Posting for administrators in Sept 2001 :
As of this writing, we are wondering
if the tXXXYYZZ convention of naming files will work. The first problem
is that the "site code" (ZZ) is not always known at the time the probe
is being initialized. If the probe is being initialized by someone other
than the person that will be deploying it, they often do not know if the
probe will be redeployed in a new geographic site or put back
on an existing site. The second problem with this convention is that the probes are often moved around to multiple sites between downloads so that the "site code" it was initialize as only refers to the first site it was deployed at.
In the case of AOLA this year, this "ZZ" code is actually represents the "batch" of processing. Each batch of processing occurs on nearly the same date. Note that, in the AOLA case, the "YY" still represents the consecutive time the probe was deployed which may or may not be the same as the "batch" code. While we lose the ability to include the site designations in the filename, in this case, there are two added benefits to batch processing of probes. 1) The operator has the ability to deploy all the probes together in a bath over night prior to reading them in order to check on the probe-to-probe bias. 2) The operator will be sending one bulk email to jmanning along with one large spreadsheet which simplifies the processing relative to set of random emails from time to time.
Despite the problems associated with
the tXXXYYZZ convention, we will continue to use it. It is better than
no convention at all (random filenames). If and when we figure out a better
convention, we will change it. We may decide, for example, that the participants
initials should be part of the standard filename.
Posting for administrators in Sept 2002 :
After nearly an hour-long discussion at the Woods Hole admin meeting (20 Sept 2002), we decided, based on suggestions primarily from the MeLA contingent, to change the file name convention to seven characters "tXXYYZZ" where "t" stands for tidbit , "XX YY" stands for the site code, and "ZZ" stands for the consecutive probe setting. The site code by convention usually includes the participants capitalized initials (XX) followed by a 2-digit # representing the consecutive number of sites that participants has occupied (YY). The "consecutive probe setting" is a 2-digit # representing the consecutive deployment of any probe at this particular site. We also decided that probes deployed above the trap (either near-surface or within the water column) will be given a distinct site code.
Participants and administrators should understand that there are three opportunities to specify filesnames. The first opportunity is during the probe initialization, the second at download time, and the third just prior to processing in Woods Hole. So, if the proper name is not assigned at first there are chances to change it. For example, the probes initialized on the morning of 20 Sept 2002 were given the old filename convention so when they get downloaded we should specify the new convention.
Often times a raw (dtf) data file will result in multiple (txt) exported files in cases where the probe was deployed at multiple sites. Since the probe serial number does not get exported to the txt file and it is no longer part of the filename, it is important to include the serial number in the documentation associated with each deployment.
While the master copy of the raw
data files are archived at Woods Hole, participants and administrators
should refrain from deleting files from their BOXCAR data directory
in case it will be necessary to recover files sometime in the future.
The space occupied by these files is insignificant in today's disks. Note
that all files have proper time stamps associated with them only if the
PC time is correct so it is important to continually check the PC time.