0 votes

I am trying to fine tune a .tec encoding converter to make it round trip safe for a Left-to-Right transliteration of a Right-to-Left language project. (This enables comparison with the original to make sure that tweaks to the orthography on the Latin script side have not introduced any changes to the text.)

The chief obstacle to this is the RTL mark that occur after colons and hyphens in verse references. It is easy of course to remove the RTL marks when moving from RTL to LTR; but inserting them on the trip home requires a bit more fiddling. I can do it with a regular expression, but I would like to make it automatic for the sake of those who are not familiar with RegexPal.

Any suggestions of how to do this?

Paratext by (175 points)

3 Answers

0 votes
Best answer

I find it odd that the rest of my email was deleted. Here are the three things to try:

  1. I do not see any \ior…\ior* markup in your screenshot. Try adding it around your references in the original project. You should be using it anyway for proper formatting.
  2. Try changing in and out of the different views – from standard to unformatted and back again in your main Paratext text window. This can trigger Paratext to add such things.
  3. It might be because a converted project is not editable by design and therefore Paratext itself is not editing it either. Try creating an auxiliary project and importing the converted text and seeing if the change happens. This is actually a good practice anyway as well because there are often a number of things you want to do to the project manually before publishing which you cannot do to the original, uneditable conversion project.
by (1.3k points)

Thanks again, Shegnada.

  1. Paratext won’t let me insert \ior in the original project. I can add it to the transliterated project, but then the number sequence displays completely LTR instead of mixed.

  2. I have cycled through all of the views, as well as all three selections under Project Advanced/Normalization (I don’t understand what that is, but it was a setting I don’t recognize, so thought it might be related.)

  3. BHDyMT2L is already an editable auxiliary project.

I tried a couple more things:
a) I tried retyping some of the number sequences, but this did not change anything
b) I tried manually adding RTL markers to specific number sequences in BHDyMT2L, but they did not seem to “stick”; the compare pane still showed an RTL on the BHD side, and none on the BHDyMT2L side.
c) I tried right clicking on a difference and selecting “revert change”, but this also did nothing

Finally, I closed PT 9 and opened up PT 8 (which wanted to do an update, which I allowed) and opened a Compare window, again with BHDyMT2L on the left and BHD on the right. Now I was seeing a lot of WORDS lit up as if they were different, but could not see any visible difference between them, even with “show invisible characters” toggled on. But I chose one of the number differences, right clicked on it, selected “revert text” and behold, ALL the number differences went away. The word differences did not (unless I did a right click-revert text on them). but decided to try again in PT 9. This time I opened the Compare pane with BHD on the left. Now the only differences showing are the real differences I was looking for in the first place, and a number of (apparently) random places where BHD has RTL marks AFTER a number, and won’t let me delete them.

So: for the moment, it does indeed seem that Paratext has finally decided to insert the RTL marks in the references of the transliteration-based auxiliary project; but the way it has arrived at this point is still way too spooky to be useful…

-DVM

I’d ask for help via Give Feedback on the Help menu. Sorry.

Blessings,

Please explain what happens when you try to add \ior markup. Does the markup turn red? Do the references not appear correct in some other way? Perhaps you just need to fine tune your style sheet.

I tried again, and this time was able to insert \ior 's. They were red because they followed \io1, which I see is now supposed to be just \io. I changed \io1 to \io and added \ior…\ior* around the references in the book intro for MAT.

This did not immediately seem to change the display of the references, though today the RTL markers seemed to be present. A “compare versions” on the two projects showed no difference, but in a visual comparison, the order of the numbers following a colon or hyphen in the transliterated project was still reversed.

I tried copying the Scripture Reference settings from the original project, and then did “Copy Books” from the back-transliterated project (BHDxMT2L) to an auxiliary project BHDyMT2L). Now the references in both projects display correctly(!)

