I propose a panel that shows/hides at the top of the mass audit that provides two lists - available and viewed - with drag and drop between them.

It would be nice if (as is generally the case) not all spots are shown, to make the last entry an ellipsis: ". . ." to denote the missing data. Also, when both ref materials and unknowns are shown, add a blank line between the two lists.

saveAction should not call save() if nothing has changed:

1. editing existing expression
2. no name change
3. no change to excel expression string

Modifiers Panel will contain "%" and ±

The floating tooltip for an expression should also show the text of the expression.

Every time a new expression is selected in the list , a large number of items (20-ish) are processed with the updateItem method ... very inefficient. (line 2410)

Squid3 Version: 0.3.0
Java Version: 1.8.0_171
OS: Windows 7 6.1

Issue details:

• add DENOMINATOR and NUMERATOR to the axes of the ratio builder grid.
• Grey out the "expression" accordion when in edit mode.
• Add functionality to set the default folder when opening/saving the different files (.xml, .squid etc)

Activate this button to allow selection of videos and then open a browser to watch.

Change button label to "HowTo Videos"

The token "NUMBER" does not provide the facility to enter a number in the pop up menu.

Should blow them away and leave only the default 10.

Squid3 Version: 0.1.8
Java Version: 1.8.0_131
OS: Windows 7 6.1

Issue details:
The visualization of the equations in the "Graph" window is still shady on a PC. Assume this is a fialure to find an internet library?

Currently, remove expression does not choose another expression and the reesult is confusing in appearance.

Similar to #34 - we need to force font to Courier New for mono-spaced look. Thanks to @cwmagee for pointing this out.

To complement the existing SQUID 3.0 0.0.1 interface:

(1) Analyses aggregated by number of species measured, summing to the total number of analyses collected, the idea being to highlight the presence of any analyses that have a different number of peaks to the 'mode'. The demo XML corresponds to the trivial case i.e. [Analyses with 10 species = 114; total analyses = 114].

(2) Analyses aggregated by number of scans, summing to the total number of analyses collected, the idea being to highlight the presence of any analyses that have a different number of scans to the 'mode'. The demo XML corresponds to the trivial case i.e. [Analyses with 6 scans = 114; total analyses = 114].

(3) A measure of total session duration, in hours, calculated by subtracting the first (chronological) analysis date/time from the final analysis date/time. The idea would be to provide a simple flag for sessions which have been inadvertently merged. This occurs most commonly when an analyst fails to log in to the SHRIMP control software at the start of their analytical session, in which case their new data is simply appended to the end of the XML file from the previous session. (Down the track, it would be worth looking at a mechanism for 'splitting' an XML file into time-delimited segments to help address this issue. We currently do it manually with a text-editor, and it's one of the main reasons most users still prefer working with PD input, rather than XML: even with SQUID's dumb XML parser, manually converting one XML file into two is fraught!).

(4) A count of the number of analyses containing the SQUID 3.0-generated suffix '-DUP-'. There should never be many; it would just be to help the interested analyst track down and amend them all.

(5) A two-stage prefix-based audit of all analysis-labels in the file (summing to the total number of analyses acquired), dictated by a user-specified number of leading characters. The first stage would correspond to what you have already done: using the minimum-length prefix to identify the set of reference zircon analyses (and remove them from further prefix-based consideration), with a summary (in the case of one-character prefix in the demo XML) like [T = 23, others = 91, total analyses = 114].

The second stage would operate solely on the set of 'others', and usually a user will have a good idea of the length of their identifier-prefixes, so could choose the number of characters that suited them best: the main idea would be to identify occasional analyses where the prefix was mistyped at the time the spot was programmed for analysis (still a vulnerable manual process). For example, for a user-defined length of 1 character, the demo XML would show [M = 1, [character] O = 16, [numeral] 0 = 49 (being the combination of 097 and 099), 1 = 25, total 'others' = 91]. Increasing to two characters would give [M2 = 1, OG = 16, 09 = 49 (being the combination of 097 and 099), 10 = 25, total analyses = 91]... note that if one of the '097' prefixes had been mistyped as '079', for example, it would show up at this point, with 07 = 1, which a user would hopefully identify as being anomalous. In the demo XML, the 'useful length' of the prefix is three characters, which would give [M25 = 1, OG1 = 16, 097 = 24, 099 = 25, 100 = 25, total analyses = 91].

It doesn't address every potential problem (e.g. analyses of one sample mislabelled as analyses of another) but it will do the job on silly typos. Note also that although the primary purpose is to permit spot-label clean-up prior to data-processing, there might be potential to store the user-anointed prefixes for later use, when the user wants to divide up the processed data for sample-specific interpretation.

We need to implement a cross-cutting concern that records whether anything has changed at any level in the project and alert the user to the need to save when exiting a project.

The math is (1-sigma abs) / value * 100.

This will occur for ratios and NU-switched expressions only (for now).

Some acquisition tables have multiple background positions, as the background count rate can change at different places in the mass spectrum. For run tables with multiple backgrounds, users will need to nominate:
-Which background to use.
-Whether to interpolate between stations (advanced- not necessary in first instance).

