VH1994-CompTech.DOC

Personal Computers 1994

Some Background & History

8 to 16 to 32 bit machines

            Better optics and processes permitted placing more and more semi conductor devices on a chip -- and to place wires closer and closer together on a mother board. Microprocessors now have millions of transistors on one chip and circuit board wiring is too closely spaced to repair by hand.

            The first PC’s (PC-XT 86/1 variety) used an 8 bit microprocessors and data bus path operating at 1 megherz.

            The next PC’s (AT 286/1 variety) used a 16 bit processors with 16 bit data path at 1 meghertz.

            The next PC (386/8,16,20 variety) used a 32 bit microprocessor with 16 bit (AT bus) data path at 8, 16, 20 meghertz options.

            The next PC (486/33,66,100 variety) use a 32 bit microprocessor with built in arithmetic with 33, 50, 66. 100 meghertz options. These work in 16 bit mode or with (2) 16 bit bus paths in 32 bit mode.

Clock Speed

            Clock speeds began at 1 meg hertz. Later 386 units went to 8, 16, 20 meg hertz. Then 486 units to 33 or 50 meg hertz mother board clocks. A clock doubling option for the microprocessor CPU (Central Processor Unit) increased speed to 66 and 100 meg hertz. The clock doubling is applicable to the microprocessor only and does not increase bus speed.

 

Binary bits, Bytes, Hexadecimal Numbers

            A digital computer is a maze of ON-OFF switches made up of transistor segments. The number system is BINARY, on or off “bits” = 0 or 1 “bit values”. A set of 8 bits was called a Byte.

            Think of a binary number as an odometer that can only count to “1”. Thus 0101 = 0005 in the “tens” system.

                        Binary=Decimal

                        0000=0000

                        0001=0001

                        0010=0002

                        0011=0003

                        0100=0004

                        0101=0005

            When displaying binary numbers on a monitor, it was found convenient to use a single column to display a set of 4 bits, half a byte = a nibbles worth. This led to the use of Hexadecimal numbers.

                        Binary=Hexadecimal=decimal

                        1001=9=9

                        1010=A=10

                        1011=B=11

 

                        1111=F=15

            Use of hex gives a 4:1 reduction in column space.

            Note: The use of “0” as a number, or place, became common practice. Device step nomenclature often begins with 0 rather than 1.

Microprocessor (CPU) Wiring Interface

            The 8 bit microprocessors used three sets of signal paths: 8 data bits, 16  address bits and X control bits. The 16 bit processors were the same except expanded to 16 data bits. The 32 bit processors can accommodate 32 bit address and data bits.

Memory & Card Addresses in Hexadecimal Numbers

            Engineers often use hexadecimal numbers as the means of defining a memory address. Input output (I/O) Cards to peripheral devices are assigned “memory addresses” for use by the microprocessor in sending data to or reading from a device. Presently a card inserted into a slot on the mother board includes user set-able switches so as to define the address for that card. The literature expresses these as a hexadecimal number. Each slot is like a seat on an airplane, the address must be set on the card and logged in the software to know where who is sitting. Future operating systems intend to read the mfgs defined address by an agreed upon code, thus eliminating conflicts. Such conflicts have caused many problems in the past.

IRQ -- Interrupt Request flag

            When installing a card it is sometimes necessary to define which of 1 to 15 Interrupt Requests are applicable to that card. The microprocessor logs an IRQ, finished an incomplete process, then services the request. Most peripheral equipment makes use of such and sometimes the user can create a conflict. Windows 3.1 software includes software called MSD = Microsoft Diagnostics which can be used to look at how the system had been configured, permitting you to select an unused IRQ or I/O (Input/Output) card address.

Memory Management

            Software is used to manage memory. DOS 6.2 provides memory manager software to help you, read the manual to understand what going on. Windows-95 is intended to solve some of the current difficulties.

            The original microprocessors would address 1 meg of memory address space. The first 640K is called lower or standard memory and the remainder is called the upper or high memory. DOS compatible software used the lower and the system used the upper for video and I/O. High Memory manager software, HIMEM.SYS -- activated by the Config.sys at boot up, loads utility software into the extra space in upper memory.

