Start a new topic

Support for Slate Desktop (once it's there ...)?

Hello Igor,

Could you please have a look at this thread.

Would you be willing to support any 'connector' for Slate Desktop?



1 person likes this idea

In anticipation of your reply and knowing how you have fulfilled many of our requests, I have taken the plunged and signed up for a CT "perk" at Indiegogo. At the same time I fear that this technique has the danger to making us shoot in our own feet. But that'll be nothing that we can stop. Especially not by ignoring it.
Plunge Shoot ourselves

Hi Hans,

It is hard to comment on something which is in the making, as I understand. If there exists such a connector I might take a look at it and provide an option to integrate such as with Google Translate and Microsoft Translator.

I think that MT has a bright future for translators but definitely not in the current form of post-editing, say, Google Translate results, which is more time consuming and interfering with the work than creating your own translation. Though in the "shallow" integration with CATs, it is great to have it as another resource like a glossary/dictionary or web resource.

As for MT technology in CafeTran, my goal is to push it to the deepest level touching on AI elements where the program will smartly suggest the given term or phrase by auto-adjusting their results depending on the context and learning from your own personal choices in your field of expertise. MT should be subtle, non-interfering, guiding, and absolutely not another statistics element (like fuzzy matches) exploited by some agencies. I hope you have noticed some early birds of it in the form of Total Recall, subsegment matching, pre-translation and auto-assembling which in fact constitute Machine Translation technology in CT. The last update does some context analyzing in sub-segment matching so the hits are not purely statistical now and may change depending on the context of the source segment. The question is whether such a contextual hit (the AI part of MT) should overwrite glossary matches and priorities during auto-assembling. Users' feedback would be highly appreciated. Thanks in advance!


2 people like this

Hi Igor,

I am the CEO of the company behind Slate Desktop. In preparation for our version 1.0 launch in January, we're running an Indiegogo campaign as a market research tool to engage possible customers early and prioritize which MT connectors we should develop first.

It's been a great experience and in this process, we uncovered an avid group of CafeTran users on ProZ.

I know little about CafeTrans, but from screen shots your customers have shared and your comments above, it looks like you do not support a connector API for 3rd party vendor. 

I also sense a health and understandable caution in your voice about MT in light of how it has been promoted for the last 10 years. Our engine runs right on the translator's local host. All aspects of the engine, its creation and usage are under the translator's control and without any connectivity to the Internet.

I'm happy to engage with you directly to bring you up to speed about how Slate Desktop works and how we are developing our interfaces.

Tom: we uncovered an avid group of CafeTran users on ProZ.


You can find that discussion here. More talk is going on in the Yahoo group help_.

My contribution (negative as always): no use for CafeTransers.


Hi Tom,

CafeTran connects to other resources ( such as Google Translate, Microsoft Translator, MyMemory) via their APIs. If you provide an API for the connection, please point me to it.


GT, MSHT, MyMemory, and other Traditional MT engines are out-of-process web-based APIs. The Slate Desktop API will be an in-process library. When we are closer to publishing, I'd like to talk to you off-line (from this forum) to arrange a dev copy for you, or we can co-develop the code. Your call.

Great. So I am looking forward to that API library for CATs (preferably in Java) to try it out.


Login to post a comment