ProZ.com Translation Forums

Greek as a source language in CAT tools: what are the actual issues? | Exploring

CafeTran Trainer — Wed, 13 May 2026 20:02:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: CafeTran Trainer
Post title: Exploring

[quote]Epameinondas Soufleros wrote:

What are you trying to do? [/quote]

I’ve noticed a plateau in CAT tool innovation. While UI and cloud integration have improved, we're missing a deeper focus on linguistic algorithms tailored to specific source and target languages.

There is significant efficiency to be gained by addressing language-specific hurdles directly within the editor, such as:

- German: Better handling of Bindestriche (hyphenation and compound nouns).
- Dutch: Solutions for los schrijf ziekte (incorrect compounding).
- Greek: Intelligent management of grammatical cases and case-sensitive capitalization.
- Polish: Integrated stemming and morphological analysis.
- Etc.

We need to explore these algorithmic solutions and integrate them into our existing toolkit, whether open-source or proprietary.

Greek as a source language in CAT tools: what are the actual issues? | What are you trying to achieve?

Epameinondas Soufleros — Wed, 13 May 2026 16:07:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Epameinondas Soufleros
Post title: What are you trying to achieve?

[quote]CafeTran Trainer wrote:

And with this table, I get: [url removed]

Now shoot ;)
[/quote]

What are you trying to do? Build yet another vibe-coded translation tool?

Greek as a source language in CAT tools: what are the actual issues? | Also in Dutch

CafeTran Trainer — Wed, 13 May 2026 13:11:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: CafeTran Trainer
Post title: Also in Dutch

[quote]Epameinondas Soufleros wrote:

Another issue that has existed for ever is automatic case conversion for Greek words that contain vowels with stress marks on them or the so-called "final sigma" (ς). Most CAT tools and word processors, as well as the corresponding CSS function (text-transform: uppercase) perform the conversion incorrectly:

άνθρωπος should become ΑΝΘΡΩΠΟΣ, not ΆΝΘΡΩΠΟΣ (notice the first letter).
And when you keep cycling through the variants by pressing Shift+F3 (the usual shortcut), you should go back to άνθρωπος, not άνθρωποσ (notice the last letter).

The thing is we don't use stress marks when we write in BLOCK CAPITALS.
(There is an exception, though: we keep the stress mark in the word Ή when it is a disjunction ("or") and not an article ("the"). I wouldn't expect any tool to handle this exception, but it could at least get the basics right.)

Also, the lowercase letter sigma is normally written as "σ" but becomes "ς" when it is the last letter of a word.

[Edited at 2026-05-12 09:51 GMT]

[Edited at 2026-05-12 12:54 GMT] [/quote]

We have that too:
Een uitzondering is één. Als één aan het begin van een zin voorkomt, vervalt het eerste accent op de hoofdletter e. Het is dus Eén, met alleen een accentteken op de tweede e.

(And, if I'm not mistaken, it's the rule for French to lose the accent when converting to uppercase.)

By the way, CafeTran Espresso correctly converts "ij" to "IJ" when toggling case ;)

Looks like this is the conversion table for Greek:
Name,Lowercase,Uppercase,Phonetic (Approx.)
Alpha,α,Α,a
Beta,β,Β,v / b
Gamma,γ,Γ,g / gh
Delta,δ,Δ,d / dh
Epsilon,ε,Ε,e (short)
Zeta,ζ,Ζ,z
Eta,η,Η,i / ee
Theta,θ,Θ,th
Iota,ι,Ι,i
Kappa,κ,Κ,k
Lambda,λ,Λ,l
Mu,μ,Μ,m
Nu,ν,Ν,n
Xi,ξ,Ξ,x
Omicron,ο,Ο,o (short)
Pi,π,Π,p
Rho,ρ,Ρ,r
Sigma,σ / ς*,Σ,s
Tau,τ,Τ,t
Upsilon,υ,Υ,y / u
Phi,φ,Φ,f / ph
Chi,χ,Χ,ch / kh
Psi,ψ,Ψ,ps
Omega,ω,Ω,o (long)