            Memory cards were added next to provide Extra Memory.

            More recently memory is added by installing memory modules to the motherboard. This is called Extended Memory. Extended Memory Manager software, EMM386.EXE --activated by CONFIG.SYS, (later described) manages the extended memory. Extended memory became common when Windows software began controlling pixels and more memory was needed.

            Ask for help if need be to set up the CONFIG.SYS file. You can make changes yourself using a text editor after reading DOS or WINDOWS manuals on what to do.

ASCII and UNICODEs

            An alpha-Numeric character requires 8 bits. The industry has used a “Look-Up table” known as the ASCII (American Standard Code version II...) as the standard means of converting keyboard inputs and monitor or printer outputs: characters to bits & from bits to characters. This was an outgrowth of telegraph codes where the lower binary numbers were Start, Stop etc. A decimal number 30 = character 0; 31=1, etc A decimal number 41=a, 42=b, etc

            In a pending “Chicago” version of Windows, to be known as WINDOWS-95 -- pending ‘95 release, the “Lookup Code will shift to a UNICODE (Universal code) of 32 bits. This will accommodate multiple languages, with plenty of range to cover ancient ones.

Binary Arithmetic -- 32 bit floating point numbers

            Arithmetic is done by the microprocessor using binary numbers. To achieve an accurate arithmetic number, especially when doing multiply and divide, it is necessary to carry the number to at least 32 bits. Arithmetic chip, companions to the microprocessor, were built to do 32 bit arithmetic. The 386 microprocessor required such an extra chip; else it uses a special algorithm to recycles 16 bit numbers with multiple adds to do a 32 bit multiply -- it worked but was slow. The 486 microprocessor has a built in arithmetic unit -- making it much faster. (Graphics work requires use of arithmetic to create displays, though no pure arithmetic is being asked for by the user.)

Machine Code or Assembly Language

            The earlier 8 bit machines were programmed in Machine, sometimes called Assembly code. I taught myself how to program in this way as part of the R&D work I was doing. It is powerful and fast but very, very tedious. The method was to use an early word processor to enter characters as a sequence of instructions to the microprocessor. The microprocessor manufacturer provided diagrams of the registers in a microprocessor and acronym commands to cause an action by the microprocessor on the input and place it in the output register. This was talking to the Machine -- get this, do that, put it there..... This written sequence was then passed though software called an Assembler which converted the text to equivalent binary code. The process was very tedious and vulnerable to the slightest error.

            To protect programmers and users Microsoft has modified the Operating systems to have various levels -- the user of a machine is not permitted to modify lower level code used by the machine.

Higher Order Languages

            To overcome the tedium of machine code, higher languages were evolved such as Cobol, Fortran, Pascal, Basic and C. The defense industry went to ADA.

            Basic was built into the early PC’s and the user could write programs to do his own work with considerable simplicity. One could write a program saying A= 10 books, B=20 dollars each and C = A * B = Cost of 10 books. Compared to Machine code it was delightful relief.

            These languages were those available when bulk data storage was on magnetic tape, cassette tapes for the early PCs.

Then came DOS

            Floppy and Hard disks required new and special software. DOS (Disk Operating System) by Microsoft evolved as the operating system of choice -- and has become the standard for IBM compatible machines. It had it’s beginning as BASIC built into early 8 bit machines which used cassette tapes for data storage. This was expanded to accommodate Floppy Disks and Hard Drives -- DOS and BASIC were built into early machines. When IBM got in the act they selected DOS, then well established as the operating system, for their machines. This launched Bill Gates of Microsoft on his way to becoming the Billionaire he is today. Machines which used DOS became known as IBM compatible.

            DOS was designed for use with machines using ASCII Lookup Table software to convert characters to bits and bits to characters. A character generator chip, also a look up table, converted binary character codes to pixel patterns which produced a character on the monitor or printer. This was fast and cost effective. Software written to deal with characters only is known as DOS compatible software.

Higher Order User Software -- DOS compatible

            Early Word Processors were coming out at the same time as high order languages. To keep moving with the R&D I was doing, I’d written a word processor program in Machine code and we used it to do special work on the B-1B. By the time I had it working the industry had produced word processor commercial software such as: XY-Write, Word Perfect, WORD, etc 

            Next followed Spreadsheets such as Lotus, Quatro Pro and Excel.

