The Qalculate! developer(s) are thankful for every possible translation and willing to answer any questions or help with any problems you may have as a translator. Please read through this document before starting the translation.

Qalculate! is split up in one library package (libqalculate, which also includes the text interface) and five GUI packages -- qalculate-kde for KDE, qalculate-gtk for GTK+, and the other for small GTK+ utilities (with only a total of 46 messages).

The GUI packages contains translations for the GUI text as usual in the po directory. The libqalculate translations is split up in two -- translation of error messages, type information and the text interface in po, and translation of definitions of functions, variables, units, etc. in po-defs (merged with the data files in data).

The best strategy is probably to start with qalculate-gtk or qalculate-kde. These contain approximately 700 messages which are essentially the same in both GUI's. The most difficult and demanding task is no doubt the translation of all definitions. It's by no means necessary to translate every package.

The source packages for libqalculate, qalculate-gtk, qalculate-bases, qalculate-currency and qalculate-units do not include any .pot files. Instead the pot file is created with "../intltool-update --pot". You can then run "msginit --locale=??" or just copy the pot-file to ??.po. The translation file is updated with "../intltool-update language code" (ex. ../intltool-update sv, for the Swedish translation) from the po directories. intltool-update is generated when ./configure is run from the top package directory.

qalculate-kde uses a slightly different system and the package includes a pot-file which is copied and edited to create a new translation.

The translation of the libqalculate message strings in po-defs require some further knowledge. These are merged with the xml data files in libqalculate/data. The different types of message strings in these files are:

Data objects: Values of text properties
Data sets: Categories, titles, names, descriptions, property titles, property names and property descriptions
Functions: Categories, titles, names, descriptions and argument titles
Variables/constants: Categories, titles and names
Units: Categories, titles and names

To distinguish equal text strings with different meaning (for example "Mercury" as a planet and as a chemical element), some text strings is preceded by "!type!" (for example "!units!" for units). Everything between and including a exclamation mark in the beginning and the second exclamation mark is not displayed in the application. Translators do not need include to include "!type!" in the translation.

Names require some special considerations from the translator. The names are used to reference an object in a mathematical expression and to display them in a result. An object can have any number of names but each must be fully unique, with the exception that functions can have names used by other types of objects (for example "min" is used as a name for the minute unit but also for a function returning smallest value in a vector). The translation does not need to have the same number of names as the the untranslated object.

There are some characters which are reserved for other purposes (mainly mathematical operators) and may not be included in names. These are
- ~ + - * / ^ & | ! <> = ' @ ? \ { } " . , ; : ( ) [ ],
- spaces, tabs and newline characters, and
- unicode signs for operators.

No name may begin with a digit and units are not allowed to contain a digit anywhere.

All names of a object is contained in one text string. The names are separated by commas (','). The order of names decides their priority, which name is by default used for display and input. Each name may be preceeded by a a number of properties before a colon (':'). The format of the names strings is thus [properties1:]name1,[properties2:]name2,... The possible properties are
- 'a': the name is an abbreviation,
- 'c': the name is case-sensitive,
- 'i': the name should not be presented to the user for input (but looks good for display),
- 'p':the name is in plural form,
- 'r': the name is used as reference,
- 's': the name ends with a suffix (the text after the last '_' will displayed as subscript with the '_' removed, often used in combination with 'i'), and
- 'u': the name contains a unicode sign.

All properties is off by default, except for names with only one sign which are interpreted as abbreviations by default, and for abbreviations which by default are case-sensitive. A property can be turned off by preceeding it with a minus '-'.

The reference property tells that the name will be available to all users regardless of the language used. If a reference name is not included in the translated list of names, it is appended at the end by the program. The translator can inlcude the reference names to select the position of the name and change all properties except 'r', 'c' and 'u'. All untranslated first names has the reference property explicitely specified, so names can easily be distinguished from other text strings by the preceeding properties and colon.

Most functions and variables have only one name. Units usually have an abbreviated name, a normal name and a name in plural form. Units should always have a name in plural form if it differs from the singular form.

Names string that is the same translated as untranslated should preferably not be copied but left empty or replaced by "-" (so that it is marked as translated), to allow the size of the definitions data files to be reduced.

Property names works like the names described above except that 'r' for reference is the only possible property.

Below follows some examples from the Swedish translation.

msgid "ar:m,meter,p:meters,metre,p:metres"
msgstr "a:m,meter"
The meter unit has the abbreviated reference name "m" (case-sensitive), a normal name "meter" and a plural form "meters". Meter does also in addition have a different spelling in English. The Swedish name for meter is the same as in English but the plural form is the same as the singular, and no alternative spelling exist.

msgid "r:mean"
msgstr "medel"
The name of the mean function is translated to "medel", but since "mean" is specified as a reference name it is appended by the program as the second name (results in "medel,r:mean").

msgid "ar:SEK"
msgstr "a-c:kr,a:SEK,krona,p:kronor"
The English version can only use the three character code for the Swedish currency since the other names can be used for other currencies, but the translation can assume that the user refers to the local currency.

msgid "-r:billion"
msgstr "miljard,p:miljarder"
The constants for large numbers follows two different systems in different countries. With the "short scale" billion means 1e9, while it with the "long scale" means 1e12. Therefor, these constants do not have any reference name. The short scale is used in most English-speaking countries, while in Sweden the long scale is used.

msgid "r:sqrt"
msgstr "-"
The untranslated string is kept.

qalc commands and options to commands in libqalculate/po/ follows the same principles as reference names, meaning that the untranslated message string is kept in addition to the translated.