Jump to content


Photo

Direction Change Indicators showing up


  • Please log in to reply
12 replies to this topic

#1 Karyn Traphagen

Karyn Traphagen

    Member

  • Active Members
  • Pip
  • 39 posts

Posted 08 August 2006 - 06:07 PM

I'm exporting to RTF a section of parsed Hebrew text with the Unicode preference selected. When I import the RTF file into Mellel it displays as Lucida Grande. When I replace the style with SBL Hebrew small directional arrows are inserted (and print) whenever there is a change in direction. For example, on each line there is some Hebrew and some English text. The Hebrew is R->L and the English is L->R, whenever there is a change in direction it is marked by a small elevated arrow. Is this due to a code that Accordance puts into the RTF file? If so, how can I get it to not display/print?

Thanks, Karyn
Karyn Traphagen
Doctoral Student
University of Stellenbosch (South Africa)

www.boulders2bits.com

#2 jpkang

jpkang

    Platinum

  • Active Members
  • PipPipPipPipPip
  • 577 posts
  • Gender:Male
  • Location:Seattle, WA
  • Interests:Comparative Semitics; ancient Hebrew semantics; Hebrew syntax; history and religion of ancient Israel; history of interpretation.
  • Accordance Version:10.x

Posted 08 August 2006 - 09:06 PM

I see what you're seeing. For the time being, you can copy each one of those arrows (easier to select with shift-arrow key than a mouse/trackpad) and do a Find/Replace on them to get rid of them. Or, you can replace one of them with a tab character ^t if you want to be able to line up the output a little better.

I took a look at the SBL_Hbrw.ttf font and saw that it actually has glyphs for the Right-to-Left Mark (x200F) and Left-to-Right Mark (x200E), whereas most other Unicode fonts do not have entries for them (since they should really be invisible). This explains why these arrows appear only with SBL Hebrew. You would need an OpenType-capable editor (e.g., FontLab) to remove those glyphs (alas, I only have the lightweight TypeTool and midweight Fontographer, neither of which can manipulate OpenType tables).

So I'm pretty sure it's a font issue, not Mellel nor Accordance.
J. P. Kang, Ph.D. (Bible)

#3 Karyn Traphagen

Karyn Traphagen

    Member

  • Active Members
  • Pip
  • 39 posts

Posted 09 August 2006 - 01:42 PM

Thanks for the very helpful explanation. I'll do the search and replace for now and then look for a longterm solution along the lines of your suggestion.
Karyn Traphagen
Doctoral Student
University of Stellenbosch (South Africa)

www.boulders2bits.com

#4 Karyn Traphagen

Karyn Traphagen

    Member

  • Active Members
  • Pip
  • 39 posts

Posted 09 August 2006 - 01:54 PM

Actually, now that I try to replace these arrows it rearranges the order of the parsed text. I think the arrows need to stay in there for it to display the direction correctly. I can delete the arrows (which yields the unsatisfactory changes in order) but I cannot change the arrows to "white" so they would seem "invisible." Any other suggestions? Thanks.

Since this now appears to be more a font issue, I'll take further discussion to the SBL Hebrew list.
Karyn Traphagen
Doctoral Student
University of Stellenbosch (South Africa)

www.boulders2bits.com

#5 jpkang

jpkang

    Platinum

  • Active Members
  • PipPipPipPipPip
  • 577 posts
  • Gender:Male
  • Location:Seattle, WA
  • Interests:Comparative Semitics; ancient Hebrew semantics; Hebrew syntax; history and religion of ancient Israel; history of interpretation.
  • Accordance Version:10.x

Posted 09 August 2006 - 02:50 PM

FYI, I replaced the right-headed arrow with the ^t sequence and deleted the left-headed arrow, and it did not switch the order of the columns of the text. See a screenshot.
J. P. Kang, Ph.D. (Bible)

#6 Thomas

Thomas

    Bronze

  • Active Members
  • PipPip
  • 81 posts
  • Gender:Male
  • Location:Pasadena, CA

Posted 19 March 2009 - 04:10 PM

The issue is the presence of the text direction indication arrows:
200E (LTR) LEFT-TO-RIGHT MARK
200F (RTL) RIGHT-TO-LEFT MARK

I. Right to Left Language Modules that exhibit the directional arrows issue.
1. BHS-W4 (NB TARG-T, TARG2-T, TARG3-T, DSSB-C, DSSB-M, QUMRAN, INSNWS, INSCRIP behave the same as BHS-W4, except the book names only appear in English).
a. Copying text including the English book name will place at RTL at the front of Hebrew text, and then each additional English book name will have a LTR, and again an RTL maker for the Hebrew.
b. Copying text including the Hebrew book name will only place one RTL mark at the beginning of the copied text.

