# [Q] Math: handling of Greek symbols in OOo 3.3 ?

19 messages
Open this post in threaded view
|

## [Q] Math: handling of Greek symbols in OOo 3.3 ?

 Hi, I'd like to ask about a decision that needs to be made for upcoming OOo 3.3 which features the new and fixed OpenSymbol font (see issue 105084).   http://www.openoffice.org/issues/show_bug.cgi?id=105084Previously (up to, and including OOo 3.2) a formula like:     %MU %PI "  " nitalic {%MU %PI} "  " italic {%MU %PI} newline     %eta %kappa "  " nitalic {%eta %kappa} "  " italic {%eta %kappa} would have been rendered as in 'mathOSF_OOo3.2.png'. Please notice that you can't get a straight %PI it is slanted or even more slanted. :-( This however did not apply to %MU since the first comes from the StarSymbol/OpenSymbol font and the latter actually from TimesNewRoman since the old OpenSymbol font did only feature those non-Latin-looking Greek characters. That of course was a bug of the old font. This problem (and others) are now fixed in the new OpenSymbol font that will come with OOo 3.0. Currently(!) with a developers build like DEV300_m84 the above formula will be rendered as in 'mathOSF_OOo3.3.png'. You can easily see that 'italic' and 'nitalic' commands now have the proper effect on the characters. Also the non-altered charatcer (see first column) are defined as slanted by default right now since that is visually the best equivalent to the old look. However that is what I'am unsure about... The question is: should the Greek lower- and upper case characters be straight or slanted by default? In every formula at least one of the sample cases     '%PI'     'nitalic PI' or     'italic PI' will look different compared to OOo 3.2 if rendered anew in OOo 3.3! (And it can't be fixed on import.) Also changing the default for Greek symbols to straight will not solve this since then only one of the other cases will turn out different. If we were to create the Greek symnbols from scratch without any regard to how it looked like in older versions we would simply take a look at about any Math school book or at LaTex. In that case our solution would be: Solution 1) - Greek uppercase characters should be straight. They are usually used for operators or as set names, and those are usually layouted straigth. - Greek lowercase characters should be slanted. They normally get used as variables and those have a slanted layout. Other possible solutions are: Solution 2) - All Greek characters should be slanted by default since that is the best visual equivalent to OOo 3.2 (at least if no modifiers like 'italic' and 'nitalic' are used). Solution 3) - All Greek characters should be straight since that is what the unmodified charcters in the font look like. Solution 4) We'll choose one of the above but still like to 'work-around' the probably rather often need to explicitly write one of the modifiers before the character. Thus we are going to introduce new 'shortcuts' for those. It could be one or both of the following   4a) For each Greek character introduce another one (prefixed with i):       E.g for %PI add %iPI, where %iPI refers to an italic uppercase PI   4b) For each Greek character introduce another one:       E.g for %PI add %nPI, where %nPI refers to a non-italic straight       version of the character. Thus the choices to make are   - one of 1), 2) and 3) and   - 4a) and/or 4b) or none of those Please tell us the favored solution. Regards, Thomas --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 Here are the links for the two images that got removed: http://www.mediafire.com/?z50tkdzovymhttp://www.mediafire.com/?myklqznynikThomas --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [Q] Math: handling of Greek symbols in OOo 3.3 ?

 In reply to this post by Thomas Lange - Sun Germany - ham02 - Hamburg I'd like to add a further option 5: 5a)    If one of 1 or 3 is chosen, a macro should be created which would "translate" any Greek letter to be prefixed with the italic modifier. This so the user can quickly change the appearance if (and only if) they do want the old 3.2 look. If option 4a is chosen, it could be changed to this new 'i##' code instead of the modifier. 5b)    If option 2 is chosen, the reverse of above should then be implemented. So the user can "quickly" change to 'nitalic' if they don't want the old look, but rather the more standardised format. This should have 2 versions: One where both upper & lower case is "straightened", the other where onlu upper case is "straightened". 5c)    Introduce both 5a & 5b irrespective of 1, 2 & 3. -- Regards Irné Barnard --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [Q] Math: handling of Greek symbols in OOo 3.3 ?

