To keep helptexts consistent in terms of content in multiple languages, a process is necessary by means of which changes can be identified in one language and the texts to be translated can be selected and the texts can be reimported into the data inventory after (external) translation.

If any content-related changes to the helptexts are made in one language on one occasion and some other language on another, it can no longer be safely decided in which language the text now actually corresponds to the functional requirements and what language entry needs to be updated accordingly.

Accordingly, the following workflow was arranged for DOKA-NG:

• The leading language for HLP texts is German.
• HLP files are submitted by developers as .XMU files.
• To the extent possible, or in the case of larger volumes of helptexts, before fine-tuning a second QA will carried out by the Secretariat staff, who will proofread, edit and correct the texts where necessary.
• The XMUs are checked in with SYSUPD.
• Import of the XMUs into the MyModelbank via SYSUPD.
• For translation purposes, the HLPs are periodically removed as a DBE file from the MyModelbank with the aid of the GENHLP transaction. Either all HLPs in DE or those starting from the last release date.
• The DBE is submitted for translation. In the process, only entries in the xml tags <NAM> and <TXT> will be translated; KEY fields, the language or version are not modified,
• The translated file is re-imported back into the MyModelbank with GENHLP; in the process, the INR will be ignored and the language changed from DE to EN. The change date is set to the date of the import; the translator's initials are arranged to be used for the User ID.
• A decision may be taken as to whether an English text whose change date is before the date of the German sentence to be overwritten or not.
• The translated and re-imported English helptexts must be checked to verify whether the maximum line length and line breaks are correct. To do so, start DBIHLP, launch INFHLP via [i] and search the sentences by entering the language, change date and User ID. Have the sentence at the very top displayed and then browse using [>] and check in each case whether the line lengths and line breaks displayed are correct. In the process, the translation itself can be proofread as well. If incorrect, take the current sentence using [Take] to DBIHLP, modify it there and then change back to INFHLP using [i]. Use [>] to browse to the next sentence without repeating a search. Again, enter the corrections again via XMU for check in.

If corrections are necessary to the existing English texts (for instance, to remedy typos), these can be submitted as an XMU file. They are also checked in using SYSUPD and imported into the MyModelbank.