World: r3wp

so if the bounding box does that for a single word, we need to use 
all chars in the alphabet to calculate the height correctly for any 
char combination. it does not do that right now

If vertical position should be in the dialect, I think, it should 
be based on a real postscript feature. If such a thing isn't found 
in postscript, it should be implemented on a higher level, shouldn't 
it?

The problem is to get the height accurately for all glyphs for a 
font (I just got that now). Some fonts have very high tops and very 
low bottoms and you want that to work for any font and letter combination 
we throw at it. it would be harder to get that information inside 
REBOL than inside postscript.

I got it rendered now so that the vertical position is at least consistent 
now, but the font is offset a few points too far down. I don't know 
why yet, but it's probably the baseline again.

there is another method which is to get the bbox information from 
the font metric file itself. it's much faster, since it's just lookup, 
but I couldn't find an example of how to read it.

so my current method is Graham's, except for calculating the entire 
alphabet. we should probably work on getting that information from 
the font metric file instead.

In Adobe documentation, they operate with "character origin", which 
is the baseline, I think. If you do:
40 50 moveto (ABC) show
, the origin of the A letter is at coordinate (40, 50).

So it doesn't matter, how high the font is.

you still need to know the size of the bbox, don't you?

A vertical positioning would be a simple calculation. Say, you want 
the baseline to be positioned at the middle of a 300 point high box 
starting at pos (100, 100). You then start by putting text at coord 
(100, 250).

The y coord is 300 / 2 plus the 100 = 250.

that assumes that the text is rendered from its own center point. 
it isn't, is it?

If you wanna build a table, you calculate the position of each line 
like this, and send the precise coords to the postscript dialect. 
It would probably be best to build a new dialect on top of the postscript 
one, and in that dialect handle the calculations.

No, text is rendered from the "character origin" (the baseline), 
which can be different than the center of the text. This should probably 
be taken care of, if needed. hm

the table is a different matter :-) it's already done, not yet as 
a dialect though. I will build one later.

The baseline problem is the one I'm hunting a solution for. About 
the font metrics, the bbox for the whole font is stored in FontBBox 
in the font dictionary.

If your goal is, that the text is completely centered vertically, 
then you have a problem. Because how do you know, what characters 
to be written? If there is no 'g' in the word, you need to position 
it a bit lower, than if you had 'g' in the text, for it to be completely 
middle aligned.

that you solve by calculating the bbox for the entire alphabet instead 
of just the words you write

this way the height will be constant

and this information shouldn't need to be calced if we can get that 
from FontBBox in the font dictionary. still hunting for how to obtain 
it. :-)

Ok, I'm a bit lost. I'm not completely sure, what you're trying to 
achieve.

accurately vertically centered text

what I did at first was just to use the font size. that doesn't work, 
because it does not take the baseline into account.

then I used Graham's method and it does practically the same thing 
(I forgot to test for letters like "g"), but you can successfully 
calculate the bbox for a text string this way, if you need it.

Combining Graham's method with getting the height from the entire 
alphabet instead of the single word would eliminate the varying height 
problem you mention. But the baseline problem turns up again.

The bbox information for the entire font is stored in each font metric 
file as a [llx, lly, urx, ury] coordinate set, so I shouldn't need 
to calculate it. This information is crucial in order to get one 
line of vertically centered text. I already got everything in place 
except that particular number. :-)

the source you are reading contains Graham's method

Isn't this centered?

PageSize A4
page [
	font [Times 40] linewidth 0
	at 0x421 "Centered text" center 595
]

well, that's hardcoded :-)

Yes, that's what I think, you should do. Build a dialect with keywords 
as top, middle, bottom etc. and make that dialect know the size of 
the paper and correctly calculate the positions. That's the way to 
do it, I think. I would do it like that.

as mentioned before, that won't work, because the font information 
is not obtainable from inside REBOL. it has to be done inside postscript. 
I've already tried that method a year ago and it failed. :-)

Remember what kind of language, PS is.

There is no information about font (other than Times 40) in my example.

The 0x421 and 595 comes from the papersize, which is 595x842 when 
using A4.

ok, tell me then the algorithm for calculating the font height inside 
REBOL. :-) you cannot possibly know without reading and parsing the 
.PFM files yourself.

and as mentioned at first, I already tried the font size... doesn't 
work.

Why would I need the font height? I didn't need that in my example 
to center the text.

no, because you probably were eyeballing it.

No, I didn't! :)

center the text
 ... are you talking about _horizontally_ centered text?

When you do things like:
at 100x200 "some text"

the character origin of the first character in the text is at position 
100x200.

No, horiz and vert!

where did you get 0x421 from?

0 is left side of paper. 421 is height of paper divided by 2. 842 
/ 2 = 421.

Then it's centered.

if that's the case, let me test the theory

what?...

why doesn't that work for smaller cells?

PageSize A4
page [
	font [Times 18]
	at 60x421 "abcdefg"
	font [Times 36]
	at 130x421 "abcdefg"
	font [Times 96]
	at 260x421 "abcdefg"
]

In my example, the baseline of all the text is vertical centered 
on the page.