I recently completed a large project using MemSource. My tool of choice is SDL Trados Studio 2011 and although I didn’t like having another tool imposed on me by my project manager, I was pleased to have the chance to put MemSource through its paces.
Briefly, MemSource is a TEnT (Translation environment tool) that stores all resources (projects, translatable files, translation memories, termbases, etc.) in the cloud. The translation process is completed in a standalone Editor installed on a local machine. I used a “linguist” account, which was the licence that my agency gave me to work on this particular project.
My function in the project was to review the translation of 55 files (a total of 60,000 words) done by two translators. I was up and running in no time. All I had to do was download and install the Editor, log on to MemSource Cloud and download the file to be reviewed. I opened the file in the Editor and the TM was automatically activated.
The advantages of MemSource:
|MemSource||SDL Trados Studio 2011|
|Shared TMs are updated in real time.||This feature is available through GroupShare|
|Fast file-opening and translation process (although QA and down/uploading take their time)||Studio isn’t renowned for its speed.|
|Join and split files. Even “linguists” can decide which files to join and when.||Studio can merge and split files, but the feature only works when you create a project. You can’t add files later and merge them.|
|Export to bilingual docx. Identical segments are greyed out, and can be autopropagated when the file is imported back into the Editor.||Export for external review has more options (including top-down or side-by-side layout), but identical segments aren’t greyed out.|
|Choice of default quote and apostrophe types in source and target.||Curly quotes can be inserted using quick insert tags. Not quite as fast.|
|TM search brings up exact matches only (great!); wildcards can be used too. Searches inside quotes bring up the exact phrase.||Concordance search brings up useful and useless matches. So a search for “vasodilation” returns “solution”. Quotes can’t be used to search for an exact phrase.|
|Subsegment match in TM||No direct leverage from subsegment matches, only through fuzzies.|
|Filtering by words in the file is very fast and simple.||Filtering is very powerful in Studio but only if you know how to use regex.|
|Segment filtering by 1st repetition (but not by number-only).||Lots of filtering options, but not by 1st repetition.|
|Lightning fast technical support for help in general and bug solving in particular.||If you don’t have a support agreement you may be left struggling.|
|MemSource||SDL Trados Studio 2011|
|Spellcheck: Internal is too limited; no real-time spellcheck. Export to docx for Word spellcheck is time-consuming.||Wiggly red underlining is a life-saver. Adding to my customised Word dictionary is very useful.|
|Autopropagation only works downwards and not for number-only segments.||Options to autopropagate from the top, ask for confirmation, etc.|
|No Goto segment||Full range of Goto options|
|Only one file can be opened at a time||Multiple file tabs|
|When files are joined, you can’t tell which is which. (Tip: click “preview translation” to see the file)||Clear file division and navigation panel within each merged file.|
|White space characters at the end of the line aren’t shown (QA will pick up “invisible” double spaces).|
|Importing of TMs hasn’t been polished yet. A small 90-segment TMX only resulted in 65 segments in MemSource. *See update in my reply below to David’s comment.|
|In-line tags are only numbered, with no other information.||Choice of tag display|
|Non-breaking space is displayed as an ordinary space. *See update in my reply below to David’s comment.|
I was pleased to discover that some features I enjoy using in Studio have also been implemented in MemSource:
- Source segment editing
- Segment locking (although I was told this is a bug and it shouldn’t be available in the “linguist” licence)
- Multiple TMs
- Font size adaptation in Editor window
- Split and merge segments (with the same no-go splitting where there are hard returns as in Studio)
- QA check
- Add terms to termbases on the fly
For translators who are new to TEnTs, I think MemSource is an excellent option, because it’s simple and fast. The free personal licence (limited to two files at a time and 10 MB file size) is great for investigating how TEnTs work, without any time limit.
For project managers, MemSource is also a good option because they can control the whole process, see who is translating what, and when. They can define QA settings, which TM to use and whether to lock segments. For post-editing work MemSource can analyse how much a target segment has been edited after machine translation and rates can be adjusted accordingly. Although I don’t do post-editing work, this is clearly an innovative alternative to charging by the hour.
For small teams MemSource also has great potential. Translators can share resources in real time.
But personally, I don’t like having my hands tied.
- As a “linguist” in a MemSource project I can’t upload translation memories or termbases (so I can’t make use of my own resources) and I can’t lock segments.
- As a freelance user of MemSource, I’d like to be able to have more say in how to use the features: I’d like to choose the way tags are displayed or decide to omit a tag (without it being autocompleted on the server when the file is generated); I’d also like to decide if and when to autopropagate segments. And I miss regex.
That said, MemSource has only been on the market for about 18 months and it’s growing fast, having just reached the 3000 user mark. I’m sure that many of the drawbacks listed in the table above will be solved soon. The developers are good listeners and they provided great support during the course of my project, so I look forward to reading about new features when they appear. In the meantime, it’s back to Studio for me.