Open this post in threaded view
|

## Re: [Q] Math: handling of Greek symbols in OOo 3.3 ?

 In reply to this post by Thomas Lange - Sun Germany - ham02 - Hamburg Thomas, I vote for Solution 3. As a researcher, I often use several greek characters in the text. While the LaTeX way makes sense for formulae, slanted characters in the middle of the text look awkward and out of place. Leaving the characters in the upright position by default and allowing the user to italicize when required sounds more logical to me. Having said that, perhaps, the defaults in formula editor can be changed to reflect the LaTeX way. Vamsi. > Solution 3) > - All Greek characters should be straight since that is what the > unmodified charcters in the font look like. --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 In reply to this post by Thomas Lange - Sun Germany - ham02 - Hamburg Thomas Lange - Sun Germany - ham02 - Hamburg schrieb: > Hi, > > I'd like to ask about a decision that needs to be made for upcoming OOo > 3.3 which features the new and fixed OpenSymbol font (see issue 105084). >   http://www.openoffice.org/issues/show_bug.cgi?id=105084I vote for option 3 (unslanted) and for adding "%i***" symbols for slanted characters. Daniel --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [Q] Math: handling of Greek symbols in OOo 3.3 ?

 In reply to this post by Thomas Lange - Sun Germany - ham02 - Hamburg Hi Thomas, Thomas Lange - Sun Germany - ham02 - Hamburg schrieb: > > Hi, > > I'd like to ask about a decision that needs to be made for upcoming OOo > 3.3 which features the new and fixed OpenSymbol font (see issue 105084). >    http://www.openoffice.org/issues/show_bug.cgi?id=105084[..] > Solution 3) > - All Greek characters should be straight since that is what the > unmodified charcters in the font look like. I vote for solution 3. For easier writing I like the idea of additional definitions %ixxx for the italic variants. kind regards Regina --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [Q] Math: handling of Greek symbols in OOo 3.3 ?

