The Unicode Consortium Discussion Forum

The Unicode Consortium Discussion Forum

 Forum Home  Unicode Home Page Code Charts Technical Reports FAQ Pages 
 
It is currently Tue Oct 21, 2014 5:22 am

All times are UTC - 6 hours [ DST ]





Post new topic Reply to topic  [ 6 posts ] 
Author Message
 Post subject: The HO property is not needed and should be removed
PostPosted: Mon Jul 02, 2012 7:54 pm 
Offline

Joined: Thu Mar 08, 2012 12:06 am
Posts: 3
Unlike the MVO and SVO properties which define default character orientation in vertical text runs, looking at the data for HO the utility of defining a the HO property is lost on me. This seems to be trying to codify something very simple, namely that Mongolian and Phags-pa are vertical only scripts and the code charts reflect the vertical version of the glyphs for each character rather than the horizontal representation used for all other codepoints.

Surely it doesn't make sense to add a data property which consists entirely of the condition "if the block is Mongolian or Phags-pa then L, else U".

If there is a typographic tradition of showing these glyphs rotated left in horizontal runs, then that can be noted. That will be sufficient as guidance to applications handling these scripts in horizontal runs. A new property is not needed nor particularly useful.


Top
 Profile  
 
 Post subject: Re: The HO property is not needed and should be removed
PostPosted: Tue Jul 03, 2012 5:53 pm 
Offline

Joined: Mon Feb 01, 2010 6:18 pm
Posts: 79
I'm actually going to disagree, in principle, though not necessarily in conclusion, with this assessment. While the wisdom of a separate HO property may be more complex and robust than is warranted - a MVO/SVO value of Vr/Vl (default vertical, rotated in horizontal) may be a simpler implementation - the ability to have this information computer parseable is the entire point of these properties. The point of these properties is to not only document the scripts that already exist, but to enable a simple data update to inform VIDI text display for scripts and characters that are yet to be encoded. If we were to instead hard-code that Mongolian and 'Phags Pa are the equivalent to HO=L, then that really precludes the easy implementation of any newly encoded scripts - or even new blocks of Mongolian/'Phags Pa - that are also default vertical and rotate in horizontal layout. The current model also allows for custom tailorization and updating of the rules for VIDI text - it may be that some Mongolian punctuation is actually favored in horizontal writing as U, rather than L, and the HO property can be simply updated in 4 years when implementations start being widespread. Having to go in and change the rules for all these implementations is an insanely high cost compared to having the next UCD update a character property.


Top
 Profile  
 
 Post subject: Re: The HO property is not needed and should be removed
PostPosted: Thu Jul 05, 2012 1:00 am 
Offline

Joined: Thu Mar 08, 2012 12:06 am
Posts: 3
The information does not need to be in property form to be "computer parseable". As it stands, the HO property is a completely verbose representation of the "upright in vertical, rotated left in horizontal" behavior. Layout systems and fonts will be designed with an eye towards how this is defined, so changing a codepoint to U will be not mean simply updating a data file, fonts designed using the previous definition would also need to change.

To simply enumerate vertical-only scripts with this behavior is much simpler at this point. If codepoint-by-codepoint variations are needed, then it's no more complex to shift to that later.


Top
 Profile  
 
 Post subject: Re: The HO property is not needed and should be removed
PostPosted: Fri Jul 06, 2012 5:46 am 
Offline

Joined: Mon Feb 01, 2010 6:18 pm
Posts: 79
jdaggett wrote:
The information does not need to be in property form to be "computer parseable". As it stands, the HO property is a completely verbose representation of the "upright in vertical, rotated left in horizontal" behavior. Layout systems and fonts will be designed with an eye towards how this is defined, so changing a codepoint to U will be not mean simply updating a data file, fonts designed using the previous definition would also need to change.

To simply enumerate vertical-only scripts with this behavior is much simpler at this point. If codepoint-by-codepoint variations are needed, then it's no more complex to shift to that later.


I'm sorry, but I don't quite get where you are coming from here. The whole point of having these data is that you don't have to update any fonts, because the data fields inform how VIDI text is laid out. The whole point is that you can get the two most common vertical presentations without having to specially program typefaces. By removing this data, you make it so that if we discover anomalies - or end users wish to define customized VIDI layout rules for special use - we have to somehow change the behavior without any properties to actually do it in, which necessitates either changing program source code or editing the fonts to do something different. That seems a recipe for chaos.


Top
 Profile  
 
 Post subject: Re: The HO property is not needed and should be removed
PostPosted: Sun Jul 08, 2012 9:14 pm 
Offline

Joined: Thu Mar 08, 2012 12:06 am
Posts: 3
vanisaac wrote:
I'm sorry, but I don't quite get where you are coming from here. The whole point of having these data is that you don't have to update any fonts, because the data fields inform how VIDI text is laid out. The whole point is that you can get the two most common vertical presentations without having to specially program typefaces. By removing this data, you make it so that if we discover anomalies - or end users wish to define customized VIDI layout rules for special use - we have to somehow change the behavior without any properties to actually do it in, which necessitates either changing program source code or editing the fonts to do something different. That seems a recipe for chaos.

You need to look more closely at the details of the scripts involved here. Mongolian and Phags-pa both require shaping, the glyph used for a given character is affected by it's position within a word just as in Arabic and ligation is required in some cases. Orientation decisions for a given character are not independent of other characters within the same script. The shaping rules in fonts that support these scripts embed the understanding that characters for these scripts share orientation. This is why I think it's far simpler to describe the nature of these scripts and not present the illusion that orientation decisions can be made on a per-codepoint basis in this case.

I think the theoretical model you have in your head is at odds with the practicalities of text layout for these scripts. Defining HO as a per-codepoint property here does introduce an abstraction but not a useful one.


Top
 Profile  
 
 Post subject: Re: The HO property is not needed and should be removed
PostPosted: Mon Jul 09, 2012 6:30 pm 
Offline

Joined: Mon Feb 01, 2010 6:18 pm
Posts: 79
jdaggett wrote:
You need to look more closely at the details of the scripts involved here. Mongolian and Phags-pa both require shaping, the glyph used for a given character is affected by it's position within a word just as in Arabic and ligation is required in some cases. Orientation decisions for a given character are not independent of other characters within the same script. The shaping rules in fonts that support these scripts embed the understanding that characters for these scripts share orientation. This is why I think it's far simpler to describe the nature of these scripts and not present the illusion that orientation decisions can be made on a per-codepoint basis in this case.

I think the theoretical model you have in your head is at odds with the practicalities of text layout for these scripts. Defining HO as a per-codepoint property here does introduce an abstraction but not a useful one.


I am well aware that both 'Phags Pa and Mongolian require contextual shaping - when you encode a complex script, you have to become an expert on the encoding and rendering model of every other script.

The problem with your perspective is that it treats all elements of the Mongolian and 'Phags Pa scripts as if they were letters that are contextually shaped. That is demonstrably false, as both also contain numerous punctuation and other non-alphabetic characters that may or may not be best treated like the alphabetic elements of those scripts. Futhermore, it precludes the simple addition of new blocks of Mongolian, 'Phags Pa, or other default vertical scripts to that VIDI behaviour. So what do we gain by eliminating this information from the VIDI properties and hard coding them? Just because something is defined on a per-character basis does not mean that it varies among characters that may share another feature, like script=Mongolian. By defining on a character basis, you simply indicate that it is a character property and not necessarily a property of every Unicode script (including script=common). There will probably be a few non-Mongolian/'Phags Pa code points with HO!=U as well, so eliminating this information will also effect other default vertical characters.


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 6 posts ] 

All times are UTC - 6 hours [ DST ]


Who is online

Users browsing this forum: Bing [Bot] and 3 guests


Quick-mod tools:
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
Template made by DEVPPL.com