General idea

It should be possible to define penalties for matches for a language resource. These penalties will then be subtracted from all matches.

So if a context match is usually a 103% match, with a penalty of 4% it would become a 99% match.

In addition there should be a penalty, if the source or target sub-language is different.

Implementation in the task to language resource assignment

The task to language resource assignment is located in the import wizard and at the individual task.

At both places there will be 2 new columns in the grid, that list the assignable language resources. Some styling shows the user, that the content of the columns is editable inline

• First new column:
  • Choose an icon for the column.
  • Tooltip:
    • DE: Abzug: Dieser Wert wird für die Analyse von der Treffer-Rate der Sprachresource abgezogen.
    • EN: Penalty: This value is subtracted from the language resources match rate for the analysis result.
  • Purpose: Allows to select a penalty for the language resource in question. Its value will be deducted from the match values of the analysis for this language resource (developer note: Should be saved in a separate location and deducted on the fly).
  • The default for the penalty will be zero as a number.
  • The default can be set differently on client level (see below).
• Second new column
  • Choose an icon for the column.
  • Tooltip:
    • DE: Sub-Sprachenabzug
      Dieser Wert wird für die Analyse von der Treffer-Rate der Sprachresource abgezogen, FALLS sich die Subsprache der Aufgabe von der Subsprache der Sprachresource unterscheidet.
      Beispiele:

• • • • Ausgangssprache Sprachresource de-DE, Ausgangssprache Aufgabe de-AT: Der Abzug greift.
      • Zielsprache Sprachresource en-GB, Zielsprache Aufgabe en-US: Der Abzug greift
      • Ausgangssprache Sprachresource de, Ausgangssprache Aufgabe de-AT: Der Abzug greift NICHT
      • Zielsprache Sprachresource de-DE, Zielsprache Aufgabe de: Der Abzug greif NICHT
      • etc.
    • EN: Sub-language penalty
      This value is subtracted from the language resource match rate for analysis IF the task's sublanguage is different from the language resource sublanguage.
      Examples:
      • Source language language resource de-DE, source language task de-AT: The deduction is made.
      • Target language language resource en-GB, Target language Task en-US: The deduction is made
      • Source language language resource de, Source language Task de-AT: The deduction is NOT made
      • Target language language resource de-DE, Target language Task de: The deduction is NOT made
      • etc.
    • Purpose: as the tooltip explains it (developer note: Should be saved in a separate location and deducted on the fly).
    • The default for the penalty will be the highest possible match-rate (currently 104, please be aware, that soon this will change) as a number. Still the language resource will be offered for assignments for tasks, even if the sub-languages do not match (what currently is not the case)
    • The default can be set differently on client level (see below).
• In both column as value it should only be selectable something between 0 and the highest possible match rate (the later should not be hard-coded, as it will be extended in the future),
  • changed values should not be saved on leaving the the cell-editor but instead - saved only when 'Start analysis' is clicked, i.e. new penalty values will be saved into db and take effect on match rate only when analysis re-run is started. Also, red text 'Analysis should be re-run for penalties to be saved and take effect' should appear (in the the bottom toolbar where 'Start analysis' button resides) since any kind of penalty is changed for any of assigned language resource and until button is pressed

• The following should only be implemented, if it does not require further messing changes to MatchAnalysis implemention. If it does, it should be moved to the issue for refactoring the MatchAnalysis:
  If a penalty is higher than
  (HighestPossibleMatchRate - runtimeOptions.LanguageResources.TM.matchPanelMatchrate)
  then it will not be possible to select such a memory as language resource to avoid unnecessary calls to that memory. If it would be selected due to default assignments, it will still not be selected. Some good mechanism will prevent its assignment in wizard and task-lang-assoc-grid and some good info message will explain the user, why

Implementation in the analysis

For each language resource a new column "penalty" will be introduced, that shows the combined penalties of general penalty and sub-language penalty that was applied and that resulted in the analysis for this resource.

Implementation in the clients management

A new tab "language resources" is created in the right panel, when a client is selected in the client management.

It contains a grid, that lists all language resources.

Above the grid there is a button "Only list assigned". When active, only the language resources assigned to the current client are shown. This button is active by default.

If the user only has clientPM roles, she/he only sees the language resources of the clients, assigned to him. If this is only one client, the previously mentioned button is not visible.

If client PM has has access to Clients tab but not to Language resource tab -> show language resource tab here anyway

As in the language resource overview, the grid is grouped by language resource type.

The grid has the following columns (each column should be filterable):

• Id
• Name
• Source language (icon like in the language resources management)
• Target language (icon like in the language resources management)
• Usable for client (check-box)
• Read access by default  (check-box)
• Write access by default  (check-box)
• Pivot by default (check-box)
• Penalty default (Number drop-down with values 1 to 104) (by default set to 0)
• Sub-languages penalty by default (Number drop-down with values 1 to 104) (by default set to 104)

The values in the grid are read from and written to the same database tables as those in the language resource management. Except for the "penalty by default", which is only available in the clients management.
Please make sure, that if values in client overview tab are changed, values in the language resource overview will also be reloaded and therefore up2date. And vice versa.

Implementation in the opened task

Penalties should become available in the segments grid as two separate 'Penalty General' and 'Penalty Sublang' columns, added after 'Match rate'-column, but both kept hidden by default despite sortable and filterable when shown. Values in those two columns should be calculated during the match-analysis process, and both values should be deducted from the original match-rate value before it's saved into database for a segment, so that value in 'Match rate'-column will be the result of such a deduction.

For the Matches-tab at the bottom of the segments grid - values for the Match-rate column should now be shown as: "X (=Y-A-B) [C]" where:

• X - the resulting match-rate
• Y - original match-rate with no penalties deducted so far
• A - General penalty
• B - Sublang penalty
• C - language resource name