Created attachment 52184 [details]
The preposition “s” may not stand in the end of line in Czech, thus no-break space is used after that
The character “NO-BREAK SPACE” (U+00A0) is incorrectly interpretated like
“fixed width no-break space”. The fixed width is impedimental there, to keep the
typographical rules and create a nice document with the block justification,
there is a need of a normal, flexible non-breaking space after some characters
(in LaTeX “~” has this function). Now, the space after U+00A0 is narrower than other spaces in the same line, which is ugly.
[This is an automated message.]
This bug was filed before the changes to Bugzilla on 2011-10-16. Thus it
started right out as NEW without ever being explicitly confirmed. The bug is
changed to state NEEDINFO for this reason. To move this bug from NEEDINFO back
to NEW please check if the bug still persists with the 3.5.0 beta1 or beta2 prereleases.
Details on how to test the 3.5.0 beta1 can be found at:
more detail on this bulk operation: http://nabble.documentfoundation.org/RFC-Operation-Spamzilla-tp3607474p3607474.html
Still present in 3.5.0.
If I remember correctly, this is a very old problem already present in OOo 3.x and maybe even older (therefore I changed the 'Version' picker back to the oldest available version). It is also not limited to Czech typography, but the Czech sample shows the severity of the problem especially well.
Writer interprets U+00A0 like a 'fixed width no-break space', but users with some experience in typesettings or DTP applications (TeX with all variants '~', QuarkXPress and others) expect that U+00A0 should behave completely like an ordinary space character (U+0020), i.e. it should be extended and compressed in justified paragraphs like an ordinary space character, it just should not break at the end of a line ...
I have never complained about this problem because I assumed it is just a clash of expecations: (many) Word processor users are used to the current behaviour, i.e. to the fact that U+00A0 behaves like a “fixed width no-break space”, while users with some experience in typesetting (like me and, obviously, dohnp5a1) would expect the other behaviour, i.e. that U+00A0 should be extended and compressed like a ordinary space. This is why I don’t use U+00A0 in Writer at all, it is just unusable in justified paragraphs for me.
But the Czech sample shows that this IS a real problem, of course, and now I know I'm not the only user who is not satisfied with the state of affairs.
So maybe we should discuss if the traditional behaviour should be changed, probably with a new compatibility setting to keep the traditional behaviour for old documents ...?!
*** Bug 49674 has been marked as a duplicate of this bug. ***
I think this should be classified as an enhancement rather than a bug. The current behaviour is in fact ancient word processing practice, predating Unicode standards. U+00A0 became the successor of the old "hard space" defined for use with ASCII codesets, and changing the treatment of U+00A0 would break countless documents which purposely use hard spaces as _fixed-width_ non-breakable spaces (with abbreviations like "Dr Freud", "i. e." or punctuation like "« Bonjour! »", "5 %" etc.). It would also be not compatible with current MS Word practise.
However, distinguishing between different forms of white space is a typographical need and should be addressed somehow. DTP software like InDesign has all sorts of spaces: em space, en space, nonbreaking space, nonbreaking space fixed width, third/quarter/sixth/hair/thin space (1/3, 1/4, 1/6, 1/24, 1/8 em space), figure space, punctuation space. LibreOffice has "space" and "hard space" (and of course Unicode spaces like U+202F and U+2009, which it handles better than MS Word).
Jan_J (bug 49674 comment 2) proposed to use the Unicode word joiner U+2060 with a normal space to get a non-fixed-width non-breakable space. But U+2060 is a zero width non-breaking space inhibiting line breaks at both sides which is "intended for disambiguation of functions for byte order mark" (Unicode 6.2). That does not sound like a good candidate for such a space (and one would need the triple U+2060 + U+0020 + U+2060, wouldn't one?).
