Sarah Ianieri

Your TM must be configured to recognize alphanumerics (which is the default). W1 and W2 are considered exchangeable values, and for efficiency reasons only one instance of your translation is stored in the TM. It should be the same with the two segments above "film in RE" and "film in PL". Same with numbers. If you want this to be different across the board, turn off "recognize alphanumeric strings" in the TM settings (which I don't recommend). If you want to force certain TUs into the TM, Press CTRL+SHIFT+U to add a segment as new translation:

Daniel

Thank you Daniel.

I tried to do this but it looks like not working. Once i select the option "add a segment as new translation", nothing changes.

Sarah Ianieri

Sarah Ianieri said:

I tried to do this

Why?  As far as the TM is concerned these two segments are the same:

classe W1

classe W2

As Daniel already explained there is no need to have thee separately in the TM and you are just adding unnecessary clutter to your TM.  The only usecase I can think of for this would be if you are exporting the TM as a TMX for use in some other tool that doesn't have this capability.  In this case you would be better off adding the TU as a new translation as Daniel also suggested.

Paul in our CMS we just have the sentence "Class W2" and when it search in the translation , no results are available, and this is an automatic process so it should be in our TM for this reason.

I tried to do what Daniel suggested me, adding the TU as new translation, but the TU does not enter in TM (see steps below):

also English one

Sarah Ianieri

Sarah Ianieri said:

in our CMS we just have the sentence "Class W2" and when it search in the translation , no results are available

If you search the TM in the TM editor you won't find it.  But if you pre-translate from that TM or use it while translating then it will always be there.  As you can see:

Why exactly do you need to have this in your TM so you are not using the features of the TM and are going to be adding dumb text that Studio will not be able to reuse for your benefit?  For example:

Toorrow you get this:

classe Y56

classe G32

Both of these will be recognised 100% or CM matches without you having to add them at all.  If you switch all the recognition off you'll get a fuzzy match on them both and have to pay (time/money) for the corrections and also deal with custom QA checks looking for and validating the correct use of these alphanumerics.

Sarah Ianieri

Sarah Ianieri said:

in our CMS we just have the sentence "Class W2" and when it search in the translation , no results are available

If you search the TM in the TM editor you won't find it.  But if you pre-translate from that TM or use it while translating then it will always be there.  As you can see:

Why exactly do you need to have this in your TM so you are not using the features of the TM and are going to be adding dumb text that Studio will not be able to reuse for your benefit?  For example:

Toorrow you get this:

classe Y56

classe G32

Both of these will be recognised 100% or CM matches without you having to add them at all.  If you switch all the recognition off you'll get a fuzzy match on them both and have to pay (time/money) for the corrections and also deal with custom QA checks looking for and validating the correct use of these alphanumerics.

Paul Filkin

if we are talking about Trados Studio environment I completely agree with you. BUt our developers, that develops a new tool for our CMS, linked with our TM, could find this TU because it is not present in our TM, and they are not working in Studio environment. So how can i solve it?

Sarah Ianieri

Sarah Ianieri said:

BUt our developers, that develops a new tool for our CMS, linked with our TM, could find this TU because it is not present in our TM, and they are not working in Studio environment.

ok - now we're getting to the crux of the problem.  I think you have several choices:

1. your developers use the Translation API to do these searches and then you should have no problem, OR
2. you add as a new TU whenever you come across a match that is not based on an exact copy of what's in the TM, OR
3. you switch off the use of alphanumerics and other features like this in Studio and treat everything as plain text with no leverage from the TM as you work at all, OR
4. you ask your developers to create your own TM Provider that doesn't support any of these leverage features at all and then you add your work through this provider

Probably 2. is the most economical and sensible approach, followed by 1, then 3 and finally 4. But only my opinion.

Paul said:

you add as a new TU whenever you come across a match that is not based on an exact copy of what's in the TM

I did this but then when i was checking in the TM "class W", just class W1 is in TM..as below.. and if you says that doing this should enter in TM i cannot understand why I don't see it

@Paul

Sarah Ianieri

ok.  So I created a small test and using default settings with a clean TM I translate as follows:

All as expected and only this in the TM:

So, now I disable Alphanumerics in the TM:

I then go back and work on the file... now I have to do this sort of thing:

Getting me this in the TM:

Bit of a mess... 

Paul said:

you add as a new TU whenever you come across a match that is not based on an exact copy of what's in the TM, OR

I found the same problem as you here.  Looks like the alphanumerics overrides this... probably because you are essentially duplicating the TU as far as Studio is concerned and it won't allow this.  So you have to disable the recognition in the TM and then you actually won't have to do it anyway.  But you will end up with this sort of thing for new alphanumerics:

I suppose it all depends on who this is going to inconvenience the most!