Christiane Charbonneau 

What confuses me is you have:
New source file
PDF Target file

I am unsure how option B would work. You have two stages to overcome.

1. Leverage from FR PDF Target file
2. Identify source changes 

In theory while you can convert the PDF Target file into a "simple plain text word document" you still need the corresponding source. Without having a true source how would you 

Christiane Charbonneau said:

simply insert the updates into this document

Meaning address your 2 stages?

They are obviously aware of the changes - so they must have a way of extracting the original source. In fact your Option A implies this - and I agree that this is the correct approach.

1. Use PDF Assistant for Trados to get a simplified target FR docx
2. Perform an alignment using what the deem is true source against your target FR docx
3. Save Alignment as an SDLXLIFF
4. Also create a TM and update
5. Create new project with new source file
6. Apply TM from step 4, so new project is part populated
7. Use Perfect Match and reference SDLXLIFF from step 2

** If they really dont have the original source (regardless of format) then you could:

• Take simplified target FR docx
• Use Language Cloud portal to drag drop document for MT translation into English
• You now have EN file (MT) based on "PDF" Target file
  
• Align to Create TM 

All this may cause more time than whats its worth compared to treating this as a new translation. 
Good luck having open chats with your client - let me know the outcome!

Lyds 

Christiane Charbonneau 

What confuses me is you have:
New source file
PDF Target file

I am unsure how option B would work. You have two stages to overcome.

1. Leverage from FR PDF Target file
2. Identify source changes 

In theory while you can convert the PDF Target file into a "simple plain text word document" you still need the corresponding source. Without having a true source how would you 

Christiane Charbonneau said:

simply insert the updates into this document

Meaning address your 2 stages?

They are obviously aware of the changes - so they must have a way of extracting the original source. In fact your Option A implies this - and I agree that this is the correct approach.

1. Use PDF Assistant for Trados to get a simplified target FR docx
2. Perform an alignment using what the deem is true source against your target FR docx
3. Save Alignment as an SDLXLIFF
4. Also create a TM and update
5. Create new project with new source file
6. Apply TM from step 4, so new project is part populated
7. Use Perfect Match and reference SDLXLIFF from step 2

** If they really dont have the original source (regardless of format) then you could:

• Take simplified target FR docx
• Use Language Cloud portal to drag drop document for MT translation into English
• You now have EN file (MT) based on "PDF" Target file
  
• Align to Create TM 

All this may cause more time than whats its worth compared to treating this as a new translation. 
Good luck having open chats with your client - let me know the outcome!

Lyds 

Hi Lydia!

Thanks so much for trying to help. So here's an update on where I'm at with this project.

After discussion with the client, we've agreed that Option A will actually be the best approach to handling this project.

However, I've come across some issues trying to align the documents, due to the complex formatting in the PDF versions of the previous content.

I tried using the PDF Assistant plugin you suggested (i.e. I downloaded the plugin then followed the steps outlined here https://community.rws.com/product-groups/trados-portfolio/rws-appstore/w/wiki/6598/pdf-assistant-for-trados), but it seems the "simplified" version of the PDF it produced contains even more formatting issues than when I tried converting the PDFs using Adobe Acrobat's conversion features. The "simplified" version created by the plugin contains all kinds of tabs that shouldn't be there, the columns are out of alignment, and every single line of text seems to be in its own separate text box). Just the same, I tried to align both of these "simplified" files using the alignment tool, but I get an error message from Trados saying it can't align the files because they are corrupt.

Attempting to align the files using the PDF source files or versions of these converted with Adobe's tools (either to a Word or txt file) produces equally frustrating results: these converted files contain random line returns throughout the text, which leads to poor segmentation.

At this point, I've reached out to the client to ask is she could do a further search within the company to find the original Word versions of the previous report (before it was converted to PDF), which would provide me with reliable versions of the content to do a proper alignment (they probably have those files somewhere in the company, but my client contact probably didn't bother searching for them because she was unaware of the formatting complexities this creates for me).

In the meantime, if you have any other ideas, feel free to let me know!

Thanks again,

Christiane

Christiane Charbonneau 

Small correction 

Christiane Charbonneau said:

unaware of the formatting complexities this creates for me

Many users would be struggling as you are. All this pre project you are putting in could be so extensive - just accepting it as new translation could be the kinder approach. Given they are asking you to work with that you have.

Well done for collaborating and being transparent to your client. 

"just accepting it as new translation could be the kinder approach. Given they are asking you to work with that you have."

Well in fact, the client does not expect me to translate everything from scratch; as a reminder, the report contains 23 500 words, and about 19 000 of this is previously translated content. Even if I put in a considerable amount of time for this "pre project" phase, I've estimated it still comes out to a fraction of the time and cost of re-translating everything, which I don't think is fair to the client.

I did manage to have a verbal discussion with my contact yesterday, to better explain the situation to her and what the various approaches to handling this project could entail for them, in terms of additional costs and delays. She is going to try on her end to find the Word version of the previous report, though I understand it's not an easy task because this version was created in 2018 by employees who no longer work for the company.

I think I have come up with a viable plan B if my client's search for the Word version is unsuccessful. In my efforts to convert the PDF reports into a workable format, I managed to extract them to plain text files. They still contain a large number of random line breaks, but I also found a tutorial on how to bulk delete many line breaks in a text file copied to Word. Using this method, I think I would have versions of the report's previous content that could be aligned effectively using Trados' alignment tool.

Thank you again for all your efforts to help me!

Very call Christiane Charbonneau 

Would love to learn more about your 

Christiane Charbonneau said:

tutorial on how to bulk delete many line breaks in a text file copied to Word

Maybe its something we can include in our PDF Converter

Lyds

Hi Lydia. This is the link to the tutorial I found: https://youtu.be/mANkNBfFvdc?si=sJyc_9WHoua8SC2M