cazier@mbunix.mitre.org (Cazier) (08/07/90)
What problems would an agency encounter if they requested ISO POSIX in place of FIPS 151-1 POSIX. Are there alternative behaviours allowed in ISO 9945 that would jeopardize maximum portable of applications following ISO vs. FIPS guidelines? Is FIPS 151-1 a proper subset of ISO 9945 or are there items in FIPS 151-1 that do not exist in the ISO version?
lewine@dg.dg.com (Don Lewine) (08/07/90)
In article <116179@linus.mitre.org> cazier@mbunix.mitre.org (Cazier) writes: > >Is FIPS 151-1 a proper subset of ISO 9945 or are there items in FIPS 151-1 >that do not exist in the ISO version? FIPS 151-1 is based on IEEE std 1003.1-1988 while ISO 9945 is based on (is identical to) IEEE std 1003.1-1990. There are a number of technical changes which could cause a program that worked under 151 to fail under ISO 9945. ISO 9945 also has various options. FIPS 151-1 mandates the setting of these options. Since ISO 9945 is newer than FIPS 151-1, there are symbols in 99945 that are not in 151-1. Donald Lewine uunet!dg!lewine
donn@hpfcdc.HP.COM (Donn Terry) (08/09/90)
The relationship is as follows: 1) FIPS 151 (not 151-1) referred to an unapproved draft of 1003.1. (Grrrr!) 2) FIPS 151-1 refers to 1003.1-1988. 3) FIPS 151-1 DOES NOT refer to, nor relate to in any way, 1003.1-1990 or IS 9945-1. (I'm presuming slightly on both titles; the official approvals havn't happened yet, but I don't see anything in the way of either.) 4) 1003.1-1990 and IS 9945-1 will be identical except for the cover pages and the fact that ISO will white out a few non-technical things like line numbers. (I don't want to get into THAT issue, please.) 5) 1003.1-1990 is slightly different than -1988, but not enough so you'd notice much (unless you're a test suite writer where you notice EVERYTHING). The major difference is tightness of langauge. 6) Roger Martin (of NIST) keeps asking me when -1990/9945-1 will be done. I have no data about how much trouble you'd get in to politically; techincally, the hard data is there in terms of the changes. (I hope/expect that the -1990 version will have change bars, but don't bet on them in the ISO version (personally, I'd buy the IEEE version)). You will have to evaluate the changes for how much they impact your situation, but you have a reasonably good chance of coming up with the answer that it really doesn't matter to you very much. There is one gotcha; there are a few little things that vendors will have to touch for the -1990 version, and you may be restricting the availablility of systems to insist on it for short-term delivery. Over time I hope that there will be a 151-2, and that vendors will provide conformant systems. Donn Terry, Chair 1003.1 These opinions are my own, and do not necessarily reflect those of IEEE, HP, P1003.1 or the guy reading this posting.