Organizations behind the standards

Organizations behind the standards

Organizations behind the standards by ROBERT FERGUSON S tandards are agreed methods of making something or doing something, described and published ...

401KB Sizes 0 Downloads 44 Views

Organizations behind the standards by ROBERT FERGUSON

S

tandards are agreed methods of making something or doing something, described and published so that the rest of the world can use the same methods. There are many advantages to be had if the rest of the world can be persuaded to make it, or do it, the same way.

ECMA,ISO, CCITT . . . all internationalbodies intent on bringinga measureof uniformityto the computingworld. Whichdoes what? How do the organizationswork?

Types of standards

Abstract: Due to the high rate of development of new products, establ~hin~ computer standards is a complex process. The article describes the relationship between the various standards bodies involved in computer standards and outlines the process of producing a computer standard. K:eywords: inte~at~~~ standards, computer standards, data processing. Robert Ferguson is international standards coordinator at Sperry Computer Systems.

~0126

no 1

january/february

1984

Standards may be p~y~icu~,so that one manufacturer can make a nut and another a bolt, and know that they will fit; or language, so that people can talk to each other; or quality or performance so that the consumer knows he/she is getting a good and safe product. The use of standards may be encouraged or enforced by law for the common good, such as the laws about radio interference and the construction of srash helmets. Many more standards are voluntarily adopted for altruistic or competitive reasons, such as production economies (variety reduction, smaller stocks, larger orders for piece parts or components); safety; protection of the average consumer who cannot always be sure what he/she is buying or the quality of it; and protection of a reputable manufacturer from less scrupulous competitors. Many standards are concerned with human communications: spoken and written languages, road signs, music notation, morse code, tape cassettes, radio wavelengths, latitude and longitude, time and date; there are also abbreviations standardized many which cut across language differences and spelling variations, like mathematical and physical symbols (rt, “C).

0011-684X/84/010013-03$03.00

0

1984 Butterworth

the field of computer standards the same is true, even more so. In fact most computer standards are to help man to communicate with machine in a way which, with appropriate training, both the man and the machine can understand and handle efficiently. Prime among language standards are the programming languages which convert human instructions. Coding standards enable not only the machine but two users, one putting data in and a different one taking it out again after storage or processing, to understand that data. Control procedures and elaborate protocols enable data to find its way between such machines, using the public telecommunications offerings of the PTTs (Post, Telegraph and Telephone Administrations) or using leased communications circuits. On the physical side, interface plugs and sockets and the nature of the signals on their pins are in many cases standardized, although this is a moving target due to continuous development. In order that portable data storage media can effectively be ported, the dimensions and the storage format are standardized for many kinds of magnetic tape, cassette and disc. Many kinds? Isn’t that a negation of standardization? Well yes, it is another moving target as new hardware and cleverer methods of packing the data on it constantly appear, and last year’s model disappears. Or, as it is fashionable to say, a new generation of hardware takes over. Until then, physical standards cut down the chaos. With the use of these standards, computer users can buy their hardware, software and communications

In

& Co (Publishers)

Ltd.

13

services from many different vendors, connect them up, and be reasonably certain it will all work. To a lesser extent there are quality standards specified for computers, notably for safety, radio frequency, and acoustic noise emanation. The ergonomic design of user operated terminals is also receiving attention, but lends itself more to guidelines than to a rigorous specification. Another aspect of quality is conformance; the extent to which a product complies with physical or language standards. For example, does a computer compiler have to comply with a22 the possible details of a programming language standard, or is it enough that it complies with less than the full set?

Figure 1. Development

,

of international

Liaison /---

Standards

organization

The three main publishers of international standards are the IS0 (International Standardization Organization), IEC (International Electrotechnical Commission) and CCITT (International Telegraph and Telephone Consultative Committee). IEC deals with electrotechnical matters (computer standards are issued by TC74 and the international special committee on radio interference (CISPR)) and IS0 with just about everything else (computer standards are issued by TC97 and TC95). The borderline between the two, which is unnecessary but historic, results in some more or less gentlemanly controversy. CCITT is the international club of the PTTs, and it writes

computer

1

standards to facilitate international telecommunications; some of its standards, e.g. the 7 bit character code, are duplicated by ISO. In 1959 there were no DP standards but a group of European companies foresaw their urgent need and formed ECMA (European Computer Manufacturers Association) whose sole objective is to write DP standards. At about the same time, cooperation also began in the USA between traditionally hostile groups, such as manufacturers, with the same objective. Transatlantic cooperation has existed ever since. Indeed, the membership of ECMA includes most of the major computer manufacturers in the world outside Japan and even that deficiency may soon be rectified.

standards.