And with this table, I get: [url removed]

Now shoot ;)

[Edited at 2026-05-13 13:47 GMT]

Greek as a source language in CAT tools: what are the actual issues? | It is relevant

Epameinondas Soufleros — Wed, 13 May 2026 09:23:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Epameinondas Soufleros
Post title: It is relevant

It is relevant, because it shows the kind of machine translation clients serve us and expect us to offer them a discount.

If anything, we should apply a surcharge for having to read and edit such nonsense instead of translating from scratch.

Greek as a source language in CAT tools: what are the actual issues? | Relevance

CafeTran Trainer — Wed, 13 May 2026 09:20:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: CafeTran Trainer
Post title: Relevance

[quote]Epameinondas Soufleros wrote:

The lazy fox source text you have there backtranslates like so:

Quickest fox of paint terra cotta, steps in favour of lazy dog.

What a nice word salad! Bon appétit!

[Edited at 2026-05-13 09:19 GMT] [/quote]

That sentence is only a placeholder to reset glossary recognition between the sentences.

It was offered by Google. Indeed without the -ai command…

Greek as a source language in CAT tools: what are the actual issues? | Kids these days with their AI toys...

Epameinondas Soufleros — Wed, 13 May 2026 09:15:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Epameinondas Soufleros
Post title: Kids these days with their AI toys...

The lazy fox source text you have there backtranslates like so:

Quickest fox of paint terra cotta, steps in favour of lazy dog.

What a nice word salad! Bon appétit!

[Edited at 2026-05-13 09:21 GMT]

Greek as a source language in CAT tools: what are the actual issues? | spaCY

CafeTran Trainer — Wed, 13 May 2026 09:10:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: CafeTran Trainer
Post title: spaCY

[quote]Li-Hsiang Hsu wrote:

spaCy is trained on statistical rules, not rigid regex.

Trados and CafeTran rely on rigid regex rules, less flexible and less reliable than spaCy for morphologically complex languages.
[/quote]

I’m afraid I couldn’t get case toggling working, neither for Greek → English nor for English → Greek. I’ll leave it at that, since this was really just an academic exercise for me and I don’t have any plans to start translating from or into Greek. I’m sure there are people here with much more experience in this area who can give better advice.

I got stemming working, but no auto-adaptation to plural of the target. I must admit that this doesn’t work for other language combinations either: I always have to add at least one plural case.

[Edited at 2026-05-13 19:35 GMT]

Greek as a source language in CAT tools: what are the actual issues? | Not only in Greek

CafeTran Trainer — Wed, 13 May 2026 08:19:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: CafeTran Trainer
Post title: Not only in Greek

[quote]Epameinondas Soufleros wrote:

Software developers are notorious for their laziness, which makes them adamant on the DRY (Don't Repeat Yourself) principle.

So, they expose this one string for translation: read. It can function as a heading, as an introduction to a list, as a button label, as a command... It can also be automatically converted to READ.[/quote]

This one is known from other language combinations too.

[Edited at 2026-05-13 08:38 GMT]

Greek as a source language in CAT tools: what are the actual issues? | These do not look like CAT tools

Epameinondas Soufleros — Wed, 13 May 2026 08:05:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Epameinondas Soufleros
Post title: These do not look like CAT tools

What are these? They don't seem to be CAT tools, so can they be compared to Trados Studio and CafeTran?

Greek as a source language in CAT tools: what are the actual issues? | Try spaCy (el_core_news_lg) and greCy for ancient Greek?

Li-Hsiang Hsu — Wed, 13 May 2026 07:14:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Li-Hsiang Hsu
Post title: Try spaCy (el_core_news_lg) and greCy for ancient Greek?

spaCy is trained on statistical rules, not rigid regex.