Users definitely also need non-breakable _fixed-width_ spaces, and if LO redefined U+00A0 as of non-fixed-width (in accordance with Unicode) - what character should be used for the classical "hard space"? MS Word displays "box characters" for symbols not defined in the active font, which should be kept in mind. (I know it cannot handle U+2009, but I haven't tested U+202F.)
A practical solution would probably be to let the user decide on a per document-basis how to interpret U+00A0: fixed width or proportional? That is, to add a configuration option under "Writer/Compatibility". But even then one should _still_ be able to use all necessary kinds of spaces at least in ODT; they may need to be converted for DOC/DOX export, however, because of MS Word limitations.
Thank you for your careful description of the situation and of possible options!
You are right, this issue is better regarded as an enhancement request; therefore I change the Importance field accordingly.
It is easy to agree with Stfhell's notion that the intervening space in expressions such as "Dr Freud" and "5 %" should be non-breaking, but I can't quite see the reasoning behind it having to be of fixed width too. By similar logic, shouldn't the spaces in "Sigmund Freud" and "five per cent" have fixed width as well? I find it rather inconsistent that a non-breaking space, which in non-justified text looks exactly like an average space, may stand out as narrower than average if the text is justified. Can you point out an authoritative source that actually recommends this? (Note that even in justified text, the difference will only be discernible on some of the lines, and in carefully typeset publications it should ideally not be discernible at all because the variation between lines is minimized by using hyphenation.)
The French spacing applied in connection with certain punctuation is a little different matter, as U+00A0 is mostly considered too wide for this purpose in professional-level typography as far as I know. A more appropriate character should be the narrow no-break space U+202F (though technical support for it may still be lacking in some environments; for a detailed, though not necessarily quite up-to-date discussion, see http://stackoverflow.com/questions/595365/how-to-render-narrow-non-breaking-spaces-in-html-for-windows).
As regards abbreviations such as "i.e." the standard way to write these seems to be without any space, at least as far as English is concerned:
So, in principle what my point boils down to is this: Is there actually a legitimate need for a fixed-width no-break space that is _only_randomly_ distinguishable from a normal space in justified text? Sure, many people have learned to expect that U+00A0 behaves like that, but from a professional typographer's perspective this expectation may be misguided, and it is clearly contradicted by the Unicode standard. (It may also be worth noting that Firefox nowadays seems compliant with the Unicode in its rendering of U+00A0.)
That said, the approach suggested by Stfhell might indeed offer a practical compromise, catering both for the Unicode-compliant view and the MS Word-compliant view.
Oh, the two dictionary links included in my previous comment should incorporate the final period, which apparently has been interpreted as sentence-ending punctuation by the Bugzilla system.
(In reply to comment #7)
> It is easy to agree with Stfhell's notion that the intervening space in
> expressions such as "Dr Freud" and "5 %" should be non-breaking, but I can't
> quite see the reasoning behind it having to be of fixed width too. By
> similar logic, shouldn't the spaces in "Sigmund Freud" and "five per cent"
> have fixed width as well? I find it rather inconsistent that a non-breaking
> space, which in non-justified text looks exactly like an average space, may
> stand out as narrower than average if the text is justified. Can you point
> out an authoritative source that actually recommends this?
Typesetting conventions are conventions, not ISO standards, and they vary with language and time and personal taste. I can direct you to the orthographic German "Duden" (following DIN 5008 for letter-writing): With office documents and e-mails use a space after abbreviation dots (z. B., u. a. m.), but not in dates (05.07.06); in word processing use a small fixed-width space in both abbreviations and dates. (What merriam-webster.com and oxforddictionaries.com do is compatible with _English_ typesetting practise and with common writer's practise, because it's the easiest way to prohibit a line break.)
