Run analysis before or after translating?
Thread poster: Nora Diaz

Nora Diaz  Identity Verified
Mexico
Local time: 18:08
Member (2002)
English to Spanish
+ ...
Aug 31, 2005

When should an analysis be run: before or after translating a document?

With an empty TM or with the one you've created?

And what are the differences when interpreting pre- and post-translation analyses?


Direct link Reply with quote
 

Burrell  Identity Verified
United Kingdom
Local time: 01:08
Member (2004)
English to Latvian
+ ...
Before Aug 31, 2005

Analysis should always be run before the translation. TM type depends on circumstances. If the client has provided you with the translation memory and is paying you accordingly, you should analyze the file against the supplied TM. If you translate from the scratch, then TM should be empty.

HTH, Cheers
Burrell


Direct link Reply with quote
 

Kirill Semenov  Identity Verified
Ukraine
Local time: 03:08
Member (2004)
English to Russian
+ ...
W/ Burrell Aug 31, 2005

Nora Diaz wrote:
When should an analysis be run: before or after translating a document?


If you are paid by source wordcount, it should be run before translating.


With an empty TM or with the one you've created?


If the client supplies you a TM, run the analysis with the TM. Otherwise, run it with an empty TM.

If you run the analysis with the TM created during translation, it will more probably show almost 100% repetitions.


And what are the differences when interpreting pre- and post-translation analyses?


If you are paid by target wordcount, run it after translating, but with an empty TM.


Direct link Reply with quote
 

Nora Diaz  Identity Verified
Mexico
Local time: 18:08
Member (2002)
English to Spanish
+ ...
TOPIC STARTER
Results are similar... is that right? Aug 31, 2005

Since I'd already run the translated files with my TM (not client-supplied) open and already had the log file, I decided to run them again with an empty TM to see what I could learn from comparing the results.

Well, it turns out, they don't seem all that different, except for the fuzzies (of course no fuzzy matches are listed when the files are analyzed with an empty memory).

Here's an example of just one file (I analyzed 15 and they all had similar results):

1. Analyzed with the TM I created while translating the file:

Match Types Segments Words Percent Placeables
Context TM 0 0 0 0
Repetitions 166 261 14 98
100% 60 60 3 30
95% - 99% 1 1 0 1
85% - 94% 2 4 0 4
75% - 84% 3 4 0 5
50% - 74% 0 0 0 0
No Match 189 1,552 83 115
Total 421 1,882 100 253

2. Same file, analyzed with an empty TM:

Match Types Segments Words Percent Placeables
Context TM 0 0 0 0
Repetitions 218 313 17 123
100% 0 0 0 0
95% - 99% 0 0 0 0
85% - 94% 0 0 0 0
75% - 84% 0 0 0 0
50% - 74% 0 0 0 0
No Match 203 1,569 83 130
Total 421 1,882 100 253

If I'm interpreting this correctly, both sets of results are basically the same, with a different breakdown.


Direct link Reply with quote
 

Jerzy Czopik  Identity Verified
Germany
Local time: 02:08
Member (2003)
Polish to German
+ ...
I do not quite get your point Aug 31, 2005

You analysed the files with your TM prior to start translating. That's clear so far.
Then you analysed the same file against an empty TM - that's clear too.
With a clean TM you should only get repetitions within the files, nothing more.
Analysing with your own TM can show some matches, however: with files of about 1500-200 words analysed file by file the chance to get a big ratio on matches in not very big.
But what could be interesting to you, is an analysis against your TM you used to translate those files - all should be 100% match or repetition.

To summarize this:
You analyse the files prior to start translating either against an empty TM (if no pretranslation is there) or against a TM with other content. So then you know how much you have to do.
You can also analyse files against an empty TM after translating. But you shouldn't analyse the translated files! The language of those files does not match the language selection of your TM!
The analysis prior to translate is basis for your offer - either with or without rebate for matches/repetitions. An analysis of untranslated files against the TM used to translate this files is worthles, as it delivers only 100% matches, and an analysis of translated files against an empty TM may make sense only, if the languages match.

What is your goal in analysing files after the translation?

Regards
Jerzy


Direct link Reply with quote
 

Nora Diaz  Identity Verified
Mexico
Local time: 18:08
Member (2002)
English to Spanish
+ ...
TOPIC STARTER
Let me try to explain myself more clearly Sep 1, 2005

Jerzy,

Actually, in both instances I analyzed the file after I finished translating it, once against an empty TM, once against the TM I used while translating the file. This is where I expected to see 100% match or repetition, like you've explained, but the results were similar to those of the empty TM (as seen in my previous post). From your post I understand that this is because the translated file's language doesn't match the target language of the TM, so that's clear now.

I'm new at analyzing files, so I'm trying to understand how this works. The reason I'm trying to figure this out is because the client has asked me to submit an analysis of each translated file along with the file itself (does that make sense?). Our agreement is based on a simple scale: 50% rate for 100% matches, full rate for everything else, based on number of target words. When I was ready to run the analysis, I started having second thoughts about how different ways of doing it could end up affecting the results, hence all my questions.

I want to understand which is the most appropriate way to do it (for this purpose) so that I can in turn explain it to the client if he asks why I decided to do it in a particular way.

Thanks!

Nora


Direct link Reply with quote
 

Hynek Palatin  Identity Verified
Czech Republic
Local time: 02:08
English to Czech
+ ...
Analysis Sep 4, 2005

Nora Diaz wrote:
Our agreement is based on a simple scale: 50% rate for 100% matches, full rate for everything else, based on number of target words.


Charging by number of target words makes things complicated, because Trados analysis is based on source words.

To find out the number of target words for 100% matches, you should analyse the final (translated and *cleaned*) documents against reversed TM created like this: Export the TM you received from your client (*not* the final TM), create a new one with reversed language pair and import the previously exported file. This analysis won't be accurate, if there are different source segments (non-100% matches with regular analysis) translated into identical target segments (100% matches with reversed analysis). Maybe there is a better solution, it just doesn't occur to me right now.

Charging by source would be much easier, because you would be able to use the standard analysis.


Direct link Reply with quote
 


To report site rules violations or get help, contact a site moderator:


You can also contact site staff by submitting a support request »

Run analysis before or after translating?

Advanced search







Protemos translation business management system
Create your account in minutes, and start working! 3-month trial for agencies, and free for freelancers!

The system lets you keep client/vendor database, with contacts and rates, manage projects and assign jobs to vendors, issue invoices, track payments, store and manage project files, generate business reports on turnover profit per client/manager etc.

More info »
PerfectIt consistency checker
Faster Checking, Greater Accuracy

PerfectIt helps deliver error-free documents. It improves consistency, ensures quality and helps to enforce style guides. It’s a powerful tool for pro users, and comes with the assurance of a 30-day money back guarantee.

More info »



Forums
  • All of ProZ.com
  • Term search
  • Jobs
  • Forums
  • Multiple search