Several data items are no longer allowed to have s.u. values

[vaitkus]

Several data items were redefined to no longer allow associated standard uncertainty (s.u.) values when migrating from the DDL1 (version 1.0.1) to the DDLm version of the dictionary. This might have been done on purpose, but I will list these items here just in case since there were similar accidental changes in the CIF_CORE dictionary that had to be resolved. Data items in question:

• _pd_meas_2theta_fixed
• _pd_meas_2theta_scan
• _pd_meas_position
• _pd_meas_step_count_time
• _pd_meas_time_of_flight
• _pd_calib_std_internal_mass_%
• _pd_peak_2theta_centroid
• _pd_peak_2theta_maximum
• _pd_peak_d_spacing
• _pd_peak_width_2theta
• _pd_peak_width_d_spacing
• _pd_phase_mass_%
• _pd_prep_cool_rate
• _pd_prep_pressure
• _pd_prep_temperature

Should the _type.purpose attribute of any of these be changed from Number to Measurand and in this way allow the associated s.u. values?

[briantoby]

Some of the following are not likely to have a standard uncertainty reported, but I do not see why it should be precluded. They are all experimentally measured quantities or are derived from them.

On Oct 10, 2022, at 8:41 AM, Antanas Vaitkus @.**@.>> wrote:

Several data items were redefined to no longer allow associated standard uncertainty (s.u.) values when migrating from the DDL1 (version 1.0.1) to the DDLm version of the dictionary. This might have been done on purpose, but I will list these items here just in case since there were similar accidental changes in the CIF_CORE dictionary that had to be resolved. Data items in question:

• _pd_meas_2theta_fixed
• _pd_meas_2theta_scan
• _pd_meas_position
• _pd_meas_step_count_time
• _pd_meas_time_of_flight
• _pd_calib_std_internalmass%
• _pd_peak_2theta_centroid
• _pd_peak_2theta_maximum
• _pd_peak_d_spacing
• _pd_peak_width_2theta
• _pd_peak_width_d_spacing
• _pd_phasemass%
• _pd_prep_cool_rate
• _pd_prep_pressure
• _pd_prep_temperature

Should the _type.purpose attribute of any of these be changed from Number to Measurand and in this way allow the associated s.u. values?

— Reply to this email directly, view it on GitHubhttps://github.com/COMCIFS/Powder_Dictionary/issues/37, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ACH7E2AN6MB6RCVTTDIBLXDWCQMKBANCNFSM6AAAAAARBLRSOY. You are receiving this because you are subscribed to this thread.Message ID: @.***>

[rowlesmr]

• _pd_meas_2theta_fixed: possible
• _pd_meas_2theta_scan: unlikely, but shouldn't be precluded
• _pd_meas_position: possible
• _pd_meas_step_count_time: unlikely, but shouldn't be precluded
• _pd_meas_time_of_flight: unlikely, but shouldn't be precluded
• _pd_calib_std_internal_mass_%: probable
• _pd_peak_2theta_centroid: probable
• _pd_peak_2theta_maximum: probable
• _pd_peak_d_spacing: probable
• _pd_peak_width_2theta: probable
• _pd_peak_width_d_spacing: probable
• _pd_phase_mass_%: probable
• _pd_prep_cool_rate: possible
• _pd_prep_pressure: possible
• _pd_prep_temperature: possible

.

Some other items that are currently Number that could be Measurand: (with various likelihoods of actually having SUs)

• _pd_calib.2theta_offset
• _pd_char.atten_coef_mu_calc
• _pd_char.atten_coef_mu_obs
• _pd_char.mass_atten_coef_mu_calc
• _pd_char.mass_atten_coef_mu_obs
• _pd_proc.2theta_corrected
• _pd_proc.d_spacing
• _pd_proc.energy_detection
• _pd_proc.energy_incident
• _pd_proc.recip_len_q
• _pd_proc.wavelength
• _pd_meas.angle_chi
• _pd_meas.angle_omega
• _pd_meas.angle_phi

I can start a PR, probably tomorrow.

All _pd_meas.counts_* are excluded from consideration as a Measurand, as their SU is defined to be square root of its value.

[jamesrhester]

I agree that all these should be Measurand. I have a little tool which will add in the su data definitions for all Measurand items so if the PR simply has the Number->Measurand change I can run that tool and update the PR with su definitions.

[rowlesmr]

little tool

Neat! I've dabbled in Julia several years ago, but the compile-every-time-you-run-a-program didn't fit how I used my programs. I should have another look.

[jamesrhester]

Still quite slow because of the compile every time thing, but it is now possible to create a system image that includes your favourite packages already compiled and then it is quite fast.