Re: [xliff-seg] a scenario

[Eiju Akahane <AKAHANE@jp.ibm.com>]

I think one of the most important reason we need "Segmentation" is the
translation memory we have in our repository depends on the segmentation
rules.  We need to perform segmentation with specific rule before applying
memories.

Are there any other scenario?

Eiju Akahane, Yamato Software Lab., Software Group, IBM

Eiju Akahane/Japan/IBM wrote on 2004/03/09 21:33:18:

> I am thinking a scenario.  This will separate the needs of
> segmentation and memory from translation editor.  This will increase
> the flexibility of the choice of translation editors.
>
> 1. Development to produce XLIFF with or without previous translation
> 2. "Segmentation" logic to make segmentation in XLIFF
>    - format will be defined in the sub-committee
>    - segmentation rule can be supplied as SRX
> 3. "Memory Insertion" logic to apply translation memories from
> memory repository by segment basis
> 4. Translator to translate it using a XLIFF aware translation editor
> 5. "Memory Extraction" logic to extract newly translated text at
> <target> and store into the memory repository
> 6. "Segment Concatenation" logic to restore original <trans-unit> --
> this step might not be required
> 7. Development team to receive the translated XLIFF
>
> Eiju Akahane, Yamato Software Lab., Software Group, IBM