On 7 Mar 02 15:32:19 GMT, kviel@GMCF.ORG (kviel) wrote: >Greetings, [ ...] > 2) On the mainframe, the base is 16. So using two bytes, I have one >full byte to store > the mantissa. However, since the largest integer is 256 (TS-654, >page 2), > I infer that it is not stored as: [...] > This is consistent with my limited knowledge of hardware: a bit can >have only 2 states, > not 16. What does it mean, then, that the mainframe is BASE 16?

you can kinda treat an ibm maineframe exponent as signed, i think a better word is its "biased"....

and there is someting special about the mantissa, i belive there is an assumed left bit set, ie the mantissa is shifted left for all left occuring zeros. these are called normalized numbers. i think this is what makes ibm floating point not fully ieee format numbers. ( i think, there is some reason they are not truely "spec")

historicly the ibm maineframe(360+ series) is refered to as base16 because its (16**e)*m, some machines the base was 2, some it was 8, ibm picked the base of the exponent at 16. it is assumed, not specified, ie "its base16".

binay, trinary, octal, decimal ,,, isnt it sexadecimal in that "base" anyway? ....ibm's have always been sex machines realy ..... not hex ....i went all thru this working with virtual machines with "tit"s, ... ...trinary digits.... i think we are all glad they settled on hex......

Besides the other web pages mentioned, i suspect you can find out more than you want at the newsgroup alt.folklore.computer, many of the true decision makers of our current computer world read and post there. i decided against cross posting this there, some threads can go 300+ posts, this could be one of them.

google has access to that was well as other newsgroups, including this as comp.soft-sys.sas for those getting sick of the email format.

look for posts by lynne (& anne) wheeler(?), he is an old old ibm'er that has great web links in his posts.

>Kevin

Huck (an x-email and x-digest reader myself) [...Trimmed for the digest readers]