Squid may want to:
initially assign nearest background to peaks AS LONG AS USER CAN CHANGE THIS later

We have plans for an ET_Redux-like set of models for parameters, but in the short term, se need an interactive window for setting values for a task. See me for instructions.

Appears in version 0.1.9.

When I open a new prawn xml file, the species are identified by their closest integer mass. This results in "duplicate" masses for species that are close to one another (e.g. 204 and 204.1 aka background). The masses in the prawn file are given to 6 decimal places, so it must be rounding and that is causing the problem. Please display the masses from the prawn file with 2 decimal places (if it were only 1 decimal place in some cases if it rounded up/down it might confuse masses, more than that is overkill).

Design, implement, and test this functionality for manual spot rejection in Weighted Means of Reference Materials.

Background

Per email to me of 10 May 2018 from @sbodorkos:
Re: weighted means, remember there aren’t all that many calculated in the automated processing (although maybe we’ll need to think about what defaults we want if people include calls to WtdAv in their Task equations). Note also that none of this applies to TukeysBiweight. All of the following relates only to invocations of WtdAv in different situations.

“Preferences”
The “Preferences”-style option for users to bypass auto-rejection procedures applies specifically to WtdMeanA calculations on the Standard (i.e. not elsewhere). Ludwig’s idea was that there ought to be no source of ‘geological’ variation in WtdMeanA, so that is the only place where auto-rejection could be uncritically applied (the corresponding situation in any given Sample, where geological variation is likely to be the primary cause of dispersion, is handled by FindCoherentGroup, a subroutine as yet untraversed).

Users requested the option of specifically ‘switching off’ the auto-rejection routines applying to WtdMeanA on the Standard, in recognition of the fact that there will always be a subclass of users who take such rejections as gospel, even while being ignorant of the reasoning (remember that SQ2.50 does not provide ‘rejection reason’, and I am really happy that SQ3.0 will change this). That measure was intended essentially to protect inexperienced users from themselves. (Ludwig’s philosophy here is odd: everywhere else in SQ2.50, he sets very stringent tests for outlier rejection, and commonly outliers are retained even when they look to the human eye as if they ‘should’ be rejected… the population defining WtdMeanA is a glaring exception, where his auto-procedures often result in ‘too many’ rejections!)

Note that the Preferences should apply equally to all potential WtdMeanA calculations e.g. in a Perm1 situation, where the range of potential WtdMeanA values spans 4corr206/238, 7corr206/238, and 8corr206/238, the Preference setting should be the same for all.

“Redo”
What we will need to deal with more broadly is the Redo functionality of WtdAv. Remember that WtdAv as implemented in Isoplot has a Boolean argument CanReject, and when this is set to TRUE, the Rejection ‘sub-subroutine’ (invoked via GOSUB in subroutine WeightedAverage) is enacted. For example, when WtdAv is invoked during a WtdMeanA calculation, CanReject is indeed set to TRUE, and the Rejection sub-subroutine is that responsible for the second round of auto-rejections (the first round being the ‘large error’ rejections independently undertaken beforehand).

But irrespective of the CanReject setting of any initial WtdAv invocation, a “Redo” occurs after an initial WtdAv has been calculated, and after the user has had to that chance to modify any auto-rejections (or FindCoherentGroup rejections) that might have happened. As a consequence of these things, when WtdAv is invoked via “Redo”, it is essential that:

(1) Auto-rejection routines (as applying to WtdMeanA; see above) must be bypassed completely

(2) CanReject must be FALSE

The combined effect of these two measures is to force WtdAv, when invoked via “Redo”, to perform its calculations solely on the ‘non-rejected’ set of analyses specified by the user.

The schema needs the notes field added and also the list of functions and operations updated to match the names in the Function and Operation class.

It seems that in this snippet from save() lines 2170-2175, that the removeExpression call should only occur twice if the the expression names are different?

        //Remove if an expression already exists with the same name
        task.removeExpression(exp);
        if (currentMode.get().equals(Mode.EDIT)) {
            task.removeExpression(selectedExpression.get());
        }
        task.addExpression(exp);

I added a new hidden operation "Value" that should not be visible to the user.

Squid3 Version: 0.1.4
Java Version: 1.8.0_131
OS: Windows 7 6.1

Issue details:
After editing duplicate spot names, all other spots starting with that initial letter are not selected when filtering by prefix. EG if two of the 28 Tem-X grains are renamed, then they are the only ones which show up when filtering.
100157.squid.zip

Provide facility to choose to display any number of differences between two masses. I propose a scroll box containing add and remove buttons and combo boxes to make selections per pair of masses.

Plots of time vs analytical conditions info from xml file. Primarily it would be nice to see a plot of time vs of primary beam intensity (nA) throughout the session as an option in Audit. There are likely other parameters too, like counts per second of essential elements (like Zr 196 in zircon).

Introduced in previous bug fix making spot names case insensitive ... need to detect duplicates also with case-insensitive filter.

