Evaluation Runs 

Use

An evaluation run creates a global worklist for each target language. The following activities are performed:

·        The domain assignment of objects is renewed

·        The objects in one or more object lists are analyzed for each target language, according to the object types defined per target language

·        The translation status is calculated for each line of every object per target language

·        The global worklist and statistics are created/updated for all users and target languages in the system

Integration

Evaluation runs can be started in transaction LXE_MASTER on the Evaluations tab page.

 

Prerequisites

·        Your translator profile contains the authorization object WRKL with the value 01.

·        The translation environment has been set up.

·        An object list has been created.

Features

Evaluation runs are started after one or more object lists have been created.

You can set the following parameters:

·        Distribution: during evaluation runs, proposals in the proposal pool can be inserted into translation lines automatically, depending on their quality. The source texts are given the status translated and do not need to be translated manually.

Texts are automatically distributed as follows:

¡        For new, modified, and/or translated source texts if distribution has been activated for new, modified, and/or translated source texts

¡        In source texts for which a proposal exists whose quality status is at least as high as the quality status defined for automatic distribution.

¡        In object types defined for distribution.

¡        In collections assigned to a translation graph

For more information, see Automatic Distribution.

·        Initial Statistics: initial statistics can be generated before the global worklist is created. The current translation status of objects in the existing global worklist is calculated for each language. Objects that may still need to be added to, or deleted from the worklist are not taken into account.

·        Statistics: after the global worklist has been created, statistics can be generated for all of the evaluated target languages. The current translation status of every object in the current global worklist is calculated for every target language. Objects that were added to, or deleted from the worklist since the last evaluation are taken into account.

·        Deletion: before a new global worklist is created, the old global worklist can be deleted or retained.

¡        If the old global worklist is deleted before the new global worklist is created, the objects in the existing global worklist are removed for the selected target languages. Objects are only included in the new global worklist if they are contained in the object list used for the evaluation. Deletion is cross-client. This means that if you translate objects from different clients in one global worklist, all objects are removed from the global worklist that are not included in the client in which the new global worklist is created.

¡        If you do not delete the existing global worklist before creating the new global worklist, new objects are added and the status of objects in previous personal worklists and statistics is updated. If objects in the old global worklist have been deleted in the system since the global worklist was created, they remain in the personal worklists. When calling up such objects, the translator is provided with the following message: “Empty List”. If you want to reduce the number of objects to be translated into a particular target language by locking an object type, for example, you must create the next global worklist with deletion. Otherwise, the objects belonging to the object type that is no longer defined as translation-relevant will remain in the global worklist.

·        Renew Domain Assignment: each object belongs to a domain that determines how terminology is assigned in the proposal pool. During a development project, new domains can be created or objects can be moved from one domain to another. To ensure that application standards can be created for the correct domain in the proposal pool, you must update the assignment of domains to objects regularly.

·        No Status Calculation: if the translation status is of no interest to you because, for example, you merely want to delete an existing worklist without generating a new one, you can start an evaluation run without calculating the translation status.

·        Object Lists:evaluations can be performed for one or more object lists.

·        Object Types:evaluations can be performed for all object types, or a selection of object types.

·        Graphs:evaluations can be performed for all translation graphs, or a selection of translation graphs.

·        Target Languages:evaluations can be performed for all target languages, or a selection of target languages.

Evaluation runs enable you to create several global worklists that you can separate one from the other by using worklist numbers. This enables you to keep several projects (that have different translation-relevant objects) separate in the same translation system. The worklist number is also the starting point for translators who want to reserve worklists.

Make sure you communicate the number of the worklist(s) containing objects relevant for translation to all parties involved in translation.

An object can only appear in one global worklist at a time. For example, if a specific object matches the criteria of two global worklists in the system, the last worklist to be created will contain the object.

Activities

For information on how to schedule evaluation runs, see Creating an Evaluation Run.

You can monitor the progress of the current evaluation run in transaction LXE_MASTER.