Maria E. Choque:

> I began using Trados last year, and eversince I've used it
> ocassionally. At that time, I created a Translation Memory
> (TM), but I had a "cleaning" problem once the document was
> completely translated. A person I knew then told me to send
> the document and he fixed it.
>
> The same client contacted me again with another document to
> translate, and I'm not sure if I should use the same TM for
> this document or create a new one. Since, I donīt even know
> how that person fixed the document, I donīt know if by using
> the same TM I will have problems with the new document.

You don't say what clean-up problem you had. It would be useful if you could
tell us exactly what the error message you got during clean-up said. You
also don't say anything about whether your document was a Word document or
some other format.

A couple of possibilities come to mind:

Since you seem to be relatively new to Trados, and if you were working in
Word, i.e. not Tag Editor, you might have damaged one or more tags (i.e. the
purple parts surrounding the segment. In

{0>Baujahr<}100{>Year of manufacture:<0}

the tags are

{0>
<}100{> and
<0}

Fixing the document would mean repairing the tags.

If you got an RTF error message, meaning Workbench wasn't able to clean up
the document at all, then your friend may have cleaned up the document using
the Word macro. The downside to this is that the most recent version of the
translation may not be in your memory since cleaning up the document in Word
doesn't update the memory. This happens if you made any changes to
individual segments of your translation without opening and closing them. If
you are able to clean up the document in the usual manner using Workbench,
then this doesn't pose a problem.

Please ask your friend what he or she did.

> Another question is about analyzing, should I analyze the
> document before I translate or after, and I donīt know how to
> interpret the listing that this feature shows.

Analyzing makes sense only if your new document is similar in content to the
old one. AND if the content of the old document is in your memory. That
would depend on the answer to the first part above.

Analyzing tells you how similar your new document is to the content in your
memory (in your case, the old document). So this is useful information
BEFORE you start work on a translation.

Take these figures, for example:

Match Types Segments Words Percent Placeables
Context TM 0 0 0 0
Repetitions 6 43 3 0
100% 72 563 44 3
95% - 99% 30 92 7 5
85% - 94% 4 48 4 0
75% - 84% 5 21 2 0
50% - 74% 32 178 14 0
No Match 38 335 26 0
Total 187 1,280 100 8


The 100% figure shows that 44% of my new document is identical to what is
already in the memory. The 7% figure given for the 95% - 99% range says that
7% of my new document is **virtually** identical to what is already in the
memory. The other ranges show percentage portions of the document that are
less similar to what is already in the memory.

"Repetitions" always refers to **internal repetitions** in the new document,
i.e. untranslated material. In the case above 3% of the new document
consists of repetitions, i.e. source segments that are identical within the
document to be translated.

100% matches always refer to segments in the new document that are identical
to segments in the memory. Of course, this does NOT mean that the
translation is correct because you always have to check this during
translation.

When working with agencies it is crucial that you understand the difference
between these two figures. The matches -- whether 100% or fuzzy matches (95%
- 99%, 85% - 94%, anything that is NOT a 100% match) -- are always in
reference to material already in the memory, material that has already been
translated, whether by you or someone else.

Repetitions are always in reference to internal repetitions in the new
document to be translated, meaning *you* translate this material.

Amy