Spaces before/after/around symbols like $ % & / « » vary a lot, but in typesetting handbooks you usually find recommendations like 1/6 or 1/8 or 0 em quad. A full and proportional space would be regarded as unprofessional typesetting in Germany. In typesetting systems, users have fixed-width spaces of all sizes (including the normal inter-word size of about 1/4 quad) for all kinds of usages (space between chapter number and title; aligning numbers like "347" and "_47" vertically; insert a space at paragraph end to avoid the last line being fully justified). They are "tools" for laying out text, not necessarily a way to encode text as information - typesetters use such things as double 1/4 quad spaces.
So fixed-width variants of normal space size do have a use (and Unicode defines them: U+2002, U+2004, U+2005 etc.). The important point is not that the fixed-width space should be distinguishable in all cases, but that it should not be extensible with proportional spacing. In good typography such spaces should in most cases be smaller than the regular space (as you say).
And, of course, you are right in that U+00A0 is _not_ defined as fixed-width. And Microsoft knows that:
But designing fonts and designing word processors are different things for Microsoft. Offering Word users a submenu with various types of spaces would be overkill for most users, and Microsoft has decided to offer them the fixed-width normal space as a single "compromise" alternative. Whether from the need to be downward-compatible with pre-Unicode documents, from misinterpretation of the Unicode standards or from conscious design principles. (Word processors are in fact used as modern typewriters, people don't want to fiddle with half a dozen spaces, and many don't even bother with hard spaces.)
In a world where only recent versions of Firefox render U+00A0 correctly, where Adobe epub-reader software cannot render a soft hyphen correctly and the most commonly used word processor renders all spaces apart from U+0020 and U+00A0 as boxes if the font doesn't define them (LibreOffice uses the glyphs from a substitution font), you cannot just follow Unicode standards blindly without regard to compatibility issues.
But of course there is other software than MS Word. InDesign imports Unicode spaces well from DOC files, and LibreOffice shouldn't let itself be limited by a word processor with modest formatting capabilities. (In InDesign, imported U+00A0 are rendered correctly. Thin spaces are fixed-width, as far as I know, in line with common typesetting practise.) But it should be a conscious decision of the user to depart from Word conventions on a per-document basis. The problem is: What space could be used for fixed-width spaces (for which there is also a definite need) if you tick that future LO box "Treat hard space as proportional"?
(In reply to comment #9)
> I can direct you to the orthographic German "Duden" (following DIN 5008
> for letter-writing): With office documents and e-mails use a space after
> abbreviation dots (z. B., u. a. m.), but not in dates (05.07.06); in word
> processing use a small fixed-width space in both abbreviations and dates.
Thank you for the reference. I got hold of a copy of the 2009 edition of "Duden: die deutsche Rechtschreibung", which defines hard spaces ("Festabstände") as fixed-width, mostly smaller ("meist kleinere") spaces that prohibit line breaks (the same definition is already included in the 2000 edition and available online: http://www.egb-buende.de/tools/EDV_Fuehrerschein_NRW/03_Grundlagen_Textverarbeitung/textverarbeitung_duden1.pdf).
So, this definition primarily seems to concern non-breakable _thin_ spaces, though the modifier "meist" leaves room for some interpretation. Furthermore, in the context of specific examples (e.g., of the use of the percent sign) it is repeatedly said that a "smaller space" should be used that is explicitly described as both hard and protected ("geschützter"; the 2000 edition isn't quite as explicit on these points). On the other hand, according to the 2009 edition, the official standard DIN 5008 speaks of a full ("ganzer") space, which apparently needs be neither fixed-width nor non-breakable. My reading of all this is that almost any space will do, but a non-breakable thin space is preferred. And this was basically what I was talking about: normally you'd want either a variable-width space or a _thin_ fixed-width space, not a fixed-width space that sometimes looks like a normal space and sometimes not.
> They are "tools" for laying out text, not necessarily a way to encode text
> as information - typesetters use such things as double 1/4 quad spaces.
> So fixed-width variants of normal space size do have a use (and Unicode
> defines them: U+2002, U+2004, U+2005 etc.). The important point is not that
> the fixed-width space should be distinguishable in all cases, but that it
> should not be extensible with proportional spacing. In good typography such
> spaces should in most cases be smaller than the regular space (as you say).
