nagler@olsen.UUCP (Robert Nagler) (07/18/87)
Could we not have a 24K posting of uncommented definition modules? Isn't there an archive that could be used as an on-demand distribution point as opposed to a flooding algorithm? Physical mail is not a bad alternative. A few questions for the BSI committee: Is there a good natural description of the proposed library so that an implementer would have a better chance at implementing it? I found that the previous Ad Hoc Proposal (listed in Modus) was quite under-documented. The descriptions of the current proposal I have seen so far seem inadequate from an implementer's standpoint. A short test suite would be wonderful. There must be at least a few example programs which have been written using the proposed standard. I seem to remember something about N81 being a list of sample programs, but I have yet to see it. In any event, I suppose all members of the committee were required to write a sample program (or two) using the proposed standard. I haven't counted, but there must be at least 25 members on the committee. So, one would guess there might be between 30 and 50 test programs floating about. It would seem rather silly to expect people to implement the library without these programs, wouldn't it? To summarize, I would expect the package of definition modules should be accompanied by the 30-50 sample programs and a short implementation guide (describing some planned implementation approaches). What do other implementers think? Rob