forum

[Mia Kalish]

Are you in New Zealand? I thought you were in Hawaii? 

 

I Absolutely Agree about getting wider support for the fonts, but so far, I have determined that reason, logic, and right-thinking don’t always work very well. It’s true we can pay to have the fonts developed, and make them available for distribution, but that can get expensive. 

 

The way I have observed seems to work best – judging from all the acquisitions – is to have something someone else wants. Like our fonts that we are happily modifying ourselves. Then, people will decide there is a way to make money from them, and they will acquire or develop support. And it’s done. Although for some, like the Apachean family, changes have to happen in the ISO standards before there can be a nice localization. And, the grammar checking doesn’t fit the linearly oriented LEX style, because heads are embedded, so we need new paradigms for both grammar and spell-checking if they are to become automated. 

 

I am impressed by your programming skills, by the way. All these years I have known you and I never knew. 

 

:-) 

Mia  

 

  _____  

From: Indigenous Languages and Technology [mailto:ILAT at LISTSERV.ARIZONA.EDU] On Behalf Of Keola Donaghy
Sent: Thursday, February 14, 2008 9:24 PM
To: ILAT at LISTSERV.ARIZONA.EDU
Subject: Re: [ILAT] forum

 

On 15 Pep. 2008, at 5:05 PM, Mia Kalish wrote:

You know what? (Of course you don’t that was So Rhetorical). . . but here’s what: 

If you own the font, you can modify it. So if you have a pile of fonts for the University, you can change them yourself . . . or have a friend to it for you (You have such a friend). That way, you can have the special character in a pile of fonts.

I imagine there is a way around this for others, because the Intellectual Property rules restrict distribution of fonts to people who don’t own them. It doesn’t say that if you own the font, you and your friends can’t modify it. The rule is that you can modify it for your own use.

Yes. We have a few hundred postscript fonts that we've modified over the years for our own, in-house use. That's how we created the HI fonts. We do have a couple that were based on some free fonts (similar to Helv, Times, Palatino and Courier), and the creator of those fonts gave us his blessing to make the fonts available to the public. Those are the four (HI Keawe, HI Manokalanipo, HI Pi'ilani, and HI Kakuhihewa) which are downloadable from our website, along with the fonts.

 

We could easily modify the fonts again to add the glottal and macrons in the correct Unicode locations, but we're more concerned with widespread support for the language - getting the foundries and system vendors to include them in the fonts that everyone gets, since so much of our work is not going web-based. We donʻt want people to have to download special fonts anymore to type and print Hawaiian.



I haven’t played with the scripting in Unicode, because the characters are all individually coded with their unique numbers. Three character compositions have 3 sequential 4-byte codes (I think), and possibly a compositing instruction code. That means that if you are manually coding Unicode, the sequence in which you specify the characters matters, with the obvious results if people choose sequences randomly, thinking that the appearance of the character is what is important.

You description sounds like the fonts were manually coded. Characters sort by their internal number; when you specified your preferred sort sequence, whoever was helping recoded the glyphs so they would sort in the specified order. Locales, lucky people <:-)> People don’t notice localization because they don’t have to live without it.

No, the characters we use are found in some, but not all, Mac and Win fonts. Many have the macron-vowel characters, few have the glottal. The common ones that do are Lucida Grande (Mac) and Lucida Sans Unicode (Win). Regarding sorting, Apple created the system-level sort routines based on our input. I've never bothered to see how the sorting worked when the Hawaiian locale in the "International" system pref was not selected. That's what triggers our preferred sorting.

 I don’t have Vista, although the changes sound intriguing. As for OneSpellCheck fits OldAndNew . . . the internal codes are different. The simplest approach is to have a utility that converts one to the other . . . this means that the utility has to tiptoe through Word codes, changing only the appropriate font codes. This is not overly difficult for a programmer, but is definitely non-trivial for many. Our problem in Athapascan with “old fonts” is that people replaced the Number sequence with the special vowels and characters. Can you see what an amazing nightmare that is for sorting, searching, databases? Aaaargh. But then, it was all people seemed to know how to do.

The Vista spell-check functionality requires that user spell check dictionaries be encoded as Unicode. I thought that this was an issue, but realized our umlaut characters are represented in Unicode as well. I simply resaved the list as Unicode (keeping the umlaut characters) and it works fine. A document typed using our HI fonts works fine. I then converted the entire word list, converted it to Hawaiian Unicode using a macro I wrote. I appended this list to the bottom of the previously mentioned list and saved. So the spell check document contains two complete lists - one in HI format and one in Unicode. That one document allows you to spell check either HI formatted or Unicode formatted documents. I did have to go through and remove duplicates - words which lacked both glottal and macron vowels which would appear in both lists - but that was pretty easy. If I could only remember how I did it ;-)

 



I took particular delight in “stepping on” characters that were not used in either English or Athapascan to create my fonts. It horrified the Unicode people, which, considering their complete non-responsiveness to Apachean issues, made my little techie heart day. (I’m so bad.)

Good for you!