Trados and CafeTran rely on rigid regex rules, less flexible and less reliable than spaCy for morphologically complex languages.

Greek as a source language in CAT tools: what are the actual issues? | Great points on Greek morphological complexity

Rodolfo Abachi — Tue, 12 May 2026 13:16:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Rodolfo Abachi
Post title: Great points on Greek morphological complexity

The examples shared here highlight the core challenge: Greek is a highly inflected language, and most CAT tools were built with Latin-script, less morphologically complex languages in mind. The fuzzy matching issue with inflected forms (like needing to add every case and number variant to a glossary) is something I've encountered when working with Portuguese, which is also inflected, though Greek takes it to another level.

For Greek-to-X translation, tools like OmegaT with proper tokenizers and lemmatizers can help, though support is inconsistent. Has anyone tried using a dedicated Greek morphological analyzer as a pre-processing step before feeding text into CAT tools? That might help with the matching issues.

Greek as a source language in CAT tools: what are the actual issues? | More tales from software localisation

Epameinondas Soufleros — Tue, 12 May 2026 10:11:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Epameinondas Soufleros
Post title: More tales from software localisation

Software developers are notorious for their laziness, which makes them adamant on the DRY (Don't Repeat Yourself) principle.

So, they expose this one string for translation: read. It can function as a heading, as an introduction to a list, as a button label, as a command... It can also be automatically converted to READ.

But Greek needs to know the exact context of this word. Where does it appear? What does it go with? Does it refer to one or multiple items? Is it a command, a notification, a popup title, a confirmation, a button, a call-to-action? (A cryptic context ID like "app.base.forms.assets.read" will make me none the wiser!)

Depending on the context, we could have these possible Greek translations of the same English word:
- διάβασε
- διάβασαν
- διαβάστε
- διαβάστηκε
- διαβάστηκαν
- διάβασμα (or: ανάγνωση)
- διαβασμένο (or: αναγνωσμένο)
- διαβασμένα (or: αναγνωσμένα)

This is at least 8 different possibilities, not counting their uppercase variants, which would need to be typed in block capitals right from the start and not rely on automatic case conversion.

[Edited at 2026-05-12 10:11 GMT]

Greek as a source language in CAT tools: what are the actual issues? | Here's another one

Epameinondas Soufleros — Tue, 12 May 2026 09:41:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Epameinondas Soufleros
Post title: Here's another one

Another issue that has existed for ever is automatic case conversion for Greek words that contain vowels with stress marks on them or the so-called "final sigma" (ς). Most CAT tools and word processors, as well as the corresponding CSS function (text-transform: uppercase) perform the conversion incorrectly:

άνθρωπος should become ΑΝΘΡΩΠΟΣ, not ΆΝΘΡΩΠΟΣ (notice the first letter).
And when you keep cycling through the variants by pressing Shift+F3 (the usual shortcut), you should go back to άνθρωπος, not άνθρωποσ (notice the last letter).

The thing is we don't use stress marks when we write in BLOCK CAPITALS.
(There is an exception, though: we keep the stress mark in the word Ή when it is a disjunction ("or") and not an article ("the"). I wouldn't expect any tool to handle this exception, but it could at least get the basics right.)

Also, the lowercase letter sigma is normally written as "σ" but becomes "ς" when it is the last letter of a word.

[Edited at 2026-05-12 09:51 GMT]

[Edited at 2026-05-12 12:54 GMT]

Greek as a source language in CAT tools: what are the actual issues? | Peculiarities of Greek

Epameinondas Soufleros — Tue, 12 May 2026 09:29:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: Epameinondas Soufleros
Post title: Peculiarities of Greek

I am not sure about Greek as source language, because I mostly work with Greek as the target language.

In any case, let's take as an example the word "human" and let's suppose we want to add it to a glossary.

In English, fuzzy matching will do the trick and recognise both "human" and "humans".

