Inkscape

Entering Arabic text

Asked by pjaj

I have a Word document that contains both Roman and Arabic font text. I can cut and paste the Arabic words from this document to an Inkscape document but whilst the direction of the letters in the individual words is retained correctly (R to L) the word order is reversed. Text in Roman fonts retain both their correct letter and word order. This applies to mixed text as well.
For example, in word

Roman Roman Roman Arabic Words Roman Roman

when pasted into Inkscape becomes

Roman Roman Roman Words Arabic Roman Roman

where "Arabic Words" are in Arabic font.

I understand that Inkscape can handle "complex text" (ie both L-R and R-L fonts) but I can find no reference to this in the on line documentation. Am I looking in the wrong place? Am I going about it the wrong way, pasting from Word? Is there some add-on I need?

The process is:
1) Arabic speaking Colleague creates bi-lingual Word document.
2) I extract parts of it and convert to graphics in Inkscape. This can involve wrapping it along paths.
3) Graphics are used for web and print.

Up until now I have been using Illustrator for similar tasks on Roman fonts, but I would have to buy a new copy (£800!) to get Middle Eastern text capability.

Question information

Language:
English Edit question
Status:
Solved
For:
Inkscape Edit question
Assignee:
No assignee Edit question
Solved by:
mahfiaz
Solved:
Last query:
Last reply:
Revision history for this message
pjaj (peter-pjajennings) said : 		#1

Really weird, could be a Word problem after all. The following are two lines cut and pasted from the Word document. I don't know how they will render in your browser, but in the first one the two Arabic words are reversed, but in the second they are the right way round! The same thing happens in Inkscape. I don't know what's going on here.

Monday ﺍﻻﺜﻨﻴﻥ ﻴﻭﻢ (yawm al-ithnayn)

Friday ﻴﻭﻢ ﺍﻟﺠﻣﻌﺔ (yawm al-jumcah)

Revision history for this message
Best mahfiaz (mahfiaz) said : 		#2

Could you try pasting to and from abiword or openoffice or similar?

Revision history for this message
pjaj (peter-pjajennings) said : 		#3

I've opened the doc in OpenOffice Writer and the offending Arabic words are already reversed. So it would appeaar that the fault lies somewhere in Word, not Inksacpe. Whilst this does not solve my problem, it exonerates Inksacpe.

Revision history for this message
pjaj (peter-pjajennings) said : 		#4

Thanks mahfiaz, that solved my question.

Revision history for this message
pbhj (pbhj) said : 		#5

Could it not be that OOo Writer also gets it wrong? It sounds to me like a RLM Unicode Direction Character marker is not being copied/imported properly. That's probably not Inkscape but the clipboard of your WM?? For me on KDE4 right-click on the text entry tab of the <ctrl>+T menu allows me to enter Unicode direction markers.

Revision history for this message
pjaj (peter-pjajennings) said : 		#6

No I don't think so. I didn't cut and paste it into OO Writer, I OPENED the doc file directly it with OO Writer which understands Office word 2003 doc files. No clipboard was involved. Note that cutting and pasting into this forum suffers from the same reversal, well I hope you see it in your browser, I'm using Firefox 3 under Windows XP Pro.

Since only about 20% of the Arabic phrases in the bi-lingual document are word reversed and it's the same ones both in OO Writer and when cut and pasted into Inkscape, The fault must lie in the source Word document; unless Inkscape, Firefox and OO share the same code library for this feature (and nobody has found this bug in any of the millions of installations of fthose 3 programs).

I've found a work-round for this document where there were relatively few Arabic phrases (The translations of the days of the week, the months and the numbers from one to thirty one; ie a calendar) Monday and 22-28 were wrong, the rest were OK. I cut and pasted the Arabic back into the correct order.

However in future there will probably be more such documents and we need to find a better, permanent solution. With or without Word 2003.