I’m interested in your issues with your ancient program . . . 

For another day... or year perhaps. It was FirstClass. They were slow to become Unicode compliant, and then to update their localization program so we could continue to translate the Client into Hawaiian and use Unicode to represent the language in menus and such. They are addressing that now and I hope to have a new client and server up this summer that will free us of our dependance on the custom fonts.

And thanks, I’m sure this discussion is or will be useful for lots of people.

Going out into the Tsaile AZ mountain snow, now coming down in huge clots.

 

No snow in NZ but getting cold, by our standards ;-)

 

Aloha,

 

Keola

 





 

 

  _____  

From: Indigenous Languages and Technology [mailto:ILAT at LISTSERV.ARIZONA.EDU] On Behalf Of Keola Donaghy
Sent: Thursday, February 14, 2008 8:47 PM
To: ILAT at LISTSERV.ARIZONA.EDU
Subject: Re: [ILAT] forum

 

Aloha Mia. I don't know if the language scripts in Unicode have anything to do with sorting routines. When we worked with Apple to get the Hawaiian keyboard into OS X back in 1991 or 1992, we provided them with our preferred sort order (vowels, Hawaiian consonants with the glottal last among those, then the borrowed consonants). We also had to define the vowel-macrons as a secondary sort, so that if there were (and there are) words that differ only based on the presence of a macron over a vowel, the word with non-macron is first, then the macron vowel. 

 

It took us several years of lobbying and it was only a chance meeting of a friend of mine who works at Apple's headquarters with some Apple system engineers that led to us getting Hawaiian support in OS X. They have since added may more languages. So we have the Hawaiian keyboard, a Hawaiian locale (which enables us to create localizations of OS X applications and have them automatically displayed if there is a localization available), as well as the custom sort routines and localized date and time strings.

 

When we created our HI fonts in the early 1990s, spell-checkers were also an issue. That is why we replaced the y-umlaut character ( ÿ ) with the glottal (‘) and replaced the vowel - umlaut combinations with the vowel macrons. In the case of the ÿ, it also had the advantage of allowing us to sort properly (in most cases). Back then ClarisWorks and Word did not accept punctuation marks and other symbols within words, so we had to find a way to make it work, and settled on ÿ. 

 

Office for Vista now handles custom spell-check documents with in Unicode fine. Office for XP and Office for Mac 2004 did not, which was an issue. I have not yet received a copy of Office 2008 for Mac so don't know if it can use spell-check documents formated in Unicode yet. We are working on such a document right now. At some point I'll have a single document that can spell-check both documents typed in our old HI font system as well as Unicode. I'll be a happy man on that day. ;-)

 

I'll look into the other Unicode glottal characters, perhaps someone here is more familiar with them than I. To be honest, I somewhat regret our decision to use the non-breaking character because it is available in so few fonts, however, that it was the one I was encouraged to use by people who knew better than I. 

 

We're in a very strange transitional place right now in our adaptation of Unicode. Because of one particular application that we depend heavily upon, we have not been able to completely abandon our old custom fonts and go to Unicode. I'm hoping to have that issue taken care of by this summer.

 

I'll get back to you on the other glottal character(s).

 

Hope this help! Feel free to email if I can be of any other assistance.

 

Keola

 

On 15 Pep. 2008, at 4:26 PM, Mia Kalish wrote:






Yes, this is exactly what I mean. I’m not familiar with U+02BB. Hyphenation is not an issue in Athapascan; I don’t think I have ever seen anything hyphenated. But, the glottal has to sort first, before the vowels it affects, ‘a is a typical form and is actually the most common glottalization of leading vowels, although other vowels are glottalized in the same way.

 

Also, since we don’t have Athapascan language and grammar, we have to use English, so we have to fool English into thinking the characters we are typing actually belong to it so we can use spell-check.

 

How do you do this in Hawaiian? I just looked at my language options for Word, and Hawaiian is a recognized language, We don’t have this luxury, and although there is now a language code for NAV and APA, there are differences in the 6 Apaches, and only one language code. Sigh. People are so stingy, aren’t they?

 

 

 

 

========================================================================

Keola Donaghy                                           

Assistant Professor of Hawaiian Studies 

Ka Haka 'Ula O Ke'elikolani             keola at leoki.uhh.hawaii.edu 

University of Hawai'i at Hilo           http://www2.hawaii.edu/~donaghy/

 

"Tír gan teanga, tír gan anam."  (Irish Gaelic saying)

A country without its language is a country without its soul.

========================================================================

 






 

 

 

 

========================================================================

Keola Donaghy                                           

Assistant Professor of Hawaiian Studies 

Ka Haka 'Ula O Ke'elikolani             keola at leoki.uhh.hawaii.edu 

University of Hawai'i at Hilo           http://www2.hawaii.edu/~donaghy/

 

"Tír gan teanga, tír gan anam."  (Irish Gaelic saying)

A country without its language is a country without its soul.

========================================================================

 





 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listserv.linguistlist.org/pipermail/ilat/attachments/20080225/cb2eb03a/attachment.htm>