Earlier release notes can be found in the archive.
The new version of Memsource Translate comes with several updates. You can now personalize Memsource Translate by adding your own generic and customizable engines from the list of supported engines. Furthermore, Memsource Translate now selects the optimal engine for a given domain as well as language pair for jobs with English source.
Supported generic engines: DeepL, Human Science, NpatMT, Yandex
Supported customizable engines: Google AutoML, Microsoft Custom Translator, Systran PNMT
It's possible to update a file previously imported to Memsource and have it pre-translated from project TMs. The older version of the same file is overwritten and cannot be restored. If a file submitted via automation widget is updated, the updated version will be delivered to the submitter. Further details are available here.
Project Manager and Administrators can now manually recalculate word counts when a file is updated and the analysis becomes outdated. An outdated analysis has a warning icon with a tooltip in front of its name. Further details can be found here.
Project Managers and Administrators can now change the vendor in project templates
Alexa Translations A.I. MT engine now supports batch pre-translation
Jobs Table: Hovering over the ellipses ( ) reveals the connector and source path of jobs imported from a repository.
When browsing a repository, entities are grouped by content models.
Support for Media Asset translation.
GIT: Automated Project Creation can now export translations to a different branch.
Some .DOTX files (MS Word template) contain broken content control links after translating. This has been fixed.
Zendesk Connector: Dynamic Content is not available for import. This has been fixed.
A new API endpoint allows the pre-translation of jobs in the case Copy Source to Target where empty is required.
The Automated Project Creation page now has a Check now button allowing users to immediately check the status of the monitored repository.
Marketo connector: Localized segmentation is now supported. Marketo segments names can be mapped to Memsource languages.
Zendesk connector: A label-based workflow functionality has been added.
Zendesk connector: An option to log debug information has been added.
Marketo: Segmentation is now supported for Emails and Landing Pages.
Zendesk: Authors can control when articles are ready for translation with new T and options in the connector configuration.
Hour is now available as a billing unit when creating a Quote and is represented in reports.
API: New parameter allows the exclusion of preview skeletons from MXLIFF files.
Filtering of multilingual CSV files was not working. This has been fixed.
XML files imported as DITA are appended with
xmlns:ditaarch. This has been fixed.
Using the URL instead of the host address when creating a connection to PangeaMT causes an error. A tooltip has been implemented to warn users of the difference.
Aligning documents for a TM is timing out or ending in an error. This has been fixed.
Zendesk: New target languages added via Zendesk Guide are added to the connector from the Guide module instead of the Support module. This enables dynamic synchronization with the connector.
Zendesk Connector can specify that jobs are only created for articles that are flagged for translation.
More project metadata filters (Owner, Client, Domain, Subdomain, Cost Center, Business Unit and Note) have been added to the Project Templates API.
Ubiqus NMT is now available as a Machine Translation Engine.
POPUP element is not imported for translation from an RC file. This has been fixed.
REST API: If a job is created without specific import settings, Global import settings are not applied. This has been fixed.
Project analysis settings gets overwritten with settings from latest analysis settings. This has been fixed.
Hidden text in MS Word files is being imported when it is specified not to be. This has been fixed.
Zendesk Connector: Draft articles are being pushed to Published instead of Draft. This has been fixed.
LQA data from Analytics can now be provided by the Analytics API.
LIST Last Login Date endpoint added to Memsource APIs.
Mirai MT setup includes additional Mirai MT optional configuration fields.
Shared drives are now supported by the Google Drive connector.
Contentful (field-level) Connector provides an optional Fallback language.
Various UI improvements.
Cost centers are not propagated from Project Templates when created via the Automation Widget or Automated Project Creation. This has been fixed.
Selection of anfor the Contentful (field-level) connector was not working in the UI. This has been fixed.
Automated Project Creation not pulling the latest files when connected to a Git repository. This has been fixed.
Amazon Translate now available for Memsource Translate.
Project page has a new filter allowing filtering of projects by status.
PO files now have options for defining translatable and non-translatable inline elements on upload.
IP address ranges specified for access are now validated before being saved.
Debugging log option added to Contentful connector configuration.
Adobe Experience Manager Plugin now accepts digital assets for translation.
SDL BeGlobal Machine Translation Engine has been updated to support the latest version of the Cloud MT API.
Brunei Malay language now supported.
For newly created organizational accounts default options have changed to increase efficiency.
ON by default:
Pre-translate on job creation
Set project to Assigned once all jobs are Accepted
Set project to Completed once all jobs are Completed
Magento Connector has been updated (1.0.6). This updates helps with connection problems to Magento version 2.3.
Job fails due to API limit. This has been fixed.
Entry-Level - restricted text values aren't pulled resulting in missing content. This has been fixed.
Adobe Experience Manager Connector:
In-Context Preview not generated if multiple connectors are configured that only differ in port number. This has been fixed.
Dropbox, FTP, BitBucket:
Changes in Continuous Jobs are not being automatically reflected in Memsource. This has been fixed.
Project owner is now visible in the Jobs View and can be used as a filter.
Project ID is now a customizable column and can be used to sort and filter.
PM no longer creates or can see another user's password. Password management is only in the hands of the user.
When batch importing users, passwords column is not required and ignored.
New languages added to Memsource
Translation requests are now based on 'last edited' date of content instead of 'last updated' date.
New Languages Supported in Amazon Translate
Connectors with open continuous jobs can be deleted with no warning. This has been changed and a warning is displayed.
Memsource employee names could be found in the properties of exported bilingual .docx files. This has been fixed.
Pre-translations exported to bilingual .docx contain incorrect fonts and languages. This has been fixed.
With trusted network configured on Access and Security, not all jobs are accessible to view in Memsource Editor for Web. This has been fixed.
Search Shared Space was not working for projects with shared jobs. This has been fixed.
Stored default paths in GitLab are incorrect. This has been fixed.
Problem with Magento Connector not making connections. This has been fixed.
Contentful Connector (field-level) jobs failing due to API limit handling. This has been fixed.
Linguist Portal Job View: Translation jobs assigned to the linguist are presented in a table with acceptance buttons and filtering options.
Added support for npatMT MT engine.
Added support for Kodesha MT engine.
Search shared space functionality improved to increase search speed.
Added support for Adobe Experience Manager 6.5
If Modify TBs is disabled for Guest users, View TBs and Approve terms functions do not work. This has been fixed.
Analysis created via API ignore the naming macro. This has been fixed.
Trusted IP Addresses can now be assigned to specific user roles.
GitLab online repositories were not maintaining original states or the full path to the repository was altered. This has been fixed.
If 'Include Tags' is selected for Google AutoML, parts between the tags are translated instead of whole segments. This has been fixed.
Have you come across any bugs?
Thank you for letting our Support team know.