Exporting Interlinear XML / Writing systems

[Alexandre Arkhipov]

Hello,

 

Here are some issues that my colleagues and I noticed with formatting and exporting interlinear texts. (Please tell me if I'm missing some features that are already implemented.)

 

1) Vote for a command to export all (or selected) interlinear texts in the project at once. The user would select all or some texts, the representation to export (gloss/analyze/...), the destination format, and the way of naming the resulting files (e.g. names based on the abbreviation field in the Info tab).

We now have some two dozen texts in a project that we analyze in FLEx and then export to XML and convert them to ELAN (and also upload on LAT server for online access). Now and then we make some changes to glosses, translations etc., so that we need to export all the texts again. It already takes some time, but will be worse later this year when the number of texts will be about fifty.

 

2) When exporting a text into Interlinear XML, some fields from the info tab are exported as well, which can be very useful, but the choice of fields is very strange. I would expect all non-empty fields (title, abbreviation, source, comment; genres, is a translation) exported by default; we're most often using at least title and abbreviation.

In fact, of all those fields, only title and abbreviation are exported from the Gloss and Analyze view, and the comment field is also added while exporting from the Print view.

 

Furthermore, these three fields can be filled in any or all of the project's writing systems. We've got five writing systems (two for the vernacular language and three for the analysis). However, only one of them is really exported for each field:

    <item type="title" lang="kjj__LING">title-vernacular-1</item>
    <item type="title-abbreviation" lang="kjj__LING">abbr-vernacular-2</item>
    <item type="description" lang="en">comment-english</item>

For title and abbreviation, it is the first writing system for which the field is non-empty (so if I fill in the text title in vernacular language, I cannot have its English title exported). Mysteriously, for the comment field always and only the English version is exported, even if it is empty.

 

3) Again with the writing systems:

Sometimes if one wrongly switches the keyboard layout while typing in a text field (e.g. free translation), the field gets the wrong writing system associated with it, and it seems impossible to restore the original one. Even if you erase the field's contents, apply the right writing system from the menu and switch the keyboard layout, you can only type with the right keyboard -- but as soon as you leave the field and return to it later, the incorrect writing system and keyboard layout are back there. This also affects the export (the lang attribute is set wrong), and if we haven't noticed such a field, our ELAN export can fail.

The only way to restore the writing system in sych cases seems to be to Delete Free Translation for a given sentence (and this means delete them in all the writing systems at once) and retype them again.

 

4) There seems no way to apply custom styles to different lines of interlinear. It is sometimes very useful to set up different fonts or font properties to different lines regardless of (or in addition to) the writing system properties. E.g. I'd like the translations to be in italic and of bigger size than the lexical gloss lines that use the same writing system.

 

Hope some of these points can be addressed in a not-too-remote release.

 

Thanks a lot for developing such a gorgeous tool,

Alexandre