            Also followed Data Bases such as DBASE IV, Paradox....

            All the above operated in character mode, being DOS compatible.

WINDOWS compatible -- Pixel control

            Apple, using a new powerful Motorola chip, took the first step in going to individual pixel control vs batches of pixels as characters. The publishing industry had already been converting a standard character into special Fonts when sent to a printer. Apple was a big hit in the publishing business.

            Microsoft also made the big jump pixel control using what they called GUI (Graphic User Interface), which has since caught on as WINDOWS, to become the current standard. This Windows operating system was superimposed on top of DOS -- both DOS character and Windows pixel software will run under Windows.

            Windows set the new programming standard -- software immediately became more powerful -- also far more difficult to program. Programming tools such as Visual Basic and Visual C++ are now the standards for Windows compatible programming.

Windows Compatible High Order Software

            Microsoft got the jump on other software providers such as Lotus, Borland, Norvel and Word Perfect; by getting on the Windows band wagon early. All the others have since taken on the difficult but user demanded task of full pixel software control.

            This permits changes to Font Size and Variety, the incorporation of Photos and graphs in text, versatility in the design of data presentation.

User Task Simplified

            The PC user no long needs to know about Machine or Assembly language. They no longer need to know how to program in Basic or C. They can concentrate on using high level tools such as Word Processors, Spreadsheets and Data Bases. This took the PC out of industry onto the desk top.

Mother Boards and Slots

            Hardware began to be standardized when IBM entered the market. IBM got into the PC market late. To enter the arena they had to package what others had already evolved. To do this IBM created a Mother board to interconnect Cards designed by others to perform various functions. This became known as “Open Architecture.”

ISA 16 bit Slots

            The IBM 16 bit bus data path first used on the AT variety, became known as the ISA bus. This became the IBM compatible motherboard slot standard -- all current motherboards include these slots.

VLB 16 bit + ISA 16 bits = 32 bit Slots

            To speed things up, a Video Local Bus (VLB) was added, Video and hard drive cards were extended in length to pick up another set of 16 bits making it possible to move two sets of 16 bits to the newer 32 bit microprocessors. A 32 bit Cache memory is used interface with (2) 16 bit buses. A higher clock rate used in conjunction with a Cache memory, the SRAM cache must be faster than the DRAM main memory. Microsoft Windows for Work Groups, Windows 3.11, provides software which will accommodate operations in a 32 bit data mode.

            Extended cards are readily available for interfacing Video Monitors and Hard Disk drives.

ISA 16 bit + PCI 32 bit Slots

            Newer machines, those using Pentium microprocessors, are shifting to a PCI (PC Internal) 32 bit plug in slots and bus. Such mother boards move data at 32 bits between these slots and the CPU. PCI slot cards are not yet readily available. Some mother boards are incorporating functions on the mother boards that had been  performed by cards.

DRAM -- Main Memory

            Conventional memory is made up of DRAM, (Dynamic Random Access Memory) chips. This is the most cost effective way of putting lots of semi conductor memory into a machine. This kind of memory looses it’s data when power is turned off. It’s the memory the processor accesses directly -- data is normally read into RAM from a bulk storage device such as a floppy or hard disk. These are called Dynamic RAM’s because they must be periodically refreshed or their data will fade away.

            Presently these are sold in SIM (System Interface Module) sets of memory chips on a common small card. These come in 30 pin and 72 pin sizes with 1 or 4 meg, etc capacities with or with out a 9th parity bit per 8 bit byte. These also come in slow or fast speeds. The speed selected must be compatible with the clock speed used. Machines today must have at least 4 meg of RAM and preferably 16 meg. Though the prices of this kind of memory has come down, it is still not cheap.

SRAM -- Cache Memory

            Static RAM is usually used for high speed operations such as a Cache memory. These are usually dual inline packages. These are faster than DRAM because they do not require a refresh cycle, they also cost more and are thus used more sparingly.

CMOS, PROM & FLASH memory

            The typical computer includes a CMOS memory chip, selected for it’s nil power consumption, as it retains it’s memory content by running off the motherboard battery. The CMOS memory is used to store the user defined hardware configuration.

