Squid3 is being developed by the Cyber Infrastructure Research and Development Lab for the Earth Sciences (CIRDLES.org) at the College of Charleston, Charleston, SC and Geoscience Australia as a re-implementation in Java of Ken Ludwig's Squid 2.5. - please contribute your expertise!
A foreseeable issue with our CSV-based reports for Unknowns is the need for users to be able to separate data for their various different Unknown samples from each other, to do more detailed sample-specific investigation and interpretation. At present, although Sample-identifiers are strongly implied by user-prefixes, the CSVs do not contain a formal, explicit record of the classification-tree prefixes actually used by Squid3 to handle the data.
We do get an indication of them when viewing the Reports inside Squid3, where the Sample-identifiers are shown in the orange rows:
but this information does not make it into the CSVs. A simple and comprehensive solution for the CSVs would be to write the selected classification-tree prefix in a separate column (to the left of Fraction!), in every row to which it applies.
I realise this measure looks kind of obsolete when applied to prefixes defined by delimiter, but I am thinking of prefixes of variable length, where the user has established a labelling hierarchy for their own purposes.
One example might be where you have two rows of reference material T on your mount, and you prefix spots from one row as TA and the other as TB. That gives you the option of calibrating your unknowns relative to RM data from either one of those rows (by specifying the appropriate two-character prefix), but it also gives you the option of calibrating your unknowns relative to the combined RM dataset from both rows (by specifying the one-character prefix 'T'). Squid3 knows exactly what choices you made in this regard: it just needs to report those in the CSVs.
A foreseeable issue with our CSV-based reports for Unknowns is the need for users to be able to separate data for their various different Unknown samples from each other, to do more detailed sample-specific investigation and interpretation. At present, although Sample-identifiers are strongly implied by user-prefixes, the CSVs do not contain a formal, explicit record of the classification-tree prefixes actually used by Squid3 to handle the data.
We do get an indication of them when viewing the Reports inside Squid3, where the Sample-identifiers are shown in the orange rows:
but this information does not make it into the CSVs. A simple and comprehensive solution for the CSVs would be to write the selected classification-tree prefix in a separate column (to the left of Fraction!), in every row to which it applies.
I realise this measure looks kind of obsolete when applied to prefixes defined by delimiter, but I am thinking of prefixes of variable length, where the user has established a labelling hierarchy for their own purposes.
One example might be where you have two rows of reference material T on your mount, and you prefix spots from one row as TA and the other as TB. That gives you the option of calibrating your unknowns relative to RM data from either one of those rows (by specifying the appropriate two-character prefix), but it also gives you the option of calibrating your unknowns relative to the combined RM dataset from both rows (by specifying the one-character prefix 'T'). Squid3 knows exactly what choices you made in this regard: it just needs to report those in the CSVs.