I've noted that the document to which many refer as "the bible" is actually not from the original manufacturer. Has anyone actually gone through the original document and compared the Philips version with it in order to ensure that there are no deviations or omissions?
I have not done this "word by word" but found no errors and no omissions. As a matter of fact I use both version, I have and use the old Intel book, but when a search is needed, I use the Philips files. The basic difference is that the Intel version has some references to T2, which the Philips version does not. My guess is that by the time the Philips version came out, the T2 was "nonstandard" (RCAP) in most current derivatives anyhow. However, since I have yet to find a datasheet for a T2 equipped device that does not include T2 it seems that the Philips "no T2 bible" is by now the standard.

I think it's very wrong for a manufacturer to make assumptions regarding what an end-user knows when creating his datasheet. The datasheet should contain the correct value of every parameter associated with every signal, and it should precisely define every detail associated with every instruction.
That would make all the "little nice additions" drown and totally become invisible in the swamp of already known information.

If it is deemed more practical to refer to a separate document common to a group of such details, that document should certainly be readily available.
well it is "readily available" any '51 user with any sense has it downloaded to a olocal disk. It can be discussed if a link should be included in every '51 datasheet, OK you may have a point there.

Erik