2012/12/7 Adam Twardoch (List) <[email protected]>: > I think proper vertical writing mode would be best, without resorting to > multiple rotations :) > > Yes, the "vert" is used for punctuation but also for Latin glyphs, which are > often proportional in "height" (ie. avtually "rotated width"). "vert" is > typically used in combination with "vkrn", which then does vertical kerning. > So if toy have the text "Typeface" set in a CJK font vertically, the "vert" > feature would map the glyphs T, y, p, o, g etc. to their counterparts which > are rotated by 90 deg to the right, and then the "vkrn" feature would move > the rotated "y" up if following a rotated "T" -- by the same amount as the > normal "Ty" kerning pair. > > Of course not all CJK fonts include rotated Latin glyphs, let alone > proportional, and let alone with vertical kerning. But some do (especially > Adobe's Japanese fonts). > > There is another set of features (fwid, pwid, hwid -- AFAIR) which switch > Latin glyphs and punctuation between full-width (monospaced), proportional > (normal from the European perspective) and half-width (narrow monospaced) > variants. Using them is up to the user's preference and discretion. Some but > not all fonts provide rotated forms for all these variants as well. > > But all this stuff is made to work with the "true" vertical typesetting > direction in mind, and has to do with things that happen within the line. > > Stuff that happens on the higher level, i.e. by which means you actually get > entire lines to be laid out in the vertical direction, is a wholly another > level. > > As I mentioned, the "@" hack is just a Windows API trick -- these special > fonts don't really exist, they're being synthetically generated by the > Windows API and are exposed to apps that use the Windows API to do text > layout and rendering. > > But XeTeX accesses font files more directly, and directly parses OpenType > Layout tables, so the "@" trick shouldn't work since these are not "real" > fonts. > > And, as I said at the beginning, I think proper vertical writing mode in > XeTeX would be the *real* solution, without resorting to OS API tricