Options for Language Templating Solutions
Please feel free to add other options to this list. _
Some things to thing about, While talking about technical solutions it doesn't hurt to have something resembling a specification…. (or wish list, whatever), so here goes:
Some libraries service diverse communitites, so we need a system that allows fror multiple languages in one koha installation which can be changed at the users discretion.
Sometimes the languages would all be displayed with the same template (ie they are in 100% the same HTML code, etc and just the text changes)
BUT! for some languages that would mean koha won't make sense (eg. some are read in directions other than right to left/then down), so Koha needs to be able to have a different template for those, as well - possibly one for each language if necessary.
(Examples of this sort of thing include: English and Chinese together, or English and Hebrew.)
Also… On this same note, images should be thought of in terms of translation if necessary to the interface, as should disabled-friendly features (eg. text only, or whatever else is necessary to accessibility). The latter feature mayn't sound all that important, but it isn't unusual for government regulations to require features of that sort.
[Note from Nick and the translation folks here: We'd naturally prefer something that would avoid us having to muck with code…. but are willing to consider things other than just flat text files to handle that – ie Kea, or otherwise. Something with the ability to preserve context – I know the ability to stick usernames, etc into output was mentioned here on the list discussion as well – is probably also a “good thing.”
[Comment from Ambrose] Images per se pose no problems with text browsers or other accessibility concerns, *provided* that we follow, e.g., section 13.8 of the HTML 4.01 standard regarding alternate text. It is just the additional effort that a translator would need to use an image editing program to finish some of the translations.
Images with roll-over or other Javascript actions is another matter completely, because creating such files require so much effort, they would probably be a real impediment to the translation. (E.g., the menu at the top-right-hand corner of the Intranet interface consists of more than a dozen layers; 30 files need to be generated, and to generate just one such file, the translators do not only need to move layers, but also change layer opacity).