[George Herman]

Posted by: George Herman
         
Hi,

I tried joining the VOLT community board, but it's been a few weeks and I haven't received an answer, so I figured I would post here.

I am trying to create some teamim within some Hebrew fonts that I have.

I am using VOLT, and I've got the basics of how to use it down, just a few things elude me.

Creating the default position for nikkud and teamim is fairly straightforward, you need to create glyph groups of the base characters and then position lookups for each mark and position them correctly. It's when they run into each where you run into problems. It seems there are two (maybe more?) ways to position the marks correctly.

1) Create pair adjustments (the pair would be the two marks) with the correct context before/after.

The problem with this method I've found is that there is no way to visually view the positioning of the two marks in contrast to the base character without compiling and proofing every time. It's a guessing game. Am I missing something?

2) Create glyphs with the base and two marks correctly positioned and make a substitution lookup to substitute the base character and marks to the glyph that I created (it would actually have to be two substitution lookups).

This way seems endless, how many glyphs to make? And which ones? And how do I know every problem situation that would need a new glyph?

Finally, I don't understand why there are two separate features, Mark To Mark feature and Mark positioning feature. Can't you do mark to mark positioning with the mark positioning feature?
         

[John Hudson]

Posted by: John Hudson
         
The current shipping version of SBL Hebrew, and previous versions, uses your method 1. As you say, it is a bit of a pain to implement, since you have to approximate and then proof the positioning. First you have to contextually adjust the position of the first mark on the base in context of following mark(s) and then position the latter relative to the first mark. Not only is this difficult, but it also doesn't provide ideal results, unless you include huge numbers of contextual lookups for discrete mark sequences, rather than grouping marks by similar width as in SBL Hebrew and Ezra SIL, the two fonts for which this method was developed. The problems with this approach are particulaly acute when mark sequences involve pre-positional characters like dehi and yetiv that are input after the characters they visually precede.

In the next version of SBl Hebrew, 2.00, I am employing a different method. The above described method is still provided as a fallback for arbitrary mark sequences, but I also include precomposed mark combination glyphs for all of the unique below base sequences that occur in the Bible text. So this is a little like your method 2, only instead of making combinations of base + marks, I am only making combinations of marks + marks. And only for combinations below the letters: the contextual behaviour of holam makes it difficult to do the same thing for above base combinations. The precomposed mark combination glyphs are applied with a ligature lookup in the feature.

Regarding the vs feature distinction: the latter feature is not normally used for Hebrew. It is very useful to have this distinction for some scripts, but for Hebrew you can do everything in the feature, and since there is a lot of contextual interaction between the rules for marks applied to bases and marks relative to marks it is easier to group the lookups within a single feature.

The OpenType Layout logic for Biblical Hebrew in the SBL font is open source. My eventual aim is to provide a template font for developers to follow, but I want to do this in a way that clearly distinguishes the layout logic from the letterform design, which is not open source. So I will create a kind of skeletal template design. In the meantime though I will be happy to provide you with a VOLT project source file for the current SBL Hebrew release. I am at a conference this week, but if you contact me by email after next Tuesday I can arrange this.
         

[George Herman]

Posted by: George Herman
         
Thank you very much. John, for the detailed reply.

Let me first shower you with compliments: The Windows Glyph Processing article you wrote for Microsoft is superb. Extremely well written and very easy to understand, I really enjoyed it. I am also very thankful to all the time you've taken to answer posters like me. I've gained a ton of knowledge browsing through forums and reading all the replies you've (in particular Typophile forums).

I didn't realize until now how difficult the first method is. I see that there is no way ( or am I mistaken?) to save move one mark relative to the positioning of the other mark, so you pretty much have to go through every letter (contextually) every time there is a collision and adjust them. The fact that you have to guess makes this ten times as hard.

Mark + mark glyphs is very clever, however as with my second method it comes with its setbacks. Being that now the two marks are one glyph the user has no way of controlling their positioning. If someone using the font decides they want, for example, the Patakh and Etnahta to be positioned a little differently they would be unable to.

You  can expect an email from me after Tuesday. I would be very grateful for the SBL Hebrew VOLT project source file.

Thank you again.