            PROM (Programmable Read Only Memory) is used for storing the Built In Operating System (BIOS) which is used to boot up the system.

            FLASH is a new kind of non-volatile read-write memory which will be used more and more in lieu of the above CMOS and PROM -- it does not require a battery to hold it’s data and it can be rewritten after installed.

            When power is turned ON, the BIOS program reads the CMOS and starts up the system. After reading the CMOS it reads the CONFIG.SYS file then the AUTOEXEC.BAT file, both of which are user programmable. By implementing these, the system is “Booted Up.”

CONFIG.SYS & AUTOEXEC.BAT files

            The user tells the machine is told how to configure itself using the CONFIG.SYS file, look up examples of what these contain in your instruction manuals of from a friend.

            The user then further tells the machine what to do using the AUTOEXEC.BAT file. There can be multiple Autoexec.bat files to perform custom set ups -- Windows now does the task of multiple special Autoexec files under DOS. Again consult the manual or a friend to cutomize your instructions.

HARD DRIVE

            The users software is stored on a Hard Drive. When configuring the system one of the hard drives is defined as the C Drive. The two floppy drives are called A & B. The DOS program FDISK is used to define the drive configuration. For example the user has the option of which drive to declare as the C drive and to partition any of the drives. For example I have (2) 540 meg hard drives partitioned into two parts each creating (4) 250 meg drives which DOS defines as C, D, E, F.

            (I have compressed drives E and F using Microsoft’s dblspace software, such that drives E and F hold 500 meg each. I use drive C for all software tools, drive D for accounting and personal files, leaving E and F for the 19th BG history task.)

            The C drive is FORMATTED to be the system drive which includes hidden files used by the system. When booting up the BIOS goes to the C drive to finish the process. The Config.sys and Autoexec.bat files are usually stored in the Root Directory of the C drive.

            Read the DOS manual on how to set up hard drives -- or have someone who knows do it for you.

Home Base

            When there was only DOS, you booted up to be in DOS and you chose what to do from there. Today WINDOWS has replaced DOS as home base. Windows is set up to have groups of Icons to represent user programs in the machine. The points with the mouse and clicks the one desired, an hour glass appears indicating it’s being loaded. There are many options on how to group and set up the window screen.

            Many have the system boot up to be in WINDOWS ready for a selection by the User.

Motherboard Standard I/O Devices: -- Keyboard, Mouse

            The Keyboard and mouse interfaces are often found on the mother board.

Standard I/O Card Devices: Parallel, Serial, Hard & Floppy and Backup Tape Drives

            Standard ISA bus I/O card will accommodate (1) parallel (LP1) interface, (2) serial (COM1 & COM2) interfaces, (1) pair of IDE type hard drives and (1) pair of floppy drives and (1) backup tape, off the floppy drive interface.

Video Card

            The current standard Video card interfaces with an ISA slot, some of these are equipped with cache memory. Video cards are also available which make use of the VLB bus to provide faster response.

(4) IDE Hard Drive Card -- Software

            The standard I/O card can be replaced with an extended card to pick up the VLB bus in addition to the ISA bus to give faster response.

            These cards are also available, including software, to accommodate up to (4) IDE hard drives. To make use of such an option you should have a Mini-Tower at least for a place to mount multiple hard drives.

Huge Memory Demands to place photos to text.

            When adding photos to text, huge demands are made on memory. The following are some of the options:

            1) Go to large hard drives. I’ve gone from 80 to 200 to 540 meg sizes in the last two years as prices have tumbled and my needs apparent. Today I’d go to 1.0 giga byte size. A 540 meg IDE now sells for $298 and an IDE 1 gig for $599. These prices will be coming down more. Price of smaller sizes have dropped just to get rid of them.

            2) Install a (4) IDE hard drive system, with this you can readily install 4 gig in one mini tower.

            3) Use compression software. With Microsoft’s dblespace you can achieve at least a 2:1 compression. (The proprietary problem was solved).

            4) Add a SCSI interface for up to 7 peripherals.