In Greek, however, in order for QA checks to work correctly and not plague us with a ton of false positives, we need to add all of the following variants:
- άνθρωπος
- ανθρώπου
- άνθρωπο
- άνθρωπε
- άνθρωποι
- ανθρώπων
- ανθρώπους

In memoQ, which supports pipes as a separator between stem and suffix, we could add the following entries:
- άνθρωπ|ος
- ανθρώπ|ου
and select the "Permissive" option for case sensitivity.

A different solution would be to use the following regular expression:
- [άα]νθρ[ωώ]π(ος|ου|ο|ε|οι|ων|ους)
and also enable case insensitivity.

There is no CAT tool that supports this, and for good reason: what would it show you in the term recognition panel? All possible matches?

So, the current solution is to use AI, which supposedly knows the different word forms (grammatical cases) of the word. Does it, though? Machine translation and artificial intelligence tools are still very bad at Greek, but people fail to realise this or, most probably, don't care about a language with such a small community of users.

Greek as a source language in CAT tools: what are the actual issues?

CafeTran Trainer — Tue, 12 May 2026 08:15:00 +0000

Forum: CAT Tools Technical Help
Topic: Greek as a source language in CAT tools: what are the actual issues?
Poster: CafeTran Trainer

I've come across mentions that many (or possibly all?) CAT tools struggle when Greek is the source language. I'd love to understand what's actually going on technically.

Specifically, I'm curious about:

- Are the problems universal across CAT tools, or do some handel Greek better than others?
- Is this a Unicode/encoding issue, a tokenization problem, or something to do with how Greek characters interact with TM matching algorithms?
- Does it affect all Greek text equally, or are there specific cases (e. g. polytonic Greek, mixed Greek / Latin content, Greek numerals) that are especially problematic?
- Are there known workarounds ?

Tools (software) to handle/modify/edit .xliffs and other CAT-generated files | Very interesting!

Jose Mota — Fri, 27 Feb 2026 14:30:00 +0000

Forum: CAT Tools Technical Help
Topic: Tools (software) to handle/modify/edit .xliffs and other CAT-generated files
Poster: Jose Mota
Post title: Very interesting!

[quote]Jorge Payan wrote:

[url removed]

[url removed]

It is just a pity that the old stand-alone SDLXLIFF Split/Merge is not available for Studio 2024.

[/quote]

Dear Jorge,
Thank you for this, it looks very interesting.

For this particular task I've used the workaround with the Toolkit, but StudioViews seems like a great option and it might be easier that way. I'll try this method next time.

Many thanks!

Tools (software) to handle/modify/edit .xliffs and other CAT-generated files | My first choice: studioViews Plugin for TRADOS 2024

Jorge Payan — Fri, 27 Feb 2026 13:49:00 +0000

Forum: CAT Tools Technical Help
Topic: Tools (software) to handle/modify/edit .xliffs and other CAT-generated files
Poster: Jorge Payan
Post title: My first choice: studioViews Plugin for TRADOS 2024

[url removed]

[url removed]

It is just a pity that the old stand-alone SDLXLIFF Split/Merge is not available for Studio 2024.

Tools (software) to handle/modify/edit .xliffs and other CAT-generated files | Thanks!

Jose Mota — Fri, 27 Feb 2026 06:43:00 +0000

Forum: CAT Tools Technical Help
Topic: Tools (software) to handle/modify/edit .xliffs and other CAT-generated files
Poster: Jose Mota
Post title: Thanks!

[quote]Nuria Morales wrote:

Hi José!

Did you try using the 'Trados Toolkit'? This is on the RWS App.

Best

[/quote]

Hi Nuria,
Thank you for your comment.

