Access Keys:
Skip to content (Access Key - 0)

Application 2.2.3 Tool 5.3 IBFB Feedback Alpha 2.0

IB FieldBook V2 Issues04/02/11

  1. The fieldbook opens with the tabular display of fieldbook components even though none has been specified yet. It should open with the introductory page with a few instructions.
  2. The other treatment factors page:


It is good that you have now identified the Other Treatment factors, but note that FERTAM is a label of FERTLE and not a factor in its own right. The first thing we need to do here is ask the user for the number of levels of the factors. Your table is ok for this but the last column should be headed "Number of Levels". (Also as it is now it will not let you type a number in the cell you have to use the dial). The next thing we need to do is ask the user for level values of any labels of any of the factors. The labels (like FERTAM) should not have a number in the 'Number of Levels" column, but rather a button with the caption "Specify Label Values" This could cause a table to pop up for any Factor which has one or more labels:

FERTLE FERTAM
1 Low*
2 Med
3 High
  • The user could enter the Fertalizer amounts here if they are known at this time or they could be left blank to be filled in at the time the trial is implemented.
  1. I suggest you move "Select Germplasm" up to position 4 in the wizard workflow. If feels more natural there.
  2. There is some problem with the assignment of columns from the Germplasm List to the Fieldbook labels. The labels for germplasm in the template are:
    ENTRY ENTRY NUMBER GERMPLASM ENTRY NUMBER ENUMERATED
    ENTRY CD ENTRY CODE GERMPLASM ENTRY CODE ASSIGNED
    DESIG ENTRY DESIGNATION GERMPLASM ID DBCV ASSIGNED
    GID GERMPLASM ID GERMPLASM ID DBID ASSIGNED

    This means that in the germplasm table:

    I am expecting four columns headed ENTRY, ENTRY CD, DESIG and GID. Because the scale of ENTRY is NUMBER I am expecting you just to number the entries 1,2, … to the number in the germplasm list (ie not to use any valye from the list, just number them). In the entry CD column I am expecting you to put the corresponding values from the ENTRYCD column of the germplasm list, not because the names are similar, but because the scale of ENTRY CD in the template is CODE. Desig and GIS are filled correctly, but again not because the names match, but because the property and scales match according to the table in the specifications. If you don't follow any of this please ask.

  3. Generate Design. We must simplify this screen. Please read carefully my notes in the specifications in step 8 on this page:

http://wiki.cimmyt.org/confluence/display/MBP/Objective+2.3+IBFB+Flow+Diagram
If you don't like my mockups we can go with yours but it must be simpler.

  1. remove the Randomization box – it makes no sense. Also remove the Unreplicated designs box and the Replicated Designs box since the distinction is artificial (Design IIs are usually replicated for example).
  2. remove Augmented Design, Develop Diallele, Develop Design II from pair.. and Develop Design II, group … Also remove Mother baby trial. Since we are not going to provide any of these in version 1.
  3. So we will have just four entries to choose from:
    1. Unreplicated design without randomization
    2. Unreplicated design with randomization
    3. Randomized complete block design
    4. Alpha lattice design

Only include d if you think you are able to link the design generator algorithm into version 1. The others only require you to be able to randomly permute a sequence of integers and I assume Java has a facility to do this.

  1. The Design box should be labelled Design Parameters. The number of sites should be indicated first, and "Sites" and "Entries" should be replaced with a little table:
    Factor Number of Levels
    TRIAL 3
    ENTRY 20
    FERTLE 2
    PDATE 2

    If there are no "Other Factors" in the template then of course this table only has the header, the TRIAL and the ENTRY line. This table is not editable since the numbers have all been specified earlier in the wizard. (Note you use the names of the factors from the template so we get away from the problem of using generic names like Sites and Entries if these are not the ones the user wants.

  2. Then, if the design is c or d above we have an entry box for "Number of Replications" and if the design is d. We have another entry box for "Number of plots per Block" Finally for any randomized design we have a entry box for "Random Seed" and you fill it with the current number of milliseconds from the system clock, but you allow the user to change it if he wants.
  3. What comes back from the Design Generator is a table for each trial instance with a column for field plot number and for each design parameter – Rep, block, (if they are there), entry and other treatment factors. These need to be mapped to the appropriate labels in the template.
  4. In the tabbed view the Factors tab is not correct: Firstly, we have decided to call it "Other Factors" or "Other Treatment Factors" and of course it should only exist if there are other treatment factors. It should also come after the Entries tab.


Thee table that is shown here has to be constructed as the cross-product of the Other Treatment Factors so it should look like this:

FERTLE FERTAM PDATE
1 Low 1
1 Low 2
2 Med 1
2 Med 2
3 High 1
3 High 2

It has one row for each other treatment factor combination. The number of other treatment factor combinations is the product of the number of levels of each other treatment factor. The user could be given the option to remove rows but let's leave that for Version.

  1. The Measurements Table must be constructed from the design for each trial instance.
    1. First find the number of plots in each trial instance. (In version 1 we are assuming each instance has the same design so they will have the same number of plots). The number of plots in trial i (Pi) is the "number of entries in trial i" X "number of other treatment factor combinations in trial " X "number of replications in trial i". Construct a table where the first column is headed by the Trial Instance Factor, and it has one row for each plot in each trial instance with the number of the instance in column 1 ie P1 1s, P2 2s and so on. You can add columns for each of the Trial Conditions with the appropriate values or blanks if they have not been specificed.
    2. Next construct unique plot numbers for each Trial Instance. First find k = Int(Log10 (Max(P1, P2, …Pm)))+1 (where Int is a truncation function, not a rounding function) then the plot number for plot j in trial instance i, pij, is i*10**k +j. (This is the meaning of the Nested Number Scale). These numbers are entered in the next column of the workspace table under the heading of the Field Plot Factor. Add columns for all the design parameters by linking the plot number (j) to the design table for trial instance i as described in 5f above – using the label names from the template.
Adaptavist Theme Builder (3.3.3-conf210) Powered by Atlassian Confluence 2.10.3, the Enterprise Wiki.
Free theme builder license