To date, only a few tasks have been investigated and as a result many expressions are not recognized and some may throw exceptions not anticipated, hence the expressions window failing to load. In addition, many expressions from imported tasks are intended to be ignored because they are really about output formatting. I concur, as a start, (per @cwmagee 's suggestion) we specify a set of say 10 tasks for a 10-peak analysis and walk through each task, deciding how Squid3 will handle each expression. We will find that there are functions not yet ported from LudwigLibrary and that there are naming conventions to be handled, etc. This will be a tedious learning process for Squid3, but we should start. Step 1 is to select these ten tasks for initial analysis and proceed to decide how to handle each contained expression. Step 2 is to implement and test handling these expressions. Step 3 is to repeat with more. We should consider a Google drive folder fo the tasks we are considering. Thoughts - @cwmagee and @sbodorkos ?

Audit of masses is good. Would also like to see audit of offsets between masses, for example the difference in the magnet mass value between 196 Zr (an autocentred peak) and 204 (non auto-centred). This should stay constant but if ONE of these drifts off mass it is easier to spot in the offset rather than the straight mass. I could provide a list of common offsets that we check, but ideally we could have some sort "offset builder" and then those are shown in an audit plot.

I opened one of our existing Prawn xml files and a accompanying task. When I inspect the task, there are 2 204corr206Pb/238U age equations under "Built in expressions". One expression (highlighted in yellow in the screen snap attached) is empty (returns 0 in the viewer). The other (which is selected in the screen snap) does return an equation but the result in wrong. The problem is that the denominator should not simply be lambda238 but lambda238*10^6 in order to return the age in Ma.

I haven't tried this experiment with one of your test files. Perhaps the issue is the parsing of my SQUID2.5 task but wanted to flag.

Squid3 Version: 0.1.7
Java Version: 1.8.0_131
OS: Windows 7 6.1

Issue details:
Help -> Online resources -> Github repository bug does not open anything in a browser. It doesn't do anything at all. Maybe a GA firewall issue?

Allow use of control key and mouse click to select any set of nodes (not contiguous) much like a multi-select list behavior and then provide option to delete, copy, move, etc.

As discussed today, I would find it more intuitive if the order of the nested options in the Report menu was:
Report --> Reference Material/Unknowns --> Within spot results/Spot average results (not exactly sure if "results" is the right word, but I find "ratio" a bit non-intuitive because these aren't just reports of ratios they are reports of the output of every expression.

Squid3 Version: 0.1.4
Java Version: 1.8.0_131
OS: Windows 7 6.1

Equation font size and line positions don't seem to match up.
Issue details:

@sbodorkos noted that Ludwig provides a function to calculate the ratios from a date - we need to determine where it is, implement it in LudwigLibrary, and make it available to the user in a future preferences menu.

Currently, importing a Squid2.5 task produces two temporary custom expressions: DUMMY and DUMMYsum, both with the value of "5" but DUMMYsum has the summary switch on. When you open the Custom Expressions accordion and click on DUMMY, then click on DUMMYsum, nothing changes in the Peek window. If you click on another expression and then on DUMMYsum, it is correct, but clicking on DUMMY has no effect. This behavior does not exist in the deprecated Expression Manager. We need to determine and fix the cause in Expression Builder.

Here is a snapshot of the error message that I get for a new .xml file. I will email you the actual xml file in case this isn't adequate to know exactly what change is need in the schema

Note: this will apply to ratios and NU-switched expressions only (for now)
Note: % = per cent unct and ± = absolute unct

Two strategies:

1. add to context menu the choices of inserting these symbols in, for example: ["204/206"]

2. add a new accordion panel "Modifiers" with these two symbols.

I favor choice 1.

1. Allow for editing, removal, export, loading, etc of "NU-switched", "Built-in", and "Unhealthy" expressions. Keep Ratios constrained.

2. Provide enabled NU-switch checkbox next to Summary Calculation checkbox - remember this field is Expression (not ExpressionTree) specific

3. Enable Summary Calculation checkbox

Specs: a new project menu item for either using a single prawn file or for joining two prawn files. Provide a context menu in the project manager's 'Manage Spots' tab for the left-hand list of all spots
to select splitting the runs at the selected run (second half begins with selected run) using the visible edited data with duplicates noted, or the original data without edits or duplicated noted.

Squid3 Version: 0.1.7
Java Version: 1.8.0_131
OS: Windows 7 6.1

Issue details:
It would be the axes ought to be labelled "numerator" and "denominator". Also, selecting an entire row or column by clicking the label would be handy. The display will have issues with run tables were decimal mass stations are used, unless some sort of auto scaling can be done.

1. Currently, an expression with the same name as an existing expression can be loaded - thus if you export an expression and then load it, there will be two expressions with the same name. We should have a warning to the user and allow them to cancel or replace or rename.

2. Also, once loaded, the focus returns to the deprecated Expression Manager instead of to the new Expression Builder. The focus should always go to the Expression Builder and then also highlight the newly loaded expression.

The popup tooltip in drag and drop mode for various nodes should show the peek contents if the element has the "summary flag" set and the top 10 spots if in detail mode.

Complete project files management - MRU, user-specific preferences, etc.