But for some reason the transliterated projects do not like having an alphabetic separator between a range of chapters. Attaching a screen shot in case you are interested…

0 votes

Hi DVM,

Paratext automatically inserts these characters. In fact it is very hard to get rid of them if you don’t want them. Have you confirmed that the automatic insertion is not working? One key could be to make sure that all references are properly mark as references using \r \ior or \xt…\xt*. I believe there are a few more options but those are the most common ones.

Blessings,

by (1.3k points)
reshown

Hi Shegnada,

Please see the attached screen shot: BHDxMT2L is created by running an encoding converter on a manually adjusted transliteration of BHD. As you can see, there are differences (RTL marks) in all number phrases involving colons or hyphens, including those following \r and \fr.

When I tested it several months ago (and maybe before moving to PT 9?) I recall that Paratext did indeed insert the RTL marks. Thinking that this was not going to be an issue, I went ahead and deleted the RTL marks from the Latin-script transliteration. But now that I have finally gotten to the point of needing the comparison, Paratext no longer seems to do this. Perhaps there is a way to turn the insertion off or on?

Dear DVM,

I haven’t seen any new behavior in this regard in 9.1, though we have asked for some more control over these characters in the past and they may have responded. You could test this by trying to add or remove the characters manually in a non-conversion RTL project.

But I have three things you could look at before reporting a problem via “Feedback” in the Help menu.

0 votes

I’ve no idea if this is helpful, but if you are using RTL marks specifically for scripture references, (eg: John 2:1), I have found the most reliable solution is to have a space alongside the colon. In English, that would be “John 2: 1”. This orders the chapter number and the verse number in the order that you expect, based on the direction of the surrounding text. Some Arabic style guides recommend this.

by (440 points)

Thanks for this suggestion. I suppose it would be possible to add spaces, although we would need to research reader acceptance. People seem happy with the current format of the published version, which has no space after the colon (or hyphen, which also affects numeral direction).

DVM,

If adding the space does the trick for you, you do not need to worry about the output. With many outputs you can make it a very, very thin space and no one will see it. But I would say that it is quite key that in Paratext all your references validate so that they can be made into links in online publishing. I am unable to reply directly to you, but know someone who can advise you on your mapping and is quite comfortable with RTL issues. If you write me directly, I will reply with their information.

Yes, I am afraid that adding spaces in the reference format might complicate the validation. A quick test just now was not promising: Scripture References dialog box accepted a colon-space combination as a chapter-verse separator, but it gets marked as invalid in the text.

Meanwhile, I have done some experimenting, and was able to find a reproducible path to get Paratext to insert RTL marks: I have to set up set up the Scripture Reference settings (which I did by copying from the original project.) If I create a new Roman to Arabic transliterated project, Paratext does not insert the RTL marks until I do this step. So that seems to be the key. I don’t know why this should be so, but it would be good for it to be documented in the help menu.

But the suggestion of inserting a space gave me an idea for the Arabic>Roman transliteration. I added a line in the .tec file that changes an RTL mark to a zero_width_space (U+220B). This makes the references in the Roman text display correctly. It seems to work the other way around as well, but this is not terribly relevant. The Roman text in this case is for the consultant to read, and will not be published.

So I think we have a workable solution for both directions of the transliteration. Thanks for all of your input!

“Scripture References dialog box accepted a colon-space combination as a chapter-verse separator, but it gets marked as invalid in the text.”

You should check your character and punctuation inventories. Perhaps the colon or the colon space sequence is marked invalid there. You should only have a problem if the colon space sequence is illegal in other types of text.

Blessings,

Turns out that when I replaced the RTL with a space in the search and replace, it put the space before the colon. When I turned that around the references show up as valid.

So yes, it seems so far that using a space would be a viable option.

Related questions

Welcome to Support Bible, where you can ask questions and receive answers from other members of the community.
Very truly I tell you, whoever accepts anyone I send accepts me; and whoever accepts me accepts the one who sent me.
John 13:20
2,477 questions
5,170 answers
4,866 comments
1,283 users