For anyone that enjoys a good shaggy-dog story – after adventures rivalling those of Indiana Jones, the rary bird stands on the cliff-top alongside its treacherous companion and faces imminent destruction with the remark it’s a long way to tip a rary – we reported in May last year (CI No 920) that, despite criticisms from principal C originator Dennis Ritchie, the ANSI X3J11 standard for the C programming language was nearing completion, with the end of the time period allowed for public comments on the proposed draft, and that a full release could be expected thereafter. However with no standard forthcoming, we heard from the British Standards Institution in July of this year that the standard would be late because of objections, raised on the grounds that as it stood X3J11/88-158 as it is known, would present a number of problems for programmers writing embedded C code. While we are now – finally – assured that the ANSI C standard will be published this week, apparently it has not changed one iota from its original guise, established 18 months ago or more.
Disgruntled
Why? It seems that after the period set aside for public comment had elapsed, one disgruntled American C programmer, Russell Hansbury, asked, quite rightly, why he had received no reply to his comments regarding the pitfalls that awaited embedded systems programmers in the standard as it stood. Very broadly Hansbury proposed changing the precedence of operators in C and although this might have improved the language for some at least – such changes would have made most C software obsolete: in the words of one observer it would have effectively meant changing C into something else, like D. The X3J11 committee found that Hansbury’s comments had been received by its technical group, but had been lost somewhere between there and its reviewing body. X3J11 eventually considered his objections, but then rejected them and sent the standard off to the American National Standards Institute – ANSI – for adoption. Hansbury, somewhat miffed at the blunt rejection of his comments, and with the support of other like-minded C-types, subsequentally wrote to ANSI with the same objections. – Standards efforts are often tortuous, but seldom has so little progress been made in so much time as in the efforts to standardise the C programming lan guage, increasingly the lingua franca of the systems programmer. William Fellows tells the frustrating story. -ANSI, somewhat wary of adopting the standard as it stood, when there was potential opposition, agreed to review the comments, but had no established appeals process to do this. In the face of the bureaucracy he was encountering, Hansbury eventually gave up and withdrew his appeal, which had already held up the standard process for some considerable time, and as a result, the ANSI C standard will be adopted this week. The standard – albeit rather belatedly – paves the way for government procurement contracts to go ahead, specifying requirements for C compilers that conform to the standard. Indeed it is thought that many companies have been waiting patiently in the wings for ANSI C to materialise, and are set to release ANSI-compliant C compilers IBM and DEC amongst them. Furthermore, insiders reckon that as yet none of the C compilers on the market would pass the conformance tests that have been established as a result. The British Standards Institution for one has a European conformance testing service for C compilers that has been lying idle for some time, and is now to begin offering its testing suite to C compiler manufacturers. At the same time, ANSI C is also about to undergo a process of revision that will address and document all the known problems with the the existing draft – without any fundamental changes – and will lead to its adoption by ISO, the International Standards Organisation. Again a period set aside for public comment will be followed by a review process, estimated to take up to a year all told. This revised draft will then be adopted by ISO and ANSI as their definitive standard for the C programmin
g language.
UK trouble brewing
It has also emerged that there is trouble brewing over UK standardisation procedures in information technology. Currently a whole host of committees meet under the umbrella of the Information Standards Technology group – IST-5 – to monitor events and establish standards processes, which are administered by the Department of Trade and Industry. These are open to all interested parties from industry, government and education alike. However, ructions are in the air as the Department of Trade & Industry, under the auspices of Project Disc, is set to change all this. It reckons that the industry spends vast sums on trying to establish its own standards, but not enough on contributing to the bodies that oversee them. Consequently, from next year, delegates will have to pay for the privilege of sitting on these committees, with those that pay the most eligible for the most representation. Insiders say that there is much opposition and bad feeling about this move, and that standards work has all but ground to a halt for the present. The British Standards Institution was in the process of trying to set up a C++ standards committee, but has given up on its efforts until the new guidelines become clear. Meanwhile the US is gaining a clear head start – ANSI has already set up a C++ language group and is pushing ahead to define a standard.
