Sounds a lot like the DDBAccess software from Christian Wittern (?). It takes the Chinese text, parses it all out, then automatically looks up the DDB, giving a basic English term for each character or simple compound.
Really only useful for people who simply do not have any clue whatsoever. The idea of simply plugging in given terms into a passage is a terrible way to, ahem, "translate", <cough cough>. One would have to identify a genre of literature, the translator too, time period, and so forth, to start getting to specific meanings in context.
As it has been said: Words do not have meaning, sentences have meaning. Or rather, sentences within paragraphs, within whole texts, etc.
Not quite. DDB Access does something very specific. It's proprietary Windows software that allows the user to interactively examine dictionary definitions for Chinese Buddhist terms according to definitions made by the people who contribute to DDB. It has a very narrow goal and expected usage. The user sits behind a computer monitor, logs in with his/her DDB login, and moves the mouse cursor over some term in the text that was pasted in.
Sanzang is a cross-platform command line utility that implements a translation algorithm. It's useful for East Asian languages such as (but not limited to!) Chinese. The idea is that the program reads a "translation table" of equivalent terms or phrases, and translates them accordingly in the text. It can also reformat texts according to their punctuation, which is generally a good idea before attempting to translate them (so words and phrases are not broken up between lines).
When I made Sanzang, I thought that dictionary programs were useful, but I disliked the user interfaces, lookup methods, output formats, etc. Interactive dictionary programs are usually rather complex, rigid, and tedious to use. Everything I've done with Sanzang is to get away from that way of doing things, to give control back to the user, and to generate succinct and helpful output.
For example, Huifeng, suppose you want to make a special translation table just for Lokaksema's vocabulary, that goes from Chinese into Gandhari, Sanskrit, and English. Dictionary programs like DDB Access cannot do that. However, Sanzang has the potential to do that easily. And when you've finished with your "Lokaksema table," you could easily share it with your colleagues as well (a translation table is just a plain text file). That's the whole idea: simple, clean, and flexible.
It would only be good as a very-rough-draft - I am not sure how one would try to use software to infer all of the English grammar that Chinese lacks and is understood by a translator (and differently in some cases from one to the next). It certainly would be an asset to me as a translator to have software that would spot all of the transliterations or other technical terms - the ordinary Chinese would not be so important.
Exactly, its output tends to be rather rough and literal. For the grammar, equivalent terms are given for each part of a sentence, but terms are not shuffled around or reordered within the sentence.
It would be quite useful to get a sense of what the text is about.
Right, I agree. In my experience, it helps bridge the gap, and it can make reading texts faster and easier.