,

I nternationai e

. . . .

European

etc

14

Manufacturers

National

. . . .

etc

National

.*

etc

Interested

I.

Bad ies

subsets

Standards

Internatlonal

Governments

Standards

Trade

Trade

Bodies

Association

Associations

Groups

data processing

policy ECMA can be seen as the forcing ground for fully international standards; for example, ECMA I, the world’s fitit ever computer standard, was a 6 bit Code, now embedded in the international 7 bit and 8 bit standards. Figure I shows how the three main standards organizations are supported by the national standards bodies and they, in turn by users, manufacturers, government bodies, etc. ECMA sits in the middle, without a vote in the other bodies, offering irs input as draft international standards and as experts sitting on, or contributing to, international committees.

Producing standards In a fast developing

industry

the

timely introduction of usetiul standards is nor easy. If a standard, e.g. for a new size of disc, is introduced too emly, it will not benefit from enough field experience. If it is inrrodnced too late, manufacmrers with a commercial need to sell their products in order to recoup development costs may have frozen their design into something incompatible with the eventual standard. There is another danger inherent in all standards-writing: the creation of a camel. A camel is a horse designed by a committee. All standards are writren by committees and due to the need for compromise between the various interests represented in a committee, especiaHy where there is a deadline to meet, a hump, or even two humps, are occasionally specified for a horse where a flat back would have been more comfortable. Standards committees are usually composed mostly of amateurs, that is to say highly knowledgeable, professional experts in rhc subject of the standard, bur amateur word-smiths. Usually, but not often enough, the secretary and a few of the others are professiona standards writers, experts in the precise use of words. In the case of international srandards this

amateurism is compounded by the involved. tongues mother many Committees mostly work in English (or French) but individual members keep an eye on how the resulting standard will. translate into their own languages. Those of us with English as a mother tongue are the lucky ones and have great respect for those who have to work in a foreign tongue. Apart from the professional secremembers are committee taries, volunteers, supported by their employers who both meet travel costs and lend the services of the experts they really hired to do something else. In some cases, trade association or government funds are available to assist with the international travel costs. Standards writing takes time. It is a process of finding and recording consensus between parties with different points of view (if they are on your side) or entrenched interests (if they are on the opposite side). The role of the chairperson is vital: h&he must sense the feeling of the meeting, judge what compromise is iikely to be unanimously acceptable, and put his/her personal or other views cm one side, The experts do their best ar a meeting, but when the result is circulated to those who were not new facts and opinions present, emerge and have to be taken into account. Then votes are taken, with adequate time for democratic consideration: five to six months internationally, two months natioionafly, often including a built-in period for public consideration of the draft, It is human nature that in many cases proper consideration is only given under the pressure of the voting procedure; then further conflicts may have to be resolved, The elapsed time from start to publication of an international standard, or a revision of one, can be three to 10 years and is typically five years. ECMA can issue a usable standard rw+three years more quickly than the corresponding international stan-

dard, due to simpler voting procedures, limited geographical area for travel to meetings, and the commercial incentive of irs members - the manufacturers - who need results.

All the ahove applies,

of course, to standards writing of any kind. But in computer standards there are two factors which sharpen the problem: urgency and complexity. Urgency arises from the great pace of development of new products and ideas, which ensures that a computer standard which is a gear late is useless. The complexity, and the intellectual and philosophical demands on data processing standards writers may seem surprising for an industry based on arithmetic which has only two l;ltI=Oandcarry rules (1 +O= 1). BUF the pressures are there noncthe&s, leading often KO SO or 100 page documents, pieced together from numerous cantributions, so rhar it is a major c&t&al exercise even to make one book coherent. A family of even related standards produces greater complexity. There are still plenty of computer standards yet to be written, especially in the data cammunkations field7 Readers who think the lifestyle of a part-time itinerant standards writer would suit them - and it is anything but boring should contact their local standards institution, or a trade association or professional institution to which they belong. ‘They should also have a word with their bosses, and, for the lengthier international meetings, their spouses. Apart from the actual time abroad, a lot of work has to be done between meetings, such as writing draft documents. Nevertheless, the production of computer standards is an important contriburion and one from which the whole industry, as well as users, will 0 benefit. Computer Systems> Sperry Crntre> Stonebridge Park, London NW10 SLS, UK. Tel: Ol-965 0511.

Spere