The tabs dialog (e.g. Paragraph | Tabs on the menu) shows four kinds of tabs: Left, Right, Centered and Decimal. The "Left" and "Right" tabs are mis-named. When inserting a Left tab stop, in an LTR paragraph the tabbed text will be left-aligned; but in an RTL paragraph the text will be right-aligned. In other words: These are 'Start' and 'End' tab stops. LO doesn't currently support Left or Right tab stops. Please change the UI labels and the documentation accordingly. Note that this situation is the converse of what we have for paragraph alignment! There, Left and Right are what's actually implemented, but Start and End - aren't. I find that amusing somehow :-) Anyway, this is all further motivation for why bug 131192 needs to be worked on. We should have Left, Right, Start, End, Centered and Decimal as tab stop types.
Agreed. For the people ignorant of LTR/RTL duality I suggest to use "Start (Left)" and "End (Right)" resp. "Start (Right)" when the paragraph has RTL locale. Might be a bit tricky and just start/end are an improvement too. Code pointer: cui/uiconfig/ui/paratabspage.ui cui/source/tabpages/tabstpge.cxx and maybe class SvxFrameDirection
(In reply to Heiko Tietze from comment #1) > Agreed. For the people ignorant of LTR/RTL duality I suggest to use "Start > (Left)" and "End (Right)" resp. "Start (Right)" when the paragraph has RTL > locale. Might be a bit tricky Probably not that tricky, since they'll always see the same thing - such people dont typically work with RTL documents at all. Maybe you could even make the choice of labels depend on whether "complex scripts" are enabled or not; I think some RTL features depend on that. > and just start/end are an improvement too. Let's start with that then. > > Code pointer: > > cui/uiconfig/ui/paratabspage.ui > > cui/source/tabpages/tabstpge.cxx > and maybe class SvxFrameDirection I'll try and have a look this weekend maybe.
Akaki Dzidziguri committed a patch related to this issue. It has been pushed to "master": https://git.libreoffice.org/core/commit/2da684d43bbc19763ebd22300904c20a99461464 tdf#147814 Tabs dialog: rename Left/Right to Start/End It will be available in 26.8.0. The patch should be included in the daily builds available at https://dev-builds.libreoffice.org/daily/ in the next 24-48 hours. More information about daily builds can be found at: https://wiki.documentfoundation.org/Testing_Daily_Builds Affected users are encouraged to test the fix and report feedback.
Wait! We can't proceed with this patch. It turns out, that what we write aren't really Start and End tabs. We are actually using faux switching of tab directions, when switching paragraph direction. To see this, create a document with left tabs at 1, 2, 3 cm and right tabs at 4, 5, 6 cm. Then write an FODT. You'll get: <style:tab-stops> <style:tab-stop style:position="1cm"/> <style:tab-stop style:position="2cm"/> <style:tab-stop style:position="3cm"/> <style:tab-stop style:position="4.001cm" style:type="right"/> <style:tab-stop style:position="5.001cm" style:type="right"/> <style:tab-stop style:position="6.001cm" style:type="right"/> </style:tab-stops> forget about the inaccuracy. The point is that the left tabs - well, we don't know what they are, but the right tabs are actually right tabs. The same kind of switching occurs today, still, for paragraph text alignment, as a legacy from the (very recent) times when we didn't directly support Start and End. So, Akaki Dzidziguri, it would be more complicated than just switching labels. Moreover - there is also the matter of what happens when we are in tbrl or tblr writing mode, and how _that_ should be labeled.
Eyal's objection is legitimate and grounded in natural language, but given the text of the ODF spec and how it's been interpreted so far, what we have here is probably the most realistic approach. As part of this fix, I think we should also file a change request with the ODF committee, asking them to define the semantics of style:tab-stop style:type. Specifically, we should ask them to define "left" to mean aligned from the start of the text, and "right" to mean aligned from the end of the text, always relative to paragraph and layout direction. This would match our implementation, and would also make the ODF spec explicitly cover the TB cases which are presently underspecified. I'm less sure about the use cases for the actually-from-left and actually-from-right tab stops requested in the original post, but I've never even needed to use the other tab stop types before so I might not be the target audience. Maybe Eyal can elaborate on this request. (In reply to Eyal Rozenberg from comment #4) > forget about the inaccuracy. The point is that the left tabs - well, we > don't know what they are, but the right tabs are actually right tabs. 'left' is the default value, per the ODF standard.
(In reply to Jonathan Clark from comment #5) > As part of this fix, I think we should also file a change request with the > ODF committee, asking them to define the semantics of style:tab-stop > style:type. Specifically, we should ask them to define "left" to mean > aligned from the start of the text, and "right" to mean aligned from the end > of the text, always relative to paragraph and layout direction. This would > match our implementation, and would also make the ODF spec explicitly cover > the TB cases which are presently underspecified. Is it really a good idea to ask for a "counter-textual" interpretation? As opposed to a change to the ODF in its next version to say "start" and end"? > I'm less sure about the use cases for the actually-from-left and > actually-from-right tab stops requested in the original post, but I've never > even needed to use the other tab stop types before so I might not be the > target audience. Maybe Eyal can elaborate on this request. Well, I'll admit this is a bit speculative, but - this is basically like Left and Right paragraph alignment. Why do we even need those at all? * Historical reasons. * Other apps which doesn't support Start and End (and we can't be sure that we need to convert L/R into S/E) * Content in which the physical direction has other significance than the direction of text progression. Specifically: * Correspondence to background/foreground images * Correspondence to positions and directions in physical/geographical space
(In reply to Eyal Rozenberg from comment #6) > Is it really a good idea to ask for a "counter-textual" interpretation? As > opposed to a change to the ODF in its next version to say "start" and end"? Yes. The ODF standard doesn't properly specify tab stops for RTL text (see also ODF 1.3 19.512.3). Programs that implement RTL ODF have had to come up with their own ideas for how tab stops should work. We should encourage the ODF committee to formalize what existing implementations already do. In our case, we interpret tab distance and style as both relative to start, despite the ODF standard specifically saying they should both be relative to absolute left.
(In reply to Jonathan Clark from comment #5) > As part of this fix, I think we should also file a change request with the > ODF committee, asking them to define the semantics of style:tab-stop > style:type. I've taken care of filing a draft proposal to bring attention to this topic. Our left/right tabs really are start/end tabs, so I think we should follow through on the relabel. I tested the code change locally but couldn't see a difference, so it may need a second pass. Once that's done, we can mark this bug fixed.
(In reply to Jonathan Clark from comment #7) > Our left/right tabs really are start/end tabs, Functionally in LO yes, but vis-a-vis the ODT, we are using Left and Right tabs, and switching them when the paragraph direction is switched. So if we just renamed our tabs Start and End, that would be a kind of make-believe, wouldn't it?
Created attachment 205439 [details] Same tab style, LTR and RTL (In reply to Eyal Rozenberg from comment #9) > (In reply to Jonathan Clark from comment #7) > > Our left/right tabs really are start/end tabs, > > Functionally in LO yes, but vis-a-vis the ODT, we are using Left and Right > tabs, and switching them when the paragraph direction is switched. So if we > just renamed our tabs Start and End, that would be a kind of make-believe, > wouldn't it? What you're describing sounds sort of like bug 153953, but I don't think it applies here. I've attached a document with left/right/left tab stops in a base style, which is reused for both LTR and RTL. This document shows these tab types are really interpreted as start/end/start, without these values being manipulated at all. The same thing works if you use DF.
*** Bug 163675 has been marked as a duplicate of this bug. ***