Hi everybody,

With a sea of fairly good machine translation engines around, what is the market niche of SDL's Language Cloud? I'd love to hear what the community here at the forum thinks about it. I am impressed with the step up that Language Cloud took from the statistical to the neural model. It's now really useful and playing in one league with the big MT providers, although for EN-DE the style of DeepL is still unmatched.

What I hope is being developed by SDL (and offered at prices affordable to Freelance translators and SMBs) is close integration with Trados Studio and MultiTerm:

• Ability to allow custom TermBases (limited number for "lighter" plans), with direct upload
• Linguistic fuzzy terminology matching
• Usage of Termbase for prescribed translations (favoring "preferred" over "admitted", preventing "deprecated" etc.)
• Easy integration (that is already provided - it's really simple to get it to work)
• Could it be set to access the TM hits to become a deluxe "match repair" fuction? Example: "Dies ist das kleine Haus im großen Wald" -> "This is the little house in the big woods." might be in the TM. Source segment be "Dies ist der große Wagen im großen Wald." Why can't an MT engine match "kleine Haus" and "little house" and keep the rest of the TM hit intact but replace just the deviating words "große Wagen" with a machine translation? This is what I mean with close integration, all being SDL.

I know most of this is technically not possible at this moment, but as we see LC take this big quality step up, I think of the vision and direction of it - it is competing with many other strong MT engines, which are free or have a very large number of free chars per month or have exceptionally high translation quality.

I'd love to hear what other users think about this.

Daniel