2. BHS-GBS behaves like the BHS-W4
a. Since this module has footnote indicators, it behaves like 1a above, and also LTR and RTL indicators around the superscripted Latin characters.

II. Right to Left Language Modules that do not exhibit the directional arrows issue.
1. PESHNT-T
2. DHNT, MHNT

III. Untested
1. Arabic,
2. Peshitta with Hebrew characters

These findings suggest that there has been a recent change to the application since never before have the directional indicators appeared in Unicode-copied-and-pasted text out of Accordance into Mellel.

Thomas

#7 jpkang

jpkang

    Platinum

  • Active Members
  • PipPipPipPipPip
  • 577 posts
  • Gender:Male
  • Location:Seattle, WA
  • Interests:Comparative Semitics; ancient Hebrew semantics; Hebrew syntax; history and religion of ancient Israel; history of interpretation.
  • Accordance Version:10.x

Posted 19 March 2009 - 04:37 PM

If you read my explanation from August 2006 above, you'll see that in fact, this behavior was actually evident three years ago in Mellel (and still is) with certain fonts (e.g., SBL Hebrew).
J. P. Kang, Ph.D. (Bible)

#8 Thomas

Thomas

    Bronze

  • Active Members
  • PipPip
  • 81 posts
  • Gender:Male
  • Location:Pasadena, CA

Posted 19 March 2009 - 04:44 PM

Let me reframe the issue. The issue is that these directional arrows are being exported in only some right-to-left modules, but not others. If text from a module that has these arrows is copied, then they appear when pasted—which for my experience is a very recent change to exports before say maybe two months ago.

The most valid question here is if there are right-to-left language modules that do not exhibit this behavior, then should the ones that do be repaired to behave the same way?


[edited for clarity]

Edited by Thomas, 19 March 2009 - 05:11 PM.


#9 jpkang

jpkang

    Platinum

  • Active Members
  • PipPipPipPipPip
  • 577 posts
  • Gender:Male
  • Location:Seattle, WA
  • Interests:Comparative Semitics; ancient Hebrew semantics; Hebrew syntax; history and religion of ancient Israel; history of interpretation.
  • Accordance Version:10.x

Posted 19 March 2009 - 06:08 PM

Which Unicode Hebrew font(s) are you using?

I just exported from BHS-W4, BHS-GBS, DHNT, PESHNT-T, HPESHNT-T, and ARAB to Mellel, and I only saw the arrows when I used SBL Hebrew and Ezra SIL for Mellel's "Main font" (the marks vanish when I set the Main font to a Latin font with the exception of Times New Roman, and use the appropriate script and font for the Secondary font). None of the Syriac Meltho or Arabic fonts I used showed the directional glyphs.

You can confirm that the directional marks are in the export by pasting from Accordance into a text editor like BBEdit or TextWrangler and then performing a hex dump to inspect the byte pairs.

So I am still convinced that the appearance of directional arrows is a font issue (albeit one with a resolution in Mellel if you set the Main and Secondary fonts appropriately), not an Accordance module or export behavior issue.

