In Tinderbox versions 5+, Unicode is supported in note text($Text) and all system/user attributes. For v5, the text engine was updated to give better support for Unicode, especially for double-byte languages (e.g. non-Roman alphabet languages).
However, some dialogs such as Create & Rename are still not fully Unicode compliant, as they rely on MacRoman (see more below) to display data. The same problem occurs for the key attribute display table in notes, and for value edit boxes in Info view. Thus although Tinderbox can now hold Unicode data in attributes, care needs to be exercised in how it is input and displayed. More specifically, with relation to Unicode data MacRoman cannot display (N.B. the list of affected UI elements is not necessarily complete):
- Create & Rename dialogs. Non-ASCII data will appear garbled in edit boxes. If 'OK' is clicked the corrupted screen version is written over the existing attribute value. If 'Cancel' is clicked no changes occur. So, if mistakenly opening the dialog, it is important to Cancel rather than click OK.
- Info view. Non-ASCII data will appear garbled in the right column though underlying source data is unharmed. If an attribute's value is edited via the view, the edit box works in the same way as above, and the source is corrupted.
- Key attributes in note text windows. Data is displayed garbled. Setting focus - even if no active edit is made - will overwrite the existing data with the garbled data.
A safe way to view & edit non-ASCII data is to use Edit-in-Place with Outline column view. Until these parts of the UI are re-written in a newer design tool (slated for v6) the limitations above will persist.
As part of the v5 text engine change, Tinderbox started a move towards storing $Text data as RTF rather than the current styled text. At first, the $Text data will be stored as both RTF and styled text which explains why there is an increase in v5+ TBX file sizes. These changes are likely to mean that v5+ created/saved files will be incompatible in whole or part for use with pre-v5 versions.
MacRoman. Used in Apple Macs from outset (and largely falling away since the move to OSX), MacRoman encodes 256 characters - like, but differing from ASCII. MacRoman shares the first 128 characters with ASCII and in the remaining parts some characters exist in both encodings but with different characters encoding numbers (e.g. some accented Roman characters). With OS X, the Mac moved to UTF-8 encoding, but some app UIs programmed in pre-OS X tools (such as the then-popular Metrowerks) may not reflect such changes and still use the legacy MacRoman encodings.
MacRoman fonts cannot represent characters outside the MacRoman basic set - especially languages using non-Roman alphabets. In Tinderbox v5, the latter will appear as strings of question marks or simply some garbled form of MacRoman characters.