 In reply to this post by Irné Barnard Hi Irné, > I'd like to add a further option 5: > 5a)    If one of 1 or 3 is chosen, a macro should be created which would > "translate" any Greek letter to be prefixed with the italic modifier. > This so the user can quickly change the appearance if (and only if) they > do want the old 3.2 look. If option 4a is chosen, it could be changed to > this new 'i##' code instead of the modifier. > > 5b)    If option 2 is chosen, the reverse of above should then be > implemented. So the user can "quickly" change to 'nitalic' if they don't > want the old look, but rather the more standardised format. This should > have 2 versions: One where both upper & lower case is "straightened", > the other where onlu upper case is "straightened". > > 5c)    Introduce both 5a & 5b irrespective of 1, 2 & 3. A simple macro won't do since to do it correctly that macro would be required to implement a parser for Math in order to find out if a specific char should be italic or not. This is because of the 'italic' and 'nitalic' modifiers that could stand about anywhere in the formula and still effect on a character. Thus a programmatic solution is required, and even a macro would at least require some new programmatic support in order to work. We will think about this a bit more... Thomas --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 In reply to this post by Thomas Lange - Sun Germany - ham02 - Hamburg Hi all, It seems that most people prefer the Greek symbols to be straight by default, and to add en extra symbols set where the symbol names get prefixed with 'i' and are slanted. I have submitted issue     http://www.openoffice.org/issues/show_bug.cgi?id=113276for this. About quickly switching between italic and non-italic default for the Greek symbols (a prerequisite if there should be a macro to change all formulas in a Writer document): For this I have submitted issue     http://www.openoffice.org/issues/show_bug.cgi?id=113277Unfortunately QA told me that even the simple and no UI change requiring first issue is unlikely to get accepted as a show stopper, and thus I have currently set the target for both issues to OOo 3.4. :-( But feel free to argue with product managers on the releases ML about this. ^_- Regards, Thomas --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 On 07/20/10 14:04, Thomas Lange - Oracle wrote: > It seems that most people prefer the Greek symbols to be straight by > default, and to add en extra symbols set where the symbol names get > prefixed with 'i' and are slanted. Just a thought:  Is it technically correct to refer to those letter faces as "italic" (and thus using "i" as prefix), or would "oblique" or "slanted" be better?  (My understanding is that italic faces have different glyph shapes than their -- typically roman -- counterparts, whereas oblique/slanted faces are the roman glyphs just, well, slanted.) -Stephan --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 On Jul 20, 2010, at 2:26 PM, Stephan Bergmann wrote: > On 07/20/10 14:04, Thomas Lange - Oracle wrote: >> It seems that most people prefer the Greek symbols to be straight by >> default, and to add en extra symbols set where the symbol names get >> prefixed with 'i' and are slanted. > > Just a thought:  Is it technically correct to refer to those letter   > faces as "italic" (and thus using "i" as prefix), or would "oblique"   > or "slanted" be better? That's a question of "implementation vs. interface": The interface is   to ask for italic and the most common implementation is to fallback to   oblique if italic is not available. > (My understanding is that italic faces have different glyph shapes   > than their -- typically roman -- counterparts, whereas oblique/ > slanted faces are the roman glyphs just, well, slanted.) Yes. -- Herbert --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 Hi, > On Jul 20, 2010, at 2:26 PM, Stephan Bergmann wrote: > > On 07/20/10 14:04, Thomas Lange - Oracle wrote: > >> It seems that most people prefer the Greek symbols to be straight by > >> default, and to add en extra symbols set where the symbol names get > >> prefixed with 'i' and are slanted. > > > > Just a thought:  Is it technically correct to refer to those letter   > > faces as "italic" (and thus using "i" as prefix), or would "oblique"   > > or "slanted" be better? > > That's a question of "implementation vs. interface": The interface is   > to ask for italic and the most common implementation is to fallback to   > oblique if italic is not available. > > > (My understanding is that italic faces have different glyph shapes   > > than their -- typically roman -- counterparts, whereas oblique/ > > slanted faces are the roman glyphs just, well, slanted.) > > Yes. >   I that is the detail level you like to talk about: Programmatically the only values from the FontItalic enum that Math makes use of are     ITALIC_NORMAL and ITALIC_NONE Thus the keyword 'italic' in a formula text corresponds to ITALIC_NORMAL and 'nitalic' to ITALIC_NONE. Everything else depends on vcl. Since (per default) *all* of the Greek symbols now from the OpenSymbol font, is there actually a difference between ITALIC_OBLIQUE and ITALIC_NORMAL when using OpenSymbol? Are you suggesting to use ITALIC_OBLIQUE instead of ITALIC_NORMAL? Thomas --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 Hi again, >>>> It seems that most people prefer the Greek symbols to be straight   >>>> by >>>> default, and to add en extra symbols set where the symbol names get >>>> prefixed with 'i' and are slanted. >>> >>> Just a thought:  Is it technically correct to refer to those letter >>> faces as "italic" (and thus using "i" as prefix), or would "oblique" >>> or "slanted" be better? >> >> That's a question of "implementation vs. interface": The interface is >> to ask for italic and the most common implementation is to fallback   >> to >> oblique if italic is not available. > [...] > > Since (per default) *all* of the Greek symbols now from the OpenSymbol > font, is there actually a difference between ITALIC_OBLIQUE and > ITALIC_NORMAL when using OpenSymbol? > Are you suggesting to use ITALIC_OBLIQUE instead of ITALIC_NORMAL? No. Asking for is ITALIC_NORMAL fine. If there is no OpenSymbol-Italic   it gets emulated with oblique. If too many people were unhappy with the emulated oblique style we   could create an OpenSymbol-Italic font face. If we did that we   shouldn't have to constrain the glyph shapes too much: If there was a   need for such a style of OpenSymbol the reason was that look of the   oblique version glyphs were not good enough in the first place. -- Herbert --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 Agreed. The Italic_Normal and Italic_Oblique seems to be a larger and more encompassing issue. Which we might look into at a later stage (if such complaint happens). For me I've never had a problem with using Oblique as is ... but that's just me. -- Regards Irné Barnard --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 In reply to this post by Herbert Duerr-2 On 07/20/10 15:03, Herbert Duerr wrote: > On Jul 20, 2010, at 2:26 PM, Stephan Bergmann wrote: >> On 07/20/10 14:04, Thomas Lange - Oracle wrote: >>> It seems that most people prefer the Greek symbols to be straight by >>> default, and to add en extra symbols set where the symbol names get >>> prefixed with 'i' and are slanted. >> >> Just a thought: Is it technically correct to refer to those letter >> faces as "italic" (and thus using "i" as prefix), or would "oblique" >> or "slanted" be better? > > That's a question of "implementation vs. interface": The interface is to > ask for italic and the most common implementation is to fallback to > oblique if italic is not available. And my question was whether you really "ask for italic."  Maybe my use of the term "technically correct" was unfortunate; so, rephrased, my question is:  In the jargon of professional typesetting, if you set the mathematical formula "sin α," do you refer to the alpha as "italic" or as "slanted" (or "oblique")? The Chicago Manual of Style appears to be no help answering that question, and The TeXbook says "It's conventional to use unslanted letters for uppercase Greek, and slanted letters for lowercase Greek [...]" but "There's also \mit, which stands for 'math italic.' This affects uppercase Greek [...]", so appears to be somewhat undecided. When reading the OP, the term "Greek italic" suddenly struck me as odd, a combination of words that, at least when taken at face value (no pun intended), do not fit together.  That got me thinking, and that's all. -Stephan --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 Hi Stephan, > On 07/20/10 15:03, Herbert Duerr wrote: >> On Jul 20, 2010, at 2:26 PM, Stephan Bergmann wrote: >>> On 07/20/10 14:04, Thomas Lange - Oracle wrote: >>>> It seems that most people prefer the Greek symbols to be straight by >>>> default, and to add en extra symbols set where the symbol names get >>>> prefixed with 'i' and are slanted. >>> >>> Just a thought: Is it technically correct to refer to those letter >>> faces as "italic" (and thus using "i" as prefix), or would "oblique" >>> or "slanted" be better? >> >> That's a question of "implementation vs. interface": The interface is to >> ask for italic and the most common implementation is to fallback to >> oblique if italic is not available. > > And my question was whether you really "ask for italic."  Maybe my use > of the term "technically correct" was unfortunate; so, rephrased, my > question is:  In the jargon of professional typesetting, if you set the > mathematical formula "sin α," do you refer to the alpha as "italic" or > as "slanted" (or "oblique")? > > The Chicago Manual of Style appears to be no help answering that > question, and The TeXbook says "It's conventional to use unslanted > letters for uppercase Greek, and slanted letters for lowercase Greek > [...]" but "There's also \mit, which stands for 'math italic.' This > affects uppercase Greek [...]", so appears to be somewhat undecided. Yup it is as you say. Unfortunately SO implemented it wrong from the start ages ago. However I already put the question to vote and the majority wanted all the Greek letter in straight form by default AND to additionally have the slanted characters available by prefixing the 'regular' ones with 'i'. I also believe that this approach should allow for the best ease of use experience. Thus %i... will be our version of \mit. ^_- > When reading the OP, the term "Greek italic" suddenly struck me as odd, > a combination of words that, at least when taken at face value (no pun > intended), do not fit together.  That got me thinking, and that's all. Thomas BTW: good news, issue 113276 got accepted as show-stopper. --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]
Open this post in threaded view
|