Now I'm a little confused. Are you talking of the regular no-break space (U+00A0) or the _narrow_ no-break space (U+202F) here?
What I said was basically that in the ideal case there should hardly be any distinguishable difference between U+00A0 and a normal space, even if U+00A0 was treated as a fixed-width space. If a fixed-width space is not distinguishable from a normal space, it does not matter in practice whether it is of fixed width or not. A different matter is that often U+00A0 is just used as the poor man's narrow no-break space, relying on it being treated as a fixed-width space in justified text. I can see the reasoning, but this usage is not in alignment with the best practices of traditional typography as far as I can see.
Granted, Unicode defines a set of fixed-width spaces, the majority of which are, as formulated on the Microsoft page you referred to, characters corresponding to traditional typographic _space_values_ that have indeed been applied in manual typesetting. Historically, for each space between words on a line, an identical space value (typically corresponding to U+2004 or U+2005) would have been applied. For each space on other lines, a slightly different value was applicable when necessary to get all the lines justified. After punctuation, a larger-than-average value would often have been preferred, or in some special cases, a thin space.
For more details, see paragraphs 239–254 explaining technical terms in the 1st edition of the Chicago Manual of Style (published in 1906):
This is the historical background to the Unicode fixed-width spaces, and one might want to argue that many of these characters are of little practical use in the age of digital typography. Notably, most of the Unicode fixed-width spaces are _breakable_ and have no non-breakable counterparts (breakability was not a concern in manual typesetting, as each line was typeset as a separate unit). This can be seen as a deficiency in the Unicode character repertoire, or alternatively perhaps as an implicit stand that the kind of fine adjustment they were originally intended for should rather rely on different means in modern typesetting systems. Using the Unicode fixed-width spaces for manual justifying in digital typesetting would be awkward and anachronistic.
Be that as it may, there is not much LibreOffice can do to change the overall situation. If a document is first edited in LibreOffice and then opened in another application (possibly after being exported into a specific format), U+00A0 will be rendered either as a variable-width space (in Firefox) or as a fixed-width space (in MS Word). LibreOffice only has control over its own rendering of the character (and how it will be printed in some non-editable formats, such as PDF). Additionally, LibreOffice might want offer an easy short cut to entering U+202F in order to cater for finer typography, but again, there can be no guarantee that it will be rendered correctly in other applications.
(re Comment #10) I don't think it makes much sense to discuss the merits of various spaces or typography issues, Simo, especially on LO Bugzilla. The characters exist, people can use them, and LO should handle them as well as possible. The details are often just a matter of taste, or the willingness to distinguish among a dozen kinds of space characters...
I think Roman's proposal (Comment #3) to let the user configure in Options/Writer/Compatibility how the classical "hard space" (encoded as U+00A0) should be handled (fixed-width as in Word or proportional as Unicode says) is a very practical solution. The Compatibility menu gives users the choice to set an option just for the current file or use it as a default.
If you decide to go with Unicode standards and configure a proportional U+00A0, you can use the characters that Unicode has defined as fixed-width spaces: U+2000 to U+200A, U+202F, U+205F.
The problem here is interoperability with MS Word, because Word, as said, displays all characters not defined in the font as "box characters". But probably this is becoming less of a problem. I had a look at some fonts: For Windows 7, Microsoft supplies fonts that actually define the various Unicode spaces. So it would be users of older Windows versions or users of the many fonts with a more restricted character set that would see the "box character". You can, as a workaround, format all fixed-width spaces in Verdana, Times New Roman or some other Unicode font to avoid that (somewhat, at least).
You can use fileformat.info to check some font glyph sets:
(for U+202F), or you can use the character map application of the OS.