SCSI Peripheral Interface

            SCSI (Small Computer Serial Interface) was established by Apple to solve their problem of having no room for internal expansion. SCSI cards can be added to an IBM compatible machine -- however the equivalent results can be obtained without it. (I removed the one I had to reduce boot up delays)

PRINTERS

            Originally Dot Matrix printers were the ones of choice as the price was right. These have now been replaced by Laser and Ink Jet printers.

            Ink Jet printers cost less than Lasers and do an excellent job, but they are slower and use a lot of ink if a page in photo intensive. Ink refills cost more than toner refills.

            Laser printers use a toner cartridge for ink. The basic units will print 300 dots per inch and later units 600 dots per inch. Upgrade cards will permit them to go to 1200 dots per inch.

            These units cost more to achieve greater speed and increased functions such as print both sides, large numbers of envelopes, etc.

            At present it would seem that a photo scanned at 300 dots per inch, and printed at 600 dots per inch will produce adequate results for an excellent document.

            I chose to use an HP Laser jet III printer in part because of the special software they use which enhances the sharpness of an edge.

COPIERS

            The problem comes when you make copies of a high quality printer output using a commercial high speed copier doing 150 dots per inch. The result shows a definite deterioration. Hopefully the industry will come out with copiers which will accept a software input and perform a 600 dot per inch output. Such a copier would require a very high speed processor. 

SCANNERS

            For quality work one should use a fixed bed scanner. HP now puts out an excellent, up to 1200 dots per inch, scanner which will also accommodate color. (I’ve not tried it yet.) I’ve been told you do not make much gain by scanning at 600 dots per inch in lieu of a 300 dot per inch resolution. That the gain is in the quality of the printer. If this proves to be true it is a blessing because a 600 dot per inch scan would require 4X more memory than a 300 dot per inch scan.

Local Area Networks -- Windows for Work Groups

            I connected two 486/33 computers on my desk using an Intel LAN (Local Area Network) card which interfaces using a twisted pair wire method. I also installed Windows for Work Groups software which permits operation over this link. By use of such one can read the Hard Drives in one machine from the other or share a printer via another machine.

WINDOWS-95

            The pending CHICAGO version of Windows will be released in 95 as Windows-95. Among other things, this will have an improved networking interface building on Windows for Works Groups. I mention this primarily because Windows-95 will probably become the path by which most users will interface with the Super Highway INTERNET.

            I subscribe to Microsoft’s Developers Network, which comes on CD-ROM, also to The Microsoft Journal which is a magazine. Via these one can be aware of what’s pending, especially in terms of software. Hardware performance increases and lower prices are having a major impact on software demands by users. Thus it behooves the user to be prepared for what’s pending.

Word Processor Software for Windows

            Software written for Windows has standardized on button symbology and menu methodology to make it easy for the user to move from one piece of software to another. New features such as “drag and drop” have become common. For example the size of columns and rows can be changed by dragging and dropping the line to where you want it.

            Subtier programs share the many Font options installed in Windows.

            Segments can be cut from one program and pasted into another by use of a clipboard. The clipboard is very clever in being able to read one format and store in another.

            Segments can also be blocked out with a border, these can be declared as FRAMES and moved about. Frames, photos can be placed and resized by the user.

            Presently the best Word Processor software for Windows is Microsoft’s WORD 6.0. This is what I’m using. The first version I used was Word for Windows 1.0 followed by 2.0. WORD 6.0 is a huge improvement over the original.

            A key WORD 6.0 attribute is that it uses OLE (Object Linking and Embedding). An Object is a package of software in a software envelope -- it can be shipped and used by a compatible recipient because the envelope includes the essential instructions for using the object in the package. Such objects can be exchanged via a network with other machines, be sent over a modem, or be used by compatible software in your own machine. A package of such software is called an Office Suite.

Spreadsheets

            Microsoft’s EXCEL 5.0 is the one which is compatible with WORD 6.0 in that both incorporate OLE methodology. Word can call up a segment out of an Excel spreadsheet and run it, embed it within word.

Data Bases

            I have upgraded to Microsoft’s ACCESS 2.0 Data base software which has also been brought up to computability with the other two. I use ACCESS to maintain the 19th BG Assn membership data.

