Hi David

Hi David, can you provide a bit more information on what your translators actually want to do when editing the Translation Memory (TM).

For example, you mention that there are a number of outdated terms in the TM that they would like to work on. Do this mean that they want to change specific words/phrases with the TM to a "new" different one? Or, do they want to delete a range of whole translation units (TU's), which is the whole source sentence/segment linked to the associated target sentence/segment? Or, maybe it is some other type of clean-up they wish to complete.

If you can come back with some further background on this it will help us understand what the requirements are and allow us to come back with some ideas and suggestions on what can be done and how.

Look forward to hearing back.

Phil.

Hi David

Hi David

Hi David

Cleaning up large TMs is quite a tedious task. The main problem is that most TM editors are not able to process such large TMX or SDLXLIFF files. One intuitive approach would be for example to export the TM to XLIFF and then open this XLIFF in Studio, which would allow one to benefit from all the editing goodies of Studio. Unfortunately, this won't work because Studio will crash when you open such a large SDLXLIFF in its Editor. 

Having executed successfully quite a few TM cleaning projects in the past years, the only feasible tool I have found is the open source software Olifant from the Okapi Framework.

It is quite tricky to use, has not been updated for ages and crashes sometimes, but once you know how to deal with it, you can achieve good results. Its main assets are that it can process very large files and allows you to do batch delete and batch replace tasks, which the other TMX editors on the market don't allow you to. 

Best regards
Walter

Hello Phil

This is what they ask of me:

"We have a lot of old and outdated terms in there that are causing confusion for us and the freelancers and are leading to inconsistencies in our translations, so were wondering if you might have an idea or some best practices on how we could approach this best?"

David

Walter

Yes, at over 2000 pages I could see that being an issue bringing back into Trados.  I can see how to create a TMX as an export file.  How would create a SDLXLIFF file?

With the Stay at Home being extended through May it would be challenging to work with other Programs. How easy is it to learn and use Olifant?  How tech knowledgeable would one have to be?

David

Hello David

There are tools that you can use inside of Trados Studio desktop that will for, example, let you batch delete whole Translation Units or carry out search and replace on specific words and phrases within a translation memory, but as Walter says, depending on the actual size of the TM this could be a slow and long winded procedure.

Although you have said that the TM is 2,000 pages, it would be good to know exactly how many Translation Units (TUs) this actually equates to, as when you view a TM inside of Trados Studio (2019 and maybe 2017 versions) you can vary the number of entries viewed within a page. This can be 10, 50, 100, 500 or 1,000, so it would make quite a difference to the actual number of TUs you need to search through and edit in some way. You can find out this information. You can find out the number of TUs in a specific TM by opening it via the Translation Memories view in Trados Studio desktop, making usre this TM is highlighted and then either right clicking on it and selecting Settings or clicking  Settings in the Ribbon. In the main page this will then confirm the name of the TM, languages managed by this TM, where it is stored etc. and at the foot of the form it will show you the total number of TU's within this TM.

So as not to mess around with a live TM in GroupShare an old Client of mine used to export the TM from GroupShare, set up a file based TM in Trados Studio Desktop and then import the TMX file. Once this was done she could then work on the TM here at her leisure without causing any issues to ongoing work and once she had edited the TM as required, she could then use this to replace the old one back in GroupShare.

The one query I still have is what you mean by terms? Is this specific words or phrases they want to search for and edit/change within the all the TUs, or do they mean they want to delete whole TUs, for example created before a certain date, or maybe it's a combination of both? Again, this would make a difference to how you search for things and complete the task.

Cheers

Phil.

Hope this helps and if you can check and let us know exactly how many TUs there are in the TM they want to edit and whether it is individual words/phrases they want to change in the TUs or if they want to delete whole old TUs, then we can come back again with some ideas on how to achieve your goals.

Hello Phil

The total TU count is: 113488

As to what the translators are looking to clean up they did not specify, but it is likely a mixture of words and phrases.

This TM has been around since 2017. There has been, like many companies, changes in terminology. They said there are a number of outdated terms they are looking to fix or delete depending on what it is.

Hello David,

Thanks for the additional information. The TM is not a massive one, so it is something I think they could do inside of Trados Studio 2019. I'll put together an outline of the process I am thinking of and share it with you probably tomorrow.

Can I just confirm one last piece of information regarding the TM. You mentioned originally that it was a server based TM and then later said you are using Trados Studio 2019. Is this a file based TM created in Trados Studio and maintained/shared on a file-server within your business, or do you also have a SDL GroupShare system and this is a TM maintained there? This will ensure I come back with the right process :).

Thanks and have a good day.

Phil.