Last month we mentioned some apparent communication problems regarding access to the new Amiga from Commodore. We're happy to report that comments in our editorial became moot before they reached print. Commodore's new senior management team moved quickly and smoothly to see that we, along with other magazines in the industry, received even-handed treatment in access to information.
The Amiga is an important product. We see a significant, lasting change in the way personal computers will be used and programmed and, thus, in the ways we cover computers. With the introduction of the Amiga (see the story on page 16) and the ST from Atari, consumer computing will never be the same again.
Among other things, BASIC now faces its first serious challenge as the language of popular computing. When you turn on these new computers, you don't see the familiar BASIC greeting "READY." Instead, you see a Macintosh-like "desktop" screen with icons, etc. This manager is called Intuition on the Amiga, GEM on the ST. BASIC is only one of several options, several languages you could load into the computer from disk. A simple command, however, exits this environment and lands you in an IBM PC-like Amiga-DOS, said to be quite like Unix, an operating system first developed for large minicomputers. The Atari ST'S TOS will be similar. Both are command-rich systems, nearly languages in themselves.
COMPUTE! expects to continue to publish the majority of its programs in BASIC. The new machines' BASICS are large and fast. They include a generous set of graphics and sound instructions. Above all, everyone who buys an ST or an Amiga will have BASIC. That language is being shipped with, though not built into, these computers.
Interestingly, most commercial software announced so far for the ST and Amiga is not being written in machine language. Instead, it is being written in C, a language popular among professional programmers which has a reputation for portability between computers. Some have argued that this spells the end of assemblers, the end of writing machine language programs. We do not find that argument compelling.
The argument goes like this: The new machines are faster (because the microprocessor, the 68000, is more efficient) and thus maximizing speed of execution by using machine language is no longer necessary. Compiled languages like C run sufficiently quickly. Lotus 1-2-3 is written in C. Also, some new BASICS and operating systems are largely C.
The other factor in favor of machine language, its conservation of memory, is now less critical, too. Compilers can use up computer memory rapidly. Amiga BASIC, written mostly in C, is about 96K large; Commodore 64 BASIC, written entirely in machine language, uses up only 8K. Instead of having to fit everything into 64K, the maximum memory which can be easily accessed by the older 8-bit chips, the new computers can access megabytes of memory. Tecmar, an Ohio company, is developing an expansion board for the Amiga which adds up to two megabytes of memory. Hence, bulky, compiled programs don't cause much of a problem. There's memory to spare. However, even though the Amiga and ST each have 192K of ROM space, both machines' operating systems-written largely in C-have to be supplied on disk with early models. The compiled C is too big to be built into ROM until programmers can optimize and condense the code.
C has its advantages, but one fact is overlooked: Machine language is the computer's language. All other languages are compromises, less direct ways of telling the computer what you want it to do. This indirection slows the computer down for many of the same reasons that you would be slowed down in a foreign country. No matter how similar the two languages, from time to time you would be forced to resort to hand signals, symbols, even to looking things up in a dictionary. Likewise, a compiled programming language results in a more or less indirect communication with the computer. Even the best compilers produce bulkier and less efficient programs than does pure machine language.
Something similar to the current popularity of C happened when home computers were first introduced. BASIC was then the most common language for commercial programs. Spreadsheets, word processors, and games were sold which were entirely BASIC.
They were slow, had few features, and used up much of the available memory space.
Now that there is a transition from 64K to 512K, quadruple the processing speed, and far better graphics and sound-most any good program is going to be impressive. The new machines make their software look good in the same way that calculators made the early 8K Commodore PET look good. It's a whole new level of power and control. But the shock of the new doesn't last. Software companies will compete along the classic lines: They will all try to offer the fastest product with the most features. Once again we are likely to see a migration to machine language as programmers vie with each other to take their machines to the limit.
The 68000 is not a new chip, but it is new to home computers. Introduced by Motorola in 1981, it cost over $200 until recently. It is the chip in the Apple Macintosh, and sales of that computer have helped drive down the price to its current $20, making it affordable as the new consumer CPU. How does the 68000 differ from the 6502, the chip in most current popular computers (Apple, Atari, Commodore, etc.)? Essentially, things like multiplying large numbers are easier to do, fetching and storing is faster and more efficient, what took several steps to accomplish in the 6502 can now be done in a single operation.
Of course, we won't see the ultimate software the minute the new hardware is introduced. It will take time for programmers to investigate the new territory. But judging from the preliminary software we've seen, the new computers offer stunning opportunities for creative programming and-whatever languages are used-the resulting software will take us far beyond what we've experienced on today's home computers. We plan to bring you some of that stunning programming in the pages of COMPUTE! in the coming years.
Editor in Chief