I have indeed Toolkit installed, but the tool only allows searching by segment status, and in this case it's more of a semantics issue.
Although a workaround (by changing the status in those segments of a particular area) should also be possible, so thanks for the tip ;-) )

Best wishes,
Jose

[Edited at 2026-02-27 08:18 GMT]

[Edited at 2026-02-27 14:26 GMT]

Tools (software) to handle/modify/edit .xliffs and other CAT-generated files | Thanks for the tip!

Jose Mota — Fri, 27 Feb 2026 06:33:00 +0000

Forum: CAT Tools Technical Help
Topic: Tools (software) to handle/modify/edit .xliffs and other CAT-generated files
Poster: Jose Mota
Post title: Thanks for the tip!

[quote]Recep Kurt wrote:

From their site ( [url removed] ):

"With Goldpan, you can:

Create or edit TMX file containing up to 8 separate languages
Create or edit TBX file containing 2 separate languages..." [/quote]

That looks like a very useful tool which I didn't know yet. I'll give it a try, many thanks!

Tools (software) to handle/modify/edit .xliffs and other CAT-generated files | Toolkit

Nuria Morales — Thu, 26 Feb 2026 19:16:00 +0000

Forum: CAT Tools Technical Help
Topic: Tools (software) to handle/modify/edit .xliffs and other CAT-generated files
Poster: Nuria Morales
Post title: Toolkit

Hi José!

Did you try using the 'Trados Toolkit'? This is on the RWS App.

Best

Tools (software) to handle/modify/edit .xliffs and other CAT-generated files | Goldpan Editor

Recep Kurt — Thu, 26 Feb 2026 15:46:00 +0000

Forum: CAT Tools Technical Help
Topic: Tools (software) to handle/modify/edit .xliffs and other CAT-generated files
Poster: Recep Kurt
Post title: Goldpan Editor

From their site ( [url removed] ):

"With Goldpan, you can:

Create or edit TMX file containing up to 8 separate languages
Create or edit TBX file containing 2 separate languages
Work with all language pairs simultaneously in a tabular form
Search and filter segments according with different criteria and regular expressions
Select and highlight different segments, or groups of segments
Block or protect segments from editing and deletion
Run a basic set of language checks, including commonly used quality assurance checks
Apply flexible filters and regular expressions to select segments for exporting
Remove HTML, XML, or RTF tags
Export selected language pairs from TMX or TBX file (export format)
Split any huge TMX or TBX file into smaller files of specified size
Export selected segments to TMX, TBX, XLIFF (v1.2), or XLSX file formats
Import TMX, TBX, XLIFF, SDLXLIFF, SDLTM, XLSX files
Merge several TMX (TBX) files into single TMX (TBX) file
Convert UTF-8-encoded tab delimited text file to TMX, TBX or XLIFF format
Process multiple files in a batch mode"

Tools (software) to handle/modify/edit .xliffs and other CAT-generated files

Jose Mota — Thu, 26 Feb 2026 15:24:00 +0000

Forum: CAT Tools Technical Help
Topic: Tools (software) to handle/modify/edit .xliffs and other CAT-generated files
Poster: Jose Mota

Dear all,
One of my last jobs was the translation of a company website in Trados Studio 2024. As it usually happens, the text in itself consisted of different sections: one (the biggest) was strictly technical, the second one related to legal issues (including data privacy, etc.), and the third one was of a rather commercial nature.

As I find some of the translated input quite interesting, I would like to divide the .xliff file (it's only one) in different small files by section (technical, legal & commercial) and then import each of the bilinguals into one of my three general TMs on those different areas.

Is there any piece of software that would allow to 'cut' or handle .xliff files like that (and maybe also the ones generated by memoQ, etc.)?

Look forward to your replies.

Smartling TM Maintenance: In-Platform or Export to Trados 2024?

Karla Matute Parini — Fri, 13 Feb 2026 17:49:00 +0000

Forum: CAT Tools Technical Help
Topic: Smartling TM Maintenance: In-Platform or Export to Trados 2024?
Poster: Karla Matute Parini

Hello everyone!

