Companhia Siderúrgica Paulista - COSIPA, Brazil
About twenty years ago, the widespread availability of low cost computational
power that we have today would be unbelievable. For its turn, the graduation
student of today would be surprised with the huge difficulties associated
with the use of computers at that time. In fact, the most powerful computers
available in the 1970's for commercial and academic use were the famous IBM
360 and 370 mainframes. They had incredible (at that time) hundreds of
kilobytes of main memory. The older IBM 1130, a more limited machine with 32k
of core memory, were still active in those days. Hard disks and magnetic
tapes as auxiliary memory, with some megabytes of capacity, were already
available.
Direct access to these beasts was a rare privilege. Normally they were operated on "batch mode": each user had to deliver its programs as decks of punched cards or rolls of perforated paper tape to be processed. After some hours - or eventually days, due to maintenance problems or work overload - the results of the job were available. They were frequently limited to error messages. So, simple software projects, like the examples of this chapter, easily took more than weeks to be completely debugged.
The availability of software was also not very wide. Fortran and Basic compilers were (and still are!) relatively common, but its use was not very friendly. Compiling, loading and running a program demanded an user acquainted with cryptic DOS commands and codes. Word processors, spreadsheets and Internet access were, at that time, dreams of sci-fi.
And, last but not the least, computer time was very expensive. Its use was severely controlled, and eventually each user had a limited account for computer use. So, normally a program had to be fanatically revised before each run. This process normally was a terrible waste of time - but, at that occasion, human time was much less expensive than computer time!
So, computing in the 1970's was not a very popular task as it is today. Only mathematical-intensive fields of science and engineering, like Physics, Chemistry as well Civil, Mechanical and Chemical Engineering developed a tradition in the use of computers. In the particular case of Materials Science, computer applications were frequently discouraged by the high cost of computer processing and by the huge efforts necessary to develop and debug a program in that environment. Only in very specific cases - crystallographic texture analysis, for example [1] - the use of computers was justified. However, it is interesting to note that a book on Materials Science edited more than twenty years ago [2] already included some simple Fortran programs related to applications in this field.
In fact, situation was beginning to change. In the beginning of the 1970's, mini-computers had their debut. These machines - like the famous PDP-11, HP-2100A and so on - were quite more affordable than IBM mainframes, so they were accessible to most schools and small companies. However, as there is not such a thing like free lunch, the simplest models of these machines had very limited memory capacity (about 4 kBytes!!) and no hard disk or magnetic tape units. So data had to be stored in perforated paper tapes generated by teletypes, at 10 characters per second. The compilation of Fortran or Basic programs in this equipment required one or more paper tapes to be generated during the compilation process, as its main memory was not sufficient for this purpose. So, the development of a program in these machines could easily turn into a nightmare, because all that work could be wasted by a single typing error! However, the low price of the mini-computers made feasible its use in the control of some laboratory equipment, e.g., electron probe micro analysers.
However, it was the advent of the programmable pocket calculators the first sign of a new age in computing. For the first time, it was possible to execute tedious calculations in an individually and easy way, enabling a more effective use of mathematical approach in science. The most powerful pocket programmable calculators at this time,like TI-59 and HP-41, had hundreds of program steps available, as well data and program storage in magnetic cards and thermal paper printer output. The TI-59 included built-in software in replaceable ROM's, including statistical and engineering applications. However, no programs in the field of materials science were available in these software modules.
The advent of the 8080 microprocessor, in the mid-seventies, was the seed of the microcomputer revolution. In the end of that decade, personal microcomputers became a commercial reality. Of course, they were much slower than mainframes, but, in their best version, had fair amounts of memory (up to 256 kbytes), floppy disk drivers and dot matrix printers. At that time, the Apple II+ was the most popular computer, mainly because it was able to run the first spreadsheet program ever developed: VisiCalc. This program provided a friendly and intuitive interface for simple calculations, a real progress for average people, instead of the mainframe's painful environment. However, if this program was an huge success among the computer novices, the hard-to-die Fortran and Basic fans still prefered to use the microcomputer built-in Basic interpreters. Even so, there was some developments about the use of this kind of program in scientific applications [3], in spite of its limitations.
The PC microcomputer, based on the 8086 microprocessor, was the delayed answer of IBM to the menace of the microcomputers against its mainframes. It quickly became the market standard equipment, unfortunately beating more conceptually advanced machines, like the Macintosh. The greater computing power of this machine, with main memory up to 640kbytes, as well IBM support, boosted the development of professional software for it. For example, in 1985, the American Society for Metals (A.S.M.) released several metallurgical IBM-PC compatible-softwares, including an electronical version of its Metal Handbook. At the same time, the same association edited a book about engineering applications for Multiplan spreadsheets, including metallurgical case studies [4].
In the mid and late eighties, advanced versions of IBM microcomputers, like the IBM-XT and IBM-AT, included hard disk storage up to 32 Mbytes. The relative low price of these machines allowed its wide use for the control of laboratory equipment; in the field of Materials Science, they were (and still are being used!) attached to X-Ray diffractometers, mechanical testing machines, image analysers for quantitative metallography, electronic microscopes and so on.
This increasingly computing power also generated more useful software. Graphical output, extended scientific and statistical functions became standard characteristics of spreadsheet programs, making them more attractive specially to scientific people. One of the most used spreadsheet programs at that time was (and still is) Lotus 1-2-3.
The development of new microprocessors, like the 80486 and Pentium, and the continuous decrease in hardware prices made available computers with dozen of megabytes of RAM and hard disks with more then 1 Gbyte of capacity. This was enough computing power to allow the implementation of graphical interfaces in the IBM microcomputers. This concept, originally developed for MacIntosh machines, created a very friendly interface between computer and the user. In fact, it banished the traces of the DOS age still existents in the microcomputers, including its technical codes and cryptic commands. So, computing is increasingly becoming a more pleasant task, attracting people that were originally refractory to computers. This "invasion", of course, disgusted old computer users, the "Fortran IV Wizards", that generally do not like Windows, considering it a shameful tool for non-professional people. May be these people are only jealous guys, but certainly you would give them a bit of reason if you lose a 100 page text due to an unexpected GPF error!
The most used spreadsheet programs used nowadays in the Windows ambient are Excel, Quatro-Pro and Lotus 1-2-3. They included multiple resources, like more extended mathematical and statistical functions, graphical and multimidia capacity, data base management and macro programming. The macro language for Excel is Visual Basic - so, dinosaurs of the mainframe age can even use the spreadsheet only for data input or output, making all the calculations with the Visual Basic macros! Besides that, as they are general use programs, they are widely available and its price is much lower than of specific software packages. That is, modern spreadsheets present a maximum benefit:cost ratio regarding applications of data analysis and mathematical processing.
In the field of Materials Science, some papers about the use of modern spreadsheets included applications as a data base management system for the steel industry [5], for the design of exercises in a polymer laboratory course [6] and a review about their use on problems about ceramic materials [7].
In the book Spreadsheets in Science and Engineering, edited by W.G. Filby and recently published by Springer Verlag in Germany, there is a chapter where are shown some classical computer applications in Materials Science, in the form of Excel 7.0 spreadsheets. This text is an extended version of the Introduction of that chapter. The following Materials Science examples were developed in the book:
- REFERENCES
|
Last Update: 06 December 1997 | |
| © Antonio Augusto Gorni |