On 2013-12-23, at 14:37, Ian S. Worthington wrote: > > Just seen this. So it is! > > Never seen that before. Have the architecture folks finally understood that > the PoO is the main reference manual for coders, not the HLASM manuals, and > quietly slipped it in the corner? Don't think they need to be /quite/ so > reticent about it! > Clearly, the extended mnemonics are features not of the hardware, but of the assembler, and deserve to be documented not with the former, but with the latter.
But can't the same be said of the basic mnemonics? After all, the hardware defines only the binary opcodes; all else might be considered properly the realm of software documentation. The design should be based, however, on what's reasonable and useful. Originally, the operation code consisted of only the first 8 bits of any instruction; bits elsewhere could be considered modifiers of a single operation code with a single mnemonic chosen by the hardware designers. Nowadays, that's far from true. To muddle things further, the organization of the PoO implies that the various associated RX, RR, SS, and SI forms are deemed variants of a single instruction (but with distinct mnemonics). Notwithstanding, regrettably, many releases ago the PoP outgrew the capacity of Bookie and has since been available only in PDF. I suppose it might just be considered ahead of its time. -- gil
