avi@pegasus.UUCP (Avi E. Gross) (01/30/85)
When a process does a fork()/exec() (or system()), I am used to using words like "parent" and "child" to refer to the process depending on what PID was returned from the fork(). Similarly, I have used "grandparent" and "grandchild" -- and even "sibling" to refer to other relationships. Some of the processes I generate easily go 7 generations and more. I am not aware of any terminology (of an anthromorphic nature) to describe the state of a process immediately before an exec() or after. It is not really a parent-child relationship -- it is more like a sex-change operation or reincarnation. In reality, the born-again process shares many characteristics of its cognate -- such as open file descriptors and process table space, but may have a different name and program text. Can anyone suggest a reasonable (and hopefully humurous :-) set of names that can be used to talk about these processes? What does one call the result of several exec's in a row? Does it matter which flavor of exec (execl, execv, execle, execve, execlp, execvp) was used? Can such a process find out whether it was the result of an exec from another process, or whether it was the exec of the original shell (hopefully ksh)? All kidding aside, I am curious whether other people have felt the need for a way to refer to several versions of the same process. I have recently been creating such processes and using terms like "overlayed process" doesn't excite me. -- -=> Avi E. Gross @ AT&T Information Systems Laboratories (201) 576-6241 suggested paths: [ihnp4, allegra, cbosg, hogpc, ...]!pegasus!avi
mark@tove.UUCP (Mark Weiser) (02/02/85)
In article <2068@pegasus.UUCP> avi@pegasus.UUCP (Avi E. Gross) writes: >I am not aware of any terminology (of an anthromorphic nature) to describe >the state of a process immediately before an exec() or after. >Can anyone suggest a reasonable (and hopefully humurous :-) set of names that >can be used to talk about these processes? What does one call the result of >several exec's in a row? Does it matter which flavor of exec (execl, execv, >execle, execve, execlp, execvp) was used? Can such a process find out whether >it was the result of an exec from another process, or whether it was the >exec of the original shell (hopefully ksh)? After fork but before exec, a process is a blastomere. This is a biological term that refers to the individual child cells of the dividing fertilized egg, before the cells have begun any differentiation or specialization, and especially before they have done any growth before division. This is a lot like pre-execed forked processes. After execing there is not an exactly analogous biological term, so `embryo' might work ok. Alternatively, before execing processes are clones, and one can speak of the parent clone and child clone and even grandparent clone. After execing processes are independent people. Only then do the terms sibling and unqualified-parent come into play. For example: "sendmail uses mostly clones, not often people. A grandfather process acts as a demon, spawning grandchildren and killing off their parents so the grandchildren will be taken care of by the orphanage (process 1). The clones become people only when another mailer is called, in which case the grandfather spawns a child directly and does not kill the parent (itself) but instead watches to see how things go." Or: "The sendmail blastomeres are taking over the machine. Good thing they are not people or we could never tell who made them." -- Spoken: Mark Weiser ARPA: mark@maryland Phone: +1-301-454-7817 CSNet: mark@umcp-cs UUCP: {seismo,allegra}!umcp-cs!mark USPS: Computer Science Dept., University of Maryland, College Park, MD 20742
rcd@opus.UUCP (Dick Dunn) (02/12/85)
> I am not aware of any terminology (of an anthromorphic nature) to describe > the state of a process immediately before an exec() or after. It is not > really a parent-child relationship -- it is more like a sex-change operation > or reincarnation... Not particularly anthropomorphic, but how about "transmogrificand" before the exec() and "transmogrifee" after? -- Dick Dunn {hao,ucbvax,allegra}!nbires!rcd (303)444-5710 x3086 ...Cerebus for dictator!
topher@cyb-eng.UUCP (Topher Eliot) (02/19/85)
>> I am not aware of any terminology (of an anthromorphic nature) to describe >> the state of a process immediately before an exec() or after. It is not >> really a parent-child relationship -- it is more like a sex-change operation >> or reincarnation... > >Not particularly anthropomorphic, but how about "transmogrificand" before >the exec() and "transmogrifee" after? Not particularly pronounceable, either :-) Cheers, Topher Eliot Cyb Systems, Austin, TX {gatech,ihnp4,nbires,seismo,ucb-vax}!ut-sally!cyb-eng!topher