devices are those that both input information into a computer
and also carry that information out of the computer once processed.
Input-output devices include modems, and NICs.  These devices carry
information from one computer to another back and forth over phone lines,
cable lines, and wireless networks.

A network is a set of conjoined computers that can share storage devices,
peripherals, and applications. Networks may be connected directly by cable
connection, or indirectly by telephone lines or satellites, and can be part
of a small-office system or a global web of numerous other networks.  A
local area network or LAN is a network that is contained within the same
office, building, or closely related buildings.  A wide area network or WAN
is a group of computers connected to each other with a certain geographical
area.  WANs usually require special communications hook ups to operate.

For a network to operate, it needs a server, which is a computer that holds
data used by the different computers on the network. Some of the benefits of
a network connection include the ability to share document files and
expensive equipment, such as laser printers. Networks can be connected using
different combinations of topologies, protocols, software, and hardware.

In pricing a computer, understand that the fastest and the largest are not
always necessary. Lower speeds and sizes that will do the job with no
appreciable difference, but can be purchased at significantly lower price.
That last 50MHz of processing speed can often cost you an unacceptable $300
or more (see Microprocessor Speeds below).  This is a disproportionately
large amount of money for such a small incremental amount of speed.

Unless you absolutely must have the biggest and fastest, you will save
significantly by backing off on speed and size a notch or two.  Whenever a
speed or size is mentioned, pay attention to the current fastest or largest
numbers so that you have an idea what the range and benchmark are of the
currently available technology.  The short and dirty of all of this is that
the faster the microprocessor, the faster the data and numbers are crunched
and the less time you are sitting there in front of the one-eyed monster,
waiting.


Microprocessor Speeds: By the time this article is published we may have
microprocessor speeds as high at one Gigahertz. The Gigahertz threshold is
one step up from the old Megahertz standard. Gigahertz (GHz) means one
billion cycles per second while Megahertz means one million cycles per
second.  A hertz measures the frequency of electric vibrations.  These
electric vibrations correspond to the simultaneous processing of millions of
different calculations or data.

Read-Only Memory (ROM) chipsets are silicon chips that are hard wired into
the motherboard of the computer.  These ROMs have hard coded instructions
that are non-labile.  Like the information on a standard CD-ROM disk, this
ROM memory can only be read.  The advantage to ROM chips is that the
information is permanent and always quickly available to the computer when
turned on.  The disadvantage is that the chips are difficult to upgrade to
newer versions of ROM.  Examples of ROM memory include complementary
metal-oxide semiconductor (CMOS), which contains system initialization
instructions, and basic input-output system (BIOS), which controls the
startup process of the computer and other basic functions controlling the
interchange of data into and out of the computer processor.

Random-Access Memory (RAM) comes usually in easily upgradeable sockets or
expansion slots.  RAM is labile, meaning that when you turn off the power to
the computer the memory is erased like a blackboard.  RAM is often referred
to as internal memory, since it is closer to the processor, being mounted on
an expansion board on the motherboard.  Disk-based memory is referred to as
external since it resides on hard, floppy, and optical disks mounted inside
the computer case, connected to the motherboard and central processor
through an interface card and a wire.  Since these components are slower
than RAM and further away in terms of access times and wires they are also
referred to as secondary storage.  When you turn the computer on the empty
RAM (internal memory) fills with information from the hard disk or CD-ROM
drive (external memory).

There are many types of RAM available today.  Much of the RAM is specialized
for a particular function, such as laptops, desktops, or cache memory.  A
RAM cache is a bank of RAM memory that allows temporary storage of data or
instructions so that the fastest transfer of data or processing can be
achieved.  RAM cache is a kind of staging area that gets information ready
to board the communications conduits for information and instructions.
Cache also stores in convenient and quickly accessible locations information
that will very likely be used again by the computer.

Recently the arrangement of memory changed from 30- to 72-pin single in-line
memory modules (SIMMS) to 144- to 168-pin dual in-line memory modules
(DIMMS).  The reason for this is that the data bus which transmits data from
one component to another on the computer motherboard just doubled its size
from 32- to 64-bits, requiring that the memory also double its size for more
convenience and compatibility.

Memory modules come in different sizes, configurations, features, and
speeds.  You need to know these features in order to determine
compatibilities between the motherboard and the memory components.  Make
sure that someone who has knowledge of these features performs RAM upgrades.
The complexity of these features puts them beyond the scope of this simple
article.