## Re: [users] [Q] Math: handling of Greek symbols in OOo 3.3 ?

 On 07/21/10 10:37, Thomas Lange - Oracle wrote: >> On 07/20/10 15:03, Herbert Duerr wrote: >>> On Jul 20, 2010, at 2:26 PM, Stephan Bergmann wrote: >>>> On 07/20/10 14:04, Thomas Lange - Oracle wrote: >>>>> It seems that most people prefer the Greek symbols to be straight by >>>>> default, and to add en extra symbols set where the symbol names get >>>>> prefixed with 'i' and are slanted. >>>> >>>> Just a thought: Is it technically correct to refer to those letter >>>> faces as "italic" (and thus using "i" as prefix), or would "oblique" >>>> or "slanted" be better? >>> >>> That's a question of "implementation vs. interface": The interface is to >>> ask for italic and the most common implementation is to fallback to >>> oblique if italic is not available. >> >> And my question was whether you really "ask for italic."  Maybe my use >> of the term "technically correct" was unfortunate; so, rephrased, my >> question is:  In the jargon of professional typesetting, if you set the >> mathematical formula "sin α," do you refer to the alpha as "italic" or >> as "slanted" (or "oblique")? >> >> The Chicago Manual of Style appears to be no help answering that >> question, and The TeXbook says "It's conventional to use unslanted >> letters for uppercase Greek, and slanted letters for lowercase Greek >> [...]" but "There's also \mit, which stands for 'math italic.' This >> affects uppercase Greek [...]", so appears to be somewhat undecided. > > > Yup it is as you say. > Unfortunately SO implemented it wrong from the start ages ago. > > However I already put the question to vote and the majority wanted all > the Greek letter in straight form by default AND to additionally have > the slanted characters available by prefixing the 'regular' ones with 'i'. > I also believe that this approach should allow for the best ease of use > experience. Thus %i... will be our version of \mit. ^_- Thomas, you still don't get me, I'm afraid.  I did not want to discuss whether or not Greek letters should be slanted when used in formulas.  I only wanted to discuss whether it is correct to call such slanted Greek letters "italic" (and thus use that "i" prefix). -Stephan --------------------------------------------------------------------- To unsubscribe, e-mail: [hidden email] For additional commands, e-mail: [hidden email]