For reference, I used Accordance 8.2.1, Mellel 2.6.1, SBL Hebrew 1.52, Ezra SIL 2.51, Lotus Linotype 1.10, and Estrangelo Edessa 1.01.
J. P. Kang, Ph.D. (Bible)

#10 Thomas

Thomas

    Bronze

  • Active Members
  • PipPip
  • 81 posts
  • Gender:Male
  • Location:Pasadena, CA

Posted 19 March 2009 - 06:32 PM

Thank you for following up on this, JP. I have the same versions as you do.
Your mention of Time New Roman was key: "I set the Main font to a Latin font with the exception of Times New Roman" (my emphasis). There lies the issue indeed! I haven't used Times New Roman for years, and was asked to submit something in it, which is why this "issue" seemed new. And here I was looking forward to using this Latin typeface again. Well, it's back to Charis SIL for me then set at 11 pt to match Times 12 pt. I'm working on enhancing Gentium Book Basic by adding glyphs that are needed for ancient Near Eastern transliteration (e.g. āēīōū ăŏĕ šḏṭġḫḥṣṯˀˁ and the Egyptian aleph "3"). There are many that can do this in regular/normal, but then very few comes with, well, fonts in italic, bold, and bold italic.

Off topic: Since you're into ancient Near Eastern things, you'll appreciate my having made a NA input method which uses Karel Piska's NA font (plus my PUA additions since the Unicode table is missing massive amounts of Akkadian forms).

תודה רבה

#11 Thomas

Thomas

    Bronze

  • Active Members
  • PipPip
  • 81 posts
  • Gender:Male
  • Location:Pasadena, CA

Posted 20 March 2009 - 02:24 PM

Hi J.P.
I've thought about this a bit more, and want your thoughts.

I ended up adding a bunch of glyphs to GentiumBook so that I could have a comparable typeface that appears similar Times NR.
As this has solved in hiding of the characters u+200E and u+200F (the arrows), I can't help but return to an overarching important question: if some of the Accordance RTL modules do not contain the u+200E and u+200F characters, then why can't the remaining RTL modules that still do have these characters be updated by removing them? The inclusion of these characters in (Unicode) copied text curbs a wider freedom in choice of typeface within Mellel—I know, I know, how is this Accordance's problem? Just that again, there are already RTL modules in Accordance that do not have these characters, e.g. the Syriac Peshitta. I'm advocating, or will advocate, that the remaining RTL modules follow suit and see the removal of these (perceived to be) extraneous characters. Sometimes the fun in using work-arounds wears away.

What thinkest thou?

לשום לך


edit: removed quote

Edited by Thomas, 20 March 2009 - 02:24 PM.


#12 jpkang

jpkang

    Platinum

  • Active Members
  • PipPipPipPipPip
  • 577 posts
  • Gender:Male
  • Location:Seattle, WA
  • Interests:Comparative Semitics; ancient Hebrew semantics; Hebrew syntax; history and religion of ancient Israel; history of interpretation.
  • Accordance Version:10.x

Posted 21 March 2009 - 04:29 PM

Hi Thomas,

Thanks for your messages. I did a little further digging, and I can now confirm what you were saying about the RTL marks not appearing in Mellel for the Syriac. I'm sure that these are NOT in the modules themselves, but are generated at the time the text is copied onto the pasteboard, so it should be a matter of Accordance modifying its output behavior to be consistent for all modules containing right-to-left text.

In any case, Mellel does not appear to need these to render the text correctly (though they may well be in there to avoid ambiguity and/or to adhere strictly to the Unicode spec). Interestingly, the Unicode bidi algorithm itself states in 2.4 that U+200E and U+200F "act exactly like right-to-left or left-to-right characters, except that they do not display or have any other semantic effect" (which begs the question why Times New Roman and SBL Hebrew include glyphs). Perhaps Ori and Eyal can be persuaded to ignore them in Mellel's rendering routines, which would solve the problem in a different way!

For now, why not edit your copy of Times New Roman to omit U+200E and U+200F?

J. P.

p.s. I'd love to see your NA input method and modified font: jpkang [at] accordancebible [dot] com.
J. P. Kang, Ph.D. (Bible)

#13 danzac

danzac

    Gold

  • Active Members
  • PipPipPipPip
  • 275 posts
  • Gender:Male
  • Location:Acadia Divinity College
  • Interests:NT, historical Jesus, DSS, NT Backgrounds

Posted 08 April 2009 - 09:43 AM

I recently stopped having this arrow problem when pasting from Accordance into Mellel. The difference, for me anyway, has been the choice to "paste as plain text" as opposed to a standard paste. Then I highlight the text and choose my Hebrew font (SBL Hebrew). When I do this, the directional arrows disappear. If I do a regular paste, the arrows are there.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users