I have no idea how comprehensive the fonts that come with MacOS are, or if Word for Mac has the same "box character" issue as Word for Win. Would be good to know...
(In reply to comment #10)
> Using the Unicode
> fixed-width spaces for manual justifying in digital typesetting would be
> awkward and anachronistic.
I consider a good support for Unicode spaces as something essential for a word processor with advanced layout capabilities. Software like Word or LibreOffice and even InDesign is in many respects "anachronistic" in your sense of the word (typewriter-like or lead-typesetting-like), there is no other way to define the spacing you want but in the form of glyphs. With XSL transformations of XML documents (or TeX) you can have a stylesheet (instead of the document) define the spaces in a template, thus achieving a uniform handling of for example thin spaces around « Bonjour! » - but even then you need to define the spaces as Unicode characters in the template. You just needn't encode them in your document. With word processors (and DTP software), you have to set all the spaces yourself in the document text. It is somewhat anachronistic and very error-prone typesetting - but a straightforward and simple concept for users.
(In reply to comment #12)
We all more or less seem to agree that the rendering of U+00A0 as a fixed-width space is basically a bug. Therefore it is a somewhat perverted situation that the bug cannot simply be fixed without paying attention to how the incorrect behaviour can also be preserved. Is it actually worth the trouble? I am not saying that it is a complete waste of time and effort, but this is a question that deserves to be asked too.
It was not me who brought up that there are a variety of fixed-width spaces in Unicode. Nevertheless, as we are discussing whether or not U+00A0 should continue to be rendered as a fixed-width space, at least optionally, we should try to understand the background to and reasoning behind these standardized fixed-width spaces (and why U+00A0 is not one of them). Sure, some of them are still relevant today, but unquestionably some are redundant (U+2000 and U+2001 are canonically equivalent to U+2002 and U+2003 respectively). And then there are some the relevance of which can be questioned as far as modern typesetting practices are concerned.
However, nobody has suggested that the redundant or possibly archaic Unicode characters need not be handled correctly. That is not the issue here. There are many redundant, archaic and even deprecated characters in Unicode, for which the main motivation is historical. People who want to use these characters for whatever reason in their documents should certainly have the option to do so, even if it may not always be the most elegant or technically reliable choice in digital typesetting.
> > Using the Unicode
> > fixed-width spaces for manual justifying in digital typesetting would be
> > awkward and anachronistic.
> there is no
> other way to define the spacing you want but in the form of glyphs.
I'm afraid you may have missed my point, so I'll try to clarify. The example was about how text used to be _justified_ manually. Historically, to achieve this effect you would have applied specific space values between words on each line. This was in fact one of the main uses for the various space values in manual typesetting. Today, however, if you want spacing on each line to be even, you simply specify that the text should be justified and let your application software automatically adjust the width of spaces accordingly. There is no point in using various fixed-width spaces for this purpose anymore.
Historically, the first line of a paragraph would have been indented about one em space. Today, rather than inserting U+2003 (or U+2001), you can specify a fixed indentation value that will automatically be applied at the beginning of each paragraph. There is no need for a specific glyph there anymore.
Historically, certain punctuation, such as a sentence-ending period, would have been followed by an em space or a couple of three-per-em spaces. Today this is often regarded old-fashioned, but people who still want to follow the tradition simply tend to type two (or even three?) regular spaces after the period. Sure, a purist could insert a U+2003 or a couple of U+2004s instead, but I fail to see how this would make any significant improvement from a typographical point of view.
Yes, you can continue to use all the standard fixed-width spaces if you want to, but this is what you choose to do and it does not make the _software_ anachronistic. When using a word processor, my father, who is in his late seventies, still tends to break lines manually by tapping the return key at the end of each line, and he may also hyphenate the last word on a line by inserting a regular hyphen (U+002D) before the line break. This is because his paradigm of typing is of a different era. He does not take full advantage of the modern technology (in fact he still prefers a mechanical typewriter occasionally) – and that's fine, since he is retired and mostly writes for his own pleasure every now and then. But for the rest of us it is good to know that today there are more elegant methods of typing a piece of text.
