Page tree
Skip to end of metadata
Go to start of metadata

Test with attached example file


and pre-translations of


and lookup in condordance search, what is actually returned by the service and re-export of the TMX proof the following

Short-comings compared to what is defined in the TMX 1.4b standard:

bpt, ept, ph, it tags

Code encapsulated by these tags is removed on import. Attributes are preserved.

This means an imported segment that looks like this

<seg>Tap <bpt i="1" x="215">&lt;strong&gt;<sub>Elephants<ph /> are big.</sub></bpt>View<ept i="1">&lt;/strong&gt;</ept> to<ph x="1" type="lb">&lt;br/&gt;</ph> get displayed<bpt i="1" x="215">code<sub>Elephants<ph /> are big.</sub>code</bpt>View<ept i="1" /> US patents.</seg>

returns like this

<seg>Tap <bpt i="1" x="215" />View<ept i="1" /> to<ph type="lb" /> get displayed<bpt i="1" x="215" />View<ept i="1" /> US patents.</seg>

sub tag

As tests with the supplied example data shows, sub tags are not handled and supported, because they are only allowed within a bpt, ept, ph, it and ut tag - and the content of those tags is ignored by the service.

ut and hi tags

In an exported tmx they return the same, was they have been imported, including code that may be contained and attributes.

Yet in a fuzzy search or concordance search they are not returned as tags. This means, the service does not treat them as tags.

They are NOT supported by the translate5-tm-service as of now.

  • No labels