I've been trying to figure out the best way to manage and clean Translation Memories in Smartling. Does anyone have experience with this?

Do you clean up the TM directly in Smartling, or do you export the TMX to a CAT Tool like Trados Studio 2024 to edit/clean it and then re-import it? Have you encountered any segmentation issues when doing this?

Any tips or best practices would be greatly appreciated.

Irregular and unreliable Pastey | Lol

The Shredder — Sat, 03 Jan 2026 12:13:00 +0000

Forum: CAT Tools Technical Help
Topic: Irregular and unreliable Pastey
Poster: The Shredder
Post title: Lol

lol at the idea of anyone depending on Pastey for their work😱

Irregular and unreliable Pastey | Direct contact

CafeTran Trainer — Sat, 03 Jan 2026 06:10:00 +0000

Forum: CAT Tools Technical Help
Topic: Irregular and unreliable Pastey
Poster: CafeTran Trainer
Post title: Direct contact

From the FAQ:

If you would like to ask questions before checking out the forum, you can email us at [email removed] . If you have any questions about the membership tiers that give you access to Pastey, please visit go.proz.com/premium or get in touch with a member of staff.

Irregular and unreliable Pastey

MoonMartin — Fri, 02 Jan 2026 12:31:00 +0000

Forum: CAT Tools Technical Help
Topic: Irregular and unreliable Pastey
Poster: MoonMartin

Yet again Pastey lets me down by telling me my trial period is over, regardless of that I'm a paying Plus customer.
There's not any forum for it under software support which is surprising.
As the translation business has turned into a slave ship with falling prices, badly prepared projects and greedier multi-billion profit companies it is really important that tools we pay for work as expected.
Hopefully this continuing bug will be ironed out, but for now it has put me up sh*t creek.
Has anyone here help to offer?

Edit keyboard shortcuts in Translation Workspace XLIFF Editor? | Never mind

Ivan Rorick — Thu, 18 Sep 2025 12:01:00 +0000

Forum: CAT Tools Technical Help
Topic: Edit keyboard shortcuts in Translation Workspace XLIFF Editor?
Poster: Ivan Rorick
Post title: Never mind

Found it, it's under "keyboard accelerators," not "keyboard shortcuts."

Edit keyboard shortcuts in Translation Workspace XLIFF Editor?

Ivan Rorick — Tue, 16 Sep 2025 19:29:00 +0000

Forum: CAT Tools Technical Help
Topic: Edit keyboard shortcuts in Translation Workspace XLIFF Editor?
Poster: Ivan Rorick

I haven't used this tool in a long time. When I last used it maybe five years ago, I was able to edit the keyboard shortcuts. The tool is required by the client for a new job, and after installing the program, I tried to edit the keyboard shortcuts but couldn't find this function anywhere. Has the edit keyboard shortcuts function been removed?

Best tool for translating XML files | Here is a alternative

Ronika Kashyap — Thu, 21 Aug 2025 06:44:00 +0000

Forum: CAT Tools Technical Help
Topic: Best tool for translating XML files
Poster: Ronika Kashyap
Post title: Here is a alternative

You could try: [url removed]

Best tool for translating XML files | @Hazel

Samuel Murray — Wed, 13 Aug 2025 17:45:00 +0000

Forum: CAT Tools Technical Help
Topic: Best tool for translating XML files
Poster: Samuel Murray
Post title: @Hazel

[quote]Hazel Underwood wrote:
I've been translating XML files for a client using Trados Studio, but are there better tools out there? [/quote]
This really depends on the nature of the XML files. Are they bilingual files? Do all text outside of the tags need to be translated? Does some text inside of tags also need to be translated, e.g. attributes? Are there "internal tags", i.e. are there tags that occur in mid-sentence? These types of questions affect which tools may or may not be suitable for the translation. "XML" is a very large category of files and different XML files require different approaches.