Elefas is a long-term server-based translation memory: segments are saved in a Postgres database.

Initially the plan was to develop it as an element from Ruber suite, i.e. using Culter for segmentation, Spongiae for file filtering and with possibilty to export to Exilis. But mainly Spongiae and Exilis are far from mature enough.

So finally we publish a first version of Elefas written in Perl language. Compared to the planned Ruby version, the main limitation is the format for querying: you must send a file in text (TXT) format with exactly one segment (not a paragraph!) per line. You can use Culter to produce such a file if your initial source contains paragraphs, but until we launch a version in Ruby, this cannot be automated.

To import data, you need to produce a TMX file. This should not change even in a future Ruby version: such a version should support more formats only for querying.

The version in Ruby is not abandonned, only postponed. To avoid any confusion, we will then refer as Elefas Indicus for the Perl version and Elefas Maximus for the future Ruby version. Normally it should use exactly the same SQL schema, meaning that memories created with both versions may be compatible, but we cannot say this with certitude.