Software Translations

            High quality software now has the ability to translate between one software format and another. I’ve found that ACCESS, EXCEL and WORD have excellent translator capability. For example Data Base information can be readily moved from one to another. In the past this was a very significant problem.

Software Specialty and Overlap

            Once you become well aquatinted with the above kinds of software you find each has it’s own domain of special value. Word now provides the user the ability to use a table like an arithmetic pad.

            Excel permits performing huge amounts of calculations with much versatility in creating sub files within a file and to move data from one file to another. I use it extensively for my own accounting operations -- which have evolved to being quite sophisticated, taking into account the effects of inflation and incremental additions or deletions and still account for duration of investment.

            ACCESS is excellent for designing the output of data -- I use this to query, sort & create mailing addresses. Distributed data can be arranged and presented more readily than with a spread sheet or word processor.

Information storage

            Floppy disks of the 3.5” 1.4 meg or 5.25” 1.2 meg variety are very useful for archiving small files. By use compression the amounts can be almost doubled. They are slower than a hard drive as they still use an 8 bit interface.

            Backup Tapes are excellent for holding or transferring large amounts of data. The current standard size is a DC2120 1/4” tape cartridge which using compression can hold 250 meg at a 2:1 compression ratio. However these are not guaranteed by the manufacturers to retain data purity for over 2 years. With enough time the closely packed iron oxide on the tapes can cause a change in an adjacent bit.

            Flopticals, a magnetic-optical combination, can store large amounts of data indefinitely. I’ve used these to store 40 meg of data on a 3.5” disk -- however they have not caught on and blank disks are hard to obtain.

            CD-ROM is the present preferred method of storing large amounts of Read Only data. Presently one disk can hold 600 meg without compression.

            Large industrial users now use 1 gig Magnetic-Optical methods, however these are still too costly for desktop use.

            Research labs are working on ideas of optical methods which permit multiple layers at a given bit location by changing the wave length of the optical beam to read only a selected depth at an instant.

FAX Modems

            FAX-Modems are now a part of almost every machine. Pictures (FAX) and Data (Modem) can be readily moved via telephone lines.

            I presently use a 9600 modem card to interface with PRODIGY over the phone and use it to down load market information. This and other services offer news, mail and many other services -- even interfacing to INTERNET, the Global Highway.

CD-ROM usage

            CD-ROMs are currently the fastest selling upgrade -- Hi Fi music via your computer while you work is now readily available at reasonable cost.

            CD-ROMs are now available with all kinds of information from encyclopedias, to atlases and in depth subjects especially useful for school work.

            The industry is presently at work making a 4X faster CD-ROM standard and the ability to store more data -- up to 1.75 hours of movie in compressed format. Microsoft is preparing software for translation of these data at high speed for presentation as a movie on a monitor. Such software prepares the way to connect to high speed fiberoptic highways via an upgraded INTERNET. But that’s a few years away.

IBM-MOT-APPLE vs Intel Pentium

            Apple, IBM and Motorola have teamed to come up with a new microprocessor (IBM concept, Motorola mfg) for a new Apple platform. This is a high risk venture. Recently IBM-Apple announced the intent of a joint design and Motorola additional investments, probably as a part owner of Apple. Its been decided that this processor will be compatible with WINDOWS-NT, an industrial version of WINDOWS-95.

            The above effort is to offer an option to the new Intel Pentium processor. The pentium processor is now being marketed at a premium price.

            Both Motorola and Intel are spending in the order of 2 billion on new chip making plants. These are using the best technology available. Motorola is specializing on wireless communications, and are already offering wireless communications between computers.

            Intel is specializing on making a 3 volt in lieu of the current 5 volt Pentium processor. This reduction in voltage reduces heating, less power drain on portables and increased speed.

            More powerful microprocessors will be needed to implement the anticipated boom in Multi Media via a fiberoptic INTERNET. Fiber optics offers the option of an enclosed “atmosphere” “fiberosphere” over which signals can be broadcast as if radio signals sharing the atmosphere. There can be multiple such “highways” with lickety split speed -- multi channels. Tests showed the entire encyclopedia Britanica could be sent in 1/2 second!!

            Know that what you buy today will be superseded in a couple years -- only don’t wait, join the adventure today....