Dynamic RAM (DRAM) is a type of computer memory that employs a system of
transistors and capacitors to retain data. DRAM is slower and less
dependable than static RAM (SRAM), because the capacitors cannot maintain an
electrical charge and need to be refreshed every millisecond, but it is
cheaper, takes up less space, and uses less power.  Static RAM (SRAM) is a
type of semiconductor memory that preserves stored information as long as
there is enough power flow to keep the device running. Although SRAM can
only hold one-fourth of the data a DRAM chip holds, it is still effective
because it does not need refreshing. Static RAM is quicker and more
dependable, but also is more expensive, takes up more space, and uses more
power than DRAM. Due to these characteristics, SRAM is never used as the
main RAM storage facility, and is instead used for the more speed-critical
parts of the computer, such as cache memory. Synchronous DRAM or SDRAM
performs all the operations under the control of the system clock.  By
synchronizing operations, the completion of timed tasks can be better
organized, further speeding up the completion of multiple tasks.

RAM speed is measured in nanoseconds, or ns.  The fewer the nanoseconds are,
the faster the RAM.  RAM is also configured to communicate on the computer
BUS.  A BUS in a computer, like a bus in the street carries things from one
place to another.  The BUS in the computer is the electric connection
linking electronic devices such as the internal sections of a computer
(e.g., the CPU, expansion cards, and peripherals).  Therefore newer computer
SDRAM will require PC100 or PC133 SDRAM so that the RAM can communicate on
the newer 100MHz and 133MHz computer bus.

The Windows 2000 enterprise server requires 128 megabytes of RAM and
recommends 256 megagbytes of RAM.  As a result I am recommending that most
people get 128 megabytes as a minimum RAM on their systems and consider
upgrading to 256 megabytes if you intend to perform many programs
simultaneously with many Windows open on the desktop without crashing.

Every card on the computer can also contain RAM and its own microprocessor.
Sound, video, hard drive controllers, and other expansion cards often have
their own RAM memory.  The more RAM memory that the card has the better it
will operate.  Video cards with 64MB of video RAM are the top-of-the-line
right now, while a minimum of 8MB of video memory is recommended.

Expansion slots are the openings on the computer motherboard where
additional expansion cards for sound, video, network, modem, Internet, and
other connections are made.  Most modern motherboards have at least one
Accelerated Graphics Port (AGP) slot, although, if you wish, you may use a
dual monitor configuration, which is provided for by the latest Windows 98
expansion pack; hence, you may need two AGP slots.
PCI, or 32-bit peripheral component interconnect, have replaced the older 8-
and 16-bit ISA, or industry standard architecture, expansion slots.  You
will need a PCI slot for each component you add, including sound cards,
network interface cards, modem cards, drive controller cards, and other
expansion cards.  It is always a good idea to have a few extra PCI slots
open in case you want to add additional components down the road as your
needs expand.

Many computers today are coming with DVD-ROM drives to read the higher
capacity Digital Versatile Disks, which can hold 4.5 to 9 gigabytes of
information, much more than the old CD-ROM standard (650 megabytes of
information).  Right now most of the applications of the DVD are in the area
of full motion video and movies, but in the future more and more programs
will become packaged on DVD disks, avoiding the problem loading multiple
disks, since one DVD can hold the information of up to 14 CDs.

New on the horizon but still pricey are the DVD-RAM or Digital Versatile
Disk-Random Access Memory drives.   These drives are similar to the older
CD-RW or compact disk—read write, where you can store information on the
disk similar to a floppy drive.  Many of the disks are designed for one
usage, but other disks you can write and erase as many as ten times before
the disk is rendered useless.  The advantage is that, if taken care of,
these disks are more durable (last up to 30 years) than magnetic media and
hold an immense amount of information for their size.

Smaller accounting firms may wish to consider using these devices for backup
purposes.  Larger firms will find them still too small to hold all the data
needed for a backup and will have to continue to use the larger capacity
tape streaming cartridges.  Whichever you choose for backup, you would be
wise to choose a DVD drive to go with it since these devices will become
more and more useful in the future.

Hard disk drives.  With the increase in the bus sizes of computer
motherboards, new and exotic data transfer has increased the speed of
accessing data by both SCSI (small computer systems interface) and IDE
(integrated drive electronics) hard drives.  The newer IDE drives use 66MHz
Ultra ATA (advanced technology attachment) to double the older IDE speed of
33MHz for hard drives.  As of this moment, Ultra ATA drives require a
separate controller like many SCSI drives because many of the motherboards
do not support this protocol without the expansion card.

SCSI drives remain the fastest and the most expensive drives.  The reason is
the high dependability of SCSI drives.  Many times SCSI drives last longer
than IDE drives and require less maintenance.  The newer ATA drives are very
nice and function very well when compared to SCSI.

While you may not agree with all my recommendations, I believe our
differences relate to comfort level.  Over the years, brand name components
became more important to me.  Brand name components rarely cause trouble,
because the computer hardware and software companies extensively test the
brand names because they are the recognized leaders in sales and
performance.  When posed with a problem involving brand name components the
service and development personnel are more likely to support these
components than unknown ones.  Often they will be very anxious to provide
patches and “work-around” procedures to get the brand names to work with
their product, while no names often get orphaned.  My comfort level with
these components is much higher than those “also ran” brands.