Nudi Kannada To Unicode Font Converter

[Nhyiraba Valentin]

Ihave coded my xls form with both English and Kannada language options (with Label::Kanada and hint::Kannada columns in survey and choices sheets). The Kannada text was originally typed in a non-UTF font (Nudi o1 e). I converted it in Excel by selecting each line of text and changing the font it to a UTF-8 Kannada font (Nudi o5 K) that is compatible with both Mac and PC.

In Excel and Google (when uploaded to Google drive) the xls form displays the Kannada font appropriately, when uploaded to Enketo web-preview and when placed on the tablet, the Kannada font displays in an unreadable ASCII form.

I have been looking into this for a while now and have tried saving the xls file as encoded for UTF-8 before uploading to the ODK XLS to Xform converter. Is there anything I am missing in order to get the font to display correctly?

You received this message because you are subscribed to the Google Groups "ODK Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to opendatakit...@googlegroups.com.

For more options, visit

Thank you Yaw! However, the fact that it shows up in excel and google drive, and not the enketo web viewer makes me think there is something else wrong (i.e. it needed to be typed in a unicode font, not just converted). I also ran a survey last year using a Kannada unicode font and it worked well on the same tab. Therefore, I am a little mystified. Will take a look at the image based form link as well.

You received this message because you are subscribed to the Google Groups

"ODK Community" group.

To unsubscribe from this group and stop receiving emails from it, send an

email to opendatakit...@googlegroups.com.

For more options, visit

FWIW, there is no particular issue with the Kannada script in Enketo, and

no need to adjust fonts in Excel (in fact that may actually be the cause).

The issue is with your form (but I don't know what exactly). Somehow the

strings aren't copied correctly.

I am having a problem in Indian language kannada, some of the letters are not correctly being shown in Indesign but the same thing in microsoft office 2007 is being shown correctly i have attached the sample documents for Indesign and office 2007 with the kannada language font.

The fact that the font works in Word or other editors is no guarantee that it will work in InDesign. ID is much pickier in the fonts with which it will work. You will probably need to find an OpenType version of a Unicode compliant font for this language.

The font does appear to be properly unicode encoded, but for some reason ID just isn't seeing combining forms that use more than two glyphs even though at least one of the combinations is visible in the Glyphs panel, so I'm not sure where the problem lies, but since the same codepoints work in Tunga (but not in Arial Unicode, either), my guess is it's some sort of internal coding issue in the font.

Plenty of complex-script font developers work to the requirements not of the Unicode consortium or the OpenType spec, but to the quirks of a given rendering platform, e.g. Windows' usp10.dll which (used to be? is?) was the DLL responsible for Unicode support in Windows and Office. Alternately, some render only in Apple Advanced Typography (AAT) which is apparently a superior tech, but which only works in Apple apps.

I've been poking away at it for a while, but I can't get it to work yet. Like Peter, I can get it render correctly in other fonts, and I can get partway there if I manually rekey it (laboriously, because I really know very little about Kannada) but I can't get the Shree fonts to work in ID as they do elsewhere.

Still don't know what's going wrong, but maybe you can make something out of this.

How did I duplicated the single characters? By scripting in ExtendScript/JavaScript. I selected one "entity" (is it a word?) and ran the following script:

The OpenType features are all linked to the script "Kannada", but not to a particular language. How does InDesign select a script engine? Based on the Unicode values of the used characters, perhaps? This font seems to be unsure about its own encoding; my font viewer reports conflicting names vs. unicodes. The Double Danda, for example, has the proper Unicode U+0965, but in this font it has the name "u0CAE".

@Jongware The OpenType language and script are provided by the language, as addressed by the matching text attribute. If you have the SDK around, that's ILanguage::GetOpenTypeScriptTag() and GetOpenTypeLanguageTag().

That strange glyph name smells like a conversion from an older format / encoding. It might lead to funny effects - I read that under some circumstances Acrobat back-guess characters from glyphs using their postscript names.

I lost a hefty comment on this thread. I think that Theun's "possibly irrelevant observation" is actually the answer, as is Ellis' observation on the MS Office side. A font that declared its language correctly, if used in this workflow, would allow the language to be marked in Word in such a way that it would be marked correctly in ID upon import, and (probably) would get the Kannada glyph-stacking right.

There is more than one font from this type foundry that bears this kind of evidence of font file format conversion. So Dirk's suspions are correct as well. Raghugada, if you return to this thread, I'd advise that you just send the link to the thread to the people over at Modular Infotech, who (I'm told) regularly update their vast library of fonts.

3a8082e126