xerox 19771123 Time Of Day Facilities In Pilot free download

File name 19771123_Time_Of_Day_Facilities_In_Pilot.pdf

To Pilot Interest Date November 23, 1977 From H. C. Lauer Location Palo Alto Subject Time Of Day Facilities Organization SOD/SO in Pilot XEROX SDD ARCHIVES I have read and undel.'stood XEROX Pages ____-----TO --------- Reviewer Date~.------- Filed on: TimeOfDay.memo # of Pag es _ __ Ref .11saD '38~ This memo describes our proposal for facilities for maintaining the time of day and date within a system element. This proposal is predicated on the availability of a reliable, battery-powered Time-and-Date clock in the hardware (in the form of a watch chip or equivalent), and it is characterized by the separation of the time-of-day functions into simple, uniform facilities provided by Pilot and facilities oriented to the human user provided by some form of common software. General Strategy We propose that the clock in each DO processor maintain the time and date according to Greenwich Mean Time. Pilot will implement an operation by which any client can get the current time (GMT) and a private operation by which special programs can correct the time in the case of clock drift, battery replacement, or repair of malfunction. The time will probably be returned in the unpacked format described below (unless the packed format turns out to be more convenient). Common software routines will be provided to convert this value into values of the local time and date and into formats suitable for application programs. These routines will also implement such features as Daylight Saving Time, which depend upon the locality in which the system is installed. It is the intention of this proposal that all values of the time and date stored internally within a system element or transmitted over a network in a form intelligible to DIS software be in Greenwich Mean Time. In particular, Pilot will use the Greenwich Mean Time whenever it records the time, such as the creation date or last reference date for a file. This convention will eliminate much confusion over the meaning of time-stamps and time values communicated by machines scattered over the globe. It will, to a large extent, keep the customer from having to set the time (and get it wrong), and therefore it will allow applications to depend fairly heavily on having a meaningful, consistent value of the time available on any particular system element. It will also allow us to parameterize our software to accommodate local conventions about the official time without having these cons