The failure of DEC to gain a significant foothold in the high-volume PC market would be the beginning of the end of the computer hardware industry in New England, as nearly all computer companies located there were focused on minicomputers for large organizations, from DEC to Data General, Wang, Prime, Computervision, Honeywell, and Symbolics Inc.

The system included a user-changeable ROM chip in a special casing to support their keyboard layout and language of the boot screen.

The distinguishing characteristic of the "A" model from an end-user perspective was that the earlier firmware did not support booting from a hard disk.

Other distinguishing hardware features included the three 2764 (8 KB) ROM chips holding the system firmware[4] and the case fan/power supply combinations.

[5] In addition, the 100A was unable to move its hardware interrupt vectors to avoid the conflict with MS-DOS soft INT 21, etc.

[6] The firmware allowed selection of the boot screen language and keyboard layout, eliminating the need to switch ROM.

[9] The 8088 bus was used for control of all other subsystems, including graphics, hard disk access, and communications.

Initial versions of the operating systems on the Rainbow did not allow for low-level formatting, requiring users to purchase RX50 media from Digital Equipment Corporation.

[citation needed] The unusual orientation confused many first-time users, who would complain that the machine would not read the disk.

[13] Third-party hard-disk controllers were also available, including a dual winchester support from CHS[14] The base Rainbow system was capable of displaying text in 80×24- or 132×24-character format in monochrome only.

[15] The base Rainbow system generates a TTL 15 kHz composite-video signal compatible with RS-170 (NTSC) in monochrome mode.

With the inclusion of the graphics option, the Rainbow could also output sync-on-green RGB video signals at TTL levels.

The Rainbow 100 and the other two microcomputers which DEC announced at the same time (DECmate II and Pro-350) had two quirks that annoyed conservative users.

Towards the end of the Rainbow's life, users were able to run some IBM PC-compatible MS-DOS software using an emulation application called Code Blue, though it emulated only the IBM PC's BIOS and some of the hardware, so programs that accessed the video cards directly would not work very well.

One DEC documentation pack for developers included a listing of Microsoft assembly code to handle this.

However, many significant commercial software products were writing directly to the hardware for a variety of reasons, including performance.

After the Compaq Portable and other clones, the market expectation was that all MS-DOS versions would be fully IBM PC compatible.

Later, Microsoft would stop licensing distinctive OEM versions and sell standardized MS-DOS 5.0 at retail.

While "Code Blue" did a good job at emulating the IBM BIOS, its inability to trap references to the video and other hardware limited what would run on the Rainbow.

The FOSSIL TSR allowed several terminal programs and editors to run on the IBM-PC, Rainbow, and other early 8088/8086 computers, but its limited adoption hampered its usefulness.

Such patches circulated, but new releases made these difficult to keep up with, and over time these hacks dried up (the online archives have very little new after 1991 or 1992, although some of that may be due to the RABIT TSR,[33] which solved the problem generically for all Borland products).

Instead, it had a number of expansion slots that could be used for a single purpose only: extra memory, graphics, rx-50 floppy controller.

One slot was originally designed for a DMA enabled serial card, but hard disk controllers were used there instead because DEC bet wrong on which was more important.

[35] While the software incompatibilities were bad, the lack of expansion hardware flexibility was worse, and the inability to use ISA cards, despite their flaws at the time, played a significant role in the Rainbow fading from the scene.

The group planned on starting the BBS on this machine as soon as it arrived; but, when it did, they found that the Z80 did not have access to the serial ports.