As regards French spacing in « Bonjour ! », inserting a non-breakable thin space before or after the punctuation marks may be a practical solution at the present time, but there are in fact alternative methods for this too. Smart font technology, as exemplified by "Linux Libertine G" and "Linux Biolinum G" fonts (already bundled with the recent versions of LibreOffice, even on Windows), allows automatic application of French spacing where deemed appropriate. There is no need to insert a specific space character, as LibreOffice is able to recognize the guillemets, exclamation marks, question marks etc. and take care of proper spacing by making use of additional instructions incorporated in the font itself. Unfortunately the technology is far from being universally supported (and there are several alternative ways to incorporate similar features), but with the right combination of font technology and application software it is quite functional.
For more information on this technology, see:
This kind of smart font technology could also offer an elegant way to make available the optional fixed-width U+00A0, were this feature considered important enough to be incorporated in a font. The feature would then be available in any application software able to support the font technology, and being an optional feature of the font it could be applied to any portion of a text, rather than to the document as a whole. On the downside, it would only be available with some specific fonts. This might also be criticized as an excessively sophisticated approach to implement a feature that is basically non-standard.
(In reply to comment #13)
> We all more or less seem to agree that the rendering of U+00A0 as a
> fixed-width space is basically a bug.
Yes. But a special kind of bug: a bug which has been sanctioned by the fact that it has been in Microsoft’s applications since ages, and therefore is considered as some kind of “industry standard” by many people :-(
> Therefore it is a somewhat perverted
> situation that the bug cannot simply be fixed without paying attention to
> how the incorrect behaviour can also be preserved. Is it actually worth the
IMHO we can not just fix this bug by rendering U+00A0 as a proportional space, because many people, probably: most (!) people will consider the new, proportional rendering as a bug and cry: “You don’t render my .doc files correctly anymore, fix this!” ...
Therefore the simplest solution which improves typography in LibreOffice without breaking the (wrong) assumptions of many people is (as already suggested in comment #3) to let the user configure in Options/Writer/Compatibility how the classical “hard space” (U+00A0) should be rendered: fixed-width as in Word or proportional (as Unicode and many (most?!) textbooks about typography suggest). This option should work on a per-document base, of course. So all existing documents will look like before when we open it, but can be changed to the new, better rendering by just switching that option; and the same option will also control the behaviour of documents created anew.
Adding that option should not be too difficult; I remember that similar additional compatibility options have been added in the past with relatively few lines of code ...
Sorry for just repeating my initial suggestion, but IMHO all the other, more advanced stuff -- e.g. bigger spaces after a sentence-ending period -- are different items, for which we should file special enhancement requests, if necessary. The present bug report is already far too long and complicated --
no developer who is just looking for some work to be done will understand easily the many things we have been already discussing in this single bug report ;-)
(In reply to comment #5)
> Jan_J (bug 49674 comment 2) proposed to use the Unicode word joiner U+2060
> with a normal space to get a non-fixed-width non-breakable space. But U+2060
> is a zero width non-breaking space inhibiting line breaks at both sides
> which is "intended for disambiguation of functions for byte order mark"
> (Unicode 6.2). That does not sound like a good candidate for such a space
> (and one would need the triple U+2060 + U+0020 + U+2060, wouldn't one?).
For Jan_J’s request (and related problems of the line-breaking algorithm), there is now bug 57652 - “Wrong treatment of Word Joiner (U+2060) in line breaking algorithm”.
However, I think that one solution does not necessarily invalidate the other. I.e., when the bug 57652 would get fixed, and WJ + SP (+ WJ) would act like an elastic NBSP, we could still discuss if the behaviour of U+00A0 should be changed, too (or better: if an option to choose the behaviour of U+00A0, elastic or fixed, should be added). So the present bug report is still a valid enhancement request and essentially independend from bug 57652.
** Please read this message in its entirety before responding **
To make sure we're focusing on the bugs that affect our users today, LibreOffice QA is asking bug reporters and confirmers to retest open, confirmed bugs which have not been touched for over a year.
There have been thousands of bug fixes and commits since anyone checked on this bug report. During that time, it's possible that the bug has been fixed, or the details of the problem have changed. We'd really appreciate your help in getting confirmation that the bug is still present.
If you have time, please do the following:
Test to see if the bug is still present on a currently supported version of LibreOffice (4.4.1 or later): https://www.libreoffice.org/download/
If the bug is present, please leave a comment that includes the version of LibreOffice and your operating system, and any changes you see in the bug behavior
If the bug is NOT present, please set the bug's Status field to RESOLVED-WORKSFORME and leave a short comment that includes your version of LibreOffice and Operating System
Please DO NOT
Update the version field
Reply via email (please reply directly on the bug tracker)
Set the bug's Status field to RESOLVED - FIXED (this status has a particular meaning that is not appropriate in this case)
If you want to do more to help you can test to see if your issue is a REGRESSION. To do so: 1. Download and install oldest version of LibreOffice (usually 3.3 unless your bug pertains to a feature added after 3.3)
2. Test your bug 3. Leave a comment with your results. 4a. If the bug was present with 3.3 - set version to "inherited from OOo"; 4b. If the bug was not present in 3.3 - add "regression" to keyword
Feel free to come ask questions or to say hello in our QA chat: http://webchat.freenode.net/?channels=libreoffice-qa
Thank you for your help!
-- The LibreOffice QA Team This NEW Message was generated on: 2015-07-18
The bug is still present in LibreOffice 184.108.40.206, on Ubuntu 14.04 LTS, its behavior did not undergo any changes.
Question: is this occurring in .ODT *and* .DOCX files?
And chance of uploading a test document?
Created attachment 122957 [details]
non breaking space interpreted as fixed width space
Yes, this happens for all formats: ODT, DOCX and DOC.
A sample document (with text of the attached PNG) attached.
(In reply to dohnp5a1 from comment #0)
> Created attachment 52184 [details]
> The preposition “s” may not stand in the end of line in Czech, thus no-break
> space is used after that
It's similar in polish editing style best practices, the correct behaviour would really set apart LO positively.
Anecdote from a 100-year old legal book in Australia (our "Annotated Constitution"):
- nbsp between name components such as "Mr. Smith" contain a narrow non-breaking space
- whereas nbsp between place names such as "Port Agusta" are normal or proportional-width spaces (most of the book is justified)
I.e. to properly duplicate this book requires both types of spaces.
Web browsers use U+00A0 as non-breaking not-fixed-width space too.
Also MS Office changed the behavior:
With the introduction of Word 2013, MS changed the behaviour of the ASCII 160 non-breaking space. It now conforms to the CSS space rules. This allows the space to expand/contract with justification so that all spaces on a line have the same width; the ASCII 160 behaviour could look odd with its fixed-width non-breaking spaces in such cases. For fixed-width non-breaking spaces you can use one of the other non-breaking space characters (eg Narrow No-Break Space: 202F,Alt-x).
This is biting me as well. It makes Libreoffice Writer basically unusable for any serious Czech language text.
This concerns Slovak texts as well. It prevents me using LibreOffice for book publishing.
(In reply to rysson from comment #22)
> Web browsers use U+00A0 as non-breaking not-fixed-width space too.
> Also MS Office changed the behavior:
> With the introduction of Word 2013, MS changed the behaviour of the ASCII
> 160 non-breaking space. It now conforms to the CSS space rules. This allows
> the space to expand/contract with justification so that all spaces on a line
> have the same width; the ASCII 160 behaviour could look odd with its
> fixed-width non-breaking spaces in such cases. For fixed-width non-breaking
> spaces you can use one of the other non-breaking space characters (eg Narrow
> No-Break Space: 202F,Alt-x).
Well, they seem to change it again, because in Word 2016 the non‐breaking space have a fixed‐width property again. It stopped commiting to modern standards for whatever reasson (probably the whining of long‐term users). Source:
> It stopped commiting to modern standards for whatever reasson (probably the whining of long‐term users).
Yeah. But it's not good reason to stop fixing it in LibreOffice.
At least the option can be added (fixed / non-fixed)..
Microsoft does what Microsoft does. I can tell you as native Czech language speaker and writer that the fixed size of the /U+0160 characters is contrary to Czech typography. Currently, both Word in its latest version and LibreOffice are unusable for any serious documents do to this issue. *Please* make this at least optional per document or per paragraph or such.
I managed to work around it! Now I can do book-publishing in LibreOffice in the Slovak language. I use the U+2060 (WORD JOINER [WJ]) character for a non-breakable relative-width space. Does it work for you guys as well?
If I have understood the discussion well, the compatibility with standards and other software is a problem.
But if we limit the solution to two cases:
- exporting to PDF
(which is entirely enough for me)
then the compatibility problem disappears, as LO can develop a proprietary solution and encode this character in some way for the export only.
The Unicode standard document http://unicode.org/reports/tr14/ clearly states that:
When expanding or compressing interword space according to common typographical practice, only the spaces marked by U+0020 SPACE and U+00A0 NO-BREAK SPACE are subject to compression, and only spaces marked by U+0020 SPACE, U+00A0 NO-BREAK SPACE, and occasionally spaces marked by U+2009 THIN SPACE are subject to expansion. All other space characters normally have fixed width.</quote>
Whether LibreOffice or Word, both should comply to the above standard and expand both U+0020 and U+00A0 equally. LibreOffice should not blindly mimic what (changing) behaviour Word exhibits on this score.
Please fix this! This is a real embarrassment vis a vis good typography practices.
Whoever wants fixed width spaces please use one of the remaining space characters!
Yes, yes, yes!
This bug was reported in 2010. Please, I think it's time to fix it. I know I still can use LaTeX or pure HTML to write documents. But using LibreOffice for text documents will be very nice IMO.
Now I can break typografy (in Polish, and many another languages) or see awful text like:
This is justified text with a0 space – terrible.
This is justified text with a0 space – terrible?
BTW. Happy New Year!
Unicode has that "non-breaking" property only set to some of its different spaces. But in reality, different typographic rules of countries/bodies/times, as noted, may require full repertoire of spaces of different width and properties (fixed/widening/shortening) to be breaking *and* non-breaking variants. ODF could introduce a special internal "non-breaking" character property applicable to any space character, which would override normal Unicode algorithm; that would allow for adding shortcuts for such combos. Without that, any "fix" like the one asked here by many would only fix things for vocal minority, and break things for most users who naturally don't participate in discussion here, because - well, it just works for them, and they don't look for this bug ;-)
Note that what I propose would require an ODF extension.
At first set the default behaviour of LO strictly according the Unicode definitions – it is about time after 8 years (!). As a supplement later there could be some extension with more features added.
(In reply to dohnp5a1 from comment #33)
> At first introduce incompatibility by breaking unknown number of millions of existing documents of our users, to allow *some other* users to do what they need; and only then start thinking about doing it properly
No, it doesn't work that way. While Unicode is an important standard, it's only of secondary importance to an office suite. Its primary goal is *not* creating a reference comformant implementation of the standard; rather, it should use the standard to the extent it needs to serve its users most. And if legacy requires that some statements of standard be violated to keep existing documents intact, that should be that way, until a better design is invented and implemented, which would make possible to please both sides.
I would be happy with a per-document option for how to treat NBSP-s, but the default behaviour for new documents should be configured where?