Comparing GlobalSight with WorldServer
This is an objective comparison of the Translation Management Systems GlobalSight and WorldServer.
- 1 Key concepts
- 2 High-level overview of the GlobalSight configurations
- 3 Translation Memory
- 4 Leverage penalty
- 5 Translation Memory Group
- 6 Terminology Database
- 7 Segmentation rules
- 8 Supported filters
- 9 Filter configuration and customizations
- 10 Locales and encodings
- 11 Scoping and Cost Model
- 12 Business Rules Engine and Scheduler
- 13 Workflow
- 14 CMS Integration
- 15 Off-line translation
- 16 Custom development
- 17 Machine Translation integration
- 18 Reporting
GlobalSight and WorldServer mainly differ in their approaches to how the systems create jobs, integrate to content management systems, configure settings, and process data.
The following is high-level comparison of the two systems with regard to these.
|Job creation||Push content from repository||Pull content from repository|
|CMS integration||Web Services||CMS clients|
|Configurations||File, L10N, and TM Profiles||AIS properties|
|Data processing||A job may contain multiple files||Each file is associated to an independent process|
High-level overview of the GlobalSight configurations
In GlobalSight, many configurations are set in File Profiles, L10N Profiles, TM Profiles, and Workflow. Users can choose the desired settings by selecting the appropriate File Profile.
|Assignment||Selected through File Profile > L10N Profile > TM Profile||Set an AIS-Property|
|Penalty options||Set in TM profile||Most penalty options are set in tm.properties file|
|Attributes||No custom attributes||Has custom attributes and available for TM leverage|
|Stemming||No indexing on stems||Has indexing on stems|
|Import/Export||Support TMX 1.4b||Support TMX 1.4b|
|Context Match||In Context matches||ICE matches|
|Unique ID Match||Segment ID-based matches for XML and Java Properties content||SPICE matches|
The following leverage penalty configurations are available in GlobalSight.
The type (e.g. URL, Text, Alt) of a new segment is compared to segments in the TM during leveraging. If two segments differ by type, the fuzzy match percentage is reduced by the amount in the Penalty box.
The case of a new segment is compared to segments in the TM during leveraging. If two segments differ in case, the fuzzy match percentage is reduced by the amount in the Penalty box.
Whitespace in a new segment is compared to segments in the TM during leveraging. If two segments differ in whitespace, the fuzzy match percentage is reduced by the amount in the Penalty box.
Code, such as formatting, in a new segment is compared to segments in the TM during leveraging. If two segments differ in code, the fuzzy match percentage is reduced by the amount in the Penalty box.
Multiple Exact Matches
In cases where there are more than one unique translation for an exact match:
- Latest: The most recent translation is selected as the exact match
- Oldest: The oldest translation is selected as the exact match
- Demoted: Multiple exact matches are displayed as fuzzy matches. The match percentage is reduced by the amount specified in the Penalty box
Consult the WorldServer Translation Memory Administration Guide (the “TMAG”) for penalty configurations in WorldServer. Most of the penalty settings are set in the tm.properties file.
Translation Memory Group
GlobalSight allows multiple reference TMs. The setting is configured in TM profiles. However, there is no hierarchical order on the reference TMs.
|Embedded in TM Profile||Separate TM Group feature|
|Has hierarchical order on reference TM||Has hierarchical order on reference TM|
GlobalSight does not have Terminology Group.
|Has Terminology Fields||Has custom TM attributes|
|No TD Group||TD Group|
GlobalSight support SRX 2.0. Additional segmentation rules can be configured.
|Support SRX 2.0||WorldServer-specific|
|Any number of TM Profile-specific segmentation rules configured in Segmentation Rules and referenced in TM profiles||One set of system-wide segmentation rules configured in SentenceBreaker.properties files|
The following are the Out-Of-the-Box filters that are provided by GlobalSight and WorldServer.
|Frame 5||OOB||Use MIF Filter|
|Frame 6||OOB||Use MIF Filter|
|Frame 7||OOB||Use MIF Filter|
|InDesign||OOB||Filter on INX file using XML filter|
Filter configuration and customizations
Users can configure the XML, SGML and Word filters in GlobalSight. In addition, users can create custom scripts to control filtering rules using external processes.
|Customization||Custom script to control filtering rules using external processes||Using WorldServer SDK|
Locales and encodings
In GlobalSight, encoding is not an attribute of a locale. In WorldServer, each Locale is associated to a specific encoding.
|Source encoding is specified in File Profile Default target encoding is specified in Workflow (can be overridden during a manual export)||Encoding is associated to Locales and Locales are assigned to assets as AIS Properties|
|Users select the correct default source and target encodings by selecting the appropriate File Profile. User can override the target encoding during a manual export.||Users select the correct encoding by selecting the appropriate Target Locale|
Scoping and Cost Model
GlobalSight and WorldServer have different scoping and cost model configurations. However, both systems provide roughly equivalent functionality for calculating translation charges in advance. In addition, GlobalSight also has functionality for calculating actual cost as the work is performed (particularly useful for LSPs).
|Fixed number of match ranges and range values (Same as Trados Workbench)||Configurable number of match ranges
and range values. Mhough most users configure their scoping ranges to match the Trados ranges
|Rates are configurable at each Activity (translate, review, DTP, etc.). Most users apply all cost at the translation activity.||Not configurable at Step-level (Activitylevel)|
Business Rules Engine and Scheduler
GlobalSight has a Watched Folder system. Custom code can be built to monitor the Watch Folder to create jobs automatically.
|Has Watched Folder||Has Business Rules Engine|
|No scheduler||Has schedule (Business Rules Engine)|
GlobalSight does not have programmatic steps or Activity in workflow.
|No programmatic steps (this capability is under development)||Has Automatic Actions (extensible with SDK)|
|Can duplicate workflows for different locales||Each workflow can be applied to any locales|
GlobalSight provides WebServices API for CMS integration. In addition, GlobalSight has been previously integrated with Interwoven Teamsite and Documentum.
|WebServices integration available||WebServices and Custom connector API available in SDK|
|Out-of-the-box connectors for CVS (both client- and server-side integrations). Interwoven TeamSite and Documentum have been previously integrated||Out-of-the-box connectors for file system, SQL database, Broadvision, ClearCase, CVS, Documentum, Interwoven Teamsite, Perforce, Visual SourceSafe, and X-Hive database|
|Push methodology to allow content owner to trigger jobs||Mostly pull methodology allow localization manager to trigger projects|
Users in GlobalSight can translate offline using Bilingual Trados RTF. GlobalSight will support XLIFF in a future release.
|No desktop workbench (this is under development)||WorldServer has Desktop Workbench|
|Bilingual Trados RTF||Can export Trados TTX files|
|XLIFF 1.2||Can export XLIFF files|
GlobalSight is an Open Source application. The source code is available to developers who would like to develop custom components or add-ons to the system. GlobalSight also provide WebServices API.
|Open Source with WebServices API||Has SDK|
Machine Translation integration
GlobalSight currently does not have integration with machine translation engine. However, integration with machine translation engines is a high priority on 2009 roadmap.
|Integration with Google MT and PROMT||Integration with Babel Fish, Systran, and Language Weaver|
GlobalSight is an Open Source application and the data schema is open for developing reports.
|Has about 2 dozen pre-defined reports. Data schema is open for report writing.||Has Report Center|