the execution phase - returned at /(jobs)/(jobid)/phase standard xlink references Enumeration of possible phases of job execution The first phase a job is entered into - this is where a job is being set up but no request to run has occurred. An job has been accepted for execution but is waiting in a queue An job is running An job has completed successfully Some form of error has occurred The job is in an unknown state. The job is HELD pending execution and will not automatically be executed - can occur after a PHASE=RUN request has been made (cf PENDING). The job has been suspended by the system during execution The job has been aborted, either by user request or by the server because of lack or overuse of resources. The complete representation of the state of a job this is a client supplied identifier - the UWS system does nothing other than to return it as part of the description of the job the owner (creator) of the job - this should be expressed as a string that can be parsed in accordance with IVOA security standards. If there was no authenticated job creator then this should be set to NULL. the execution phase - returned at /(jobs)/(jobid)/phase A Quote predicts when the job is likely to complete - returned at /(jobs)/(jobid)/quote "don't know" is encoded by setting to the XML null value xsi:nil="true" The instant at which the job started execution. The instant at which the job finished execution The duration (in seconds) for which the job should be allowed to run - a value of 0 is intended to mean unlimited - returned at /(jobs)/(jobid)/executionduration The time at which the whole job + records + results will be destroyed. returned at /(jobs)/(jobid)/destruction The parameters to the job (where appropriate) can also be retrieved at /(jobs)/(jobid)/parameters The results for the job - can also be retrieved at /(jobs)/(jobid)/results This is arbitrary information that can be added to the job description by the UWS implementation. The identifier for the job This is the information that is returned when a GET is made for a single job resource - i.e. /(jobs)/(jobid) The list of job references returned at /(jobs) ISSUE - do we want to have any sort of paging or selection mechanism in case the job list gets very large? Or is that an unnecessary complication... A reference to a UWS result The element returned for /(jobs)/(jobid)/results A short summary of an error - a fuller representation of the error may be retrieved from /(jobs)/(jobid)/error characterization of the type of the error if true then there is a more detailed error message available at /(jobs)/(jobid)/error the list of input parameters to the job - if the job description language does not naturally have parameters, then this list should contain one element which is the content of the original POST that created the job. if this attribute is true then the content of the parameter represents a URL to retrieve the actual parameter value. It is up to the implementation to decide if a parameter value cannot be returned directly as the content - the basic rule is that the representation of the parameter must allow the whole job element to be valid XML. If this cannot be achieved then the parameter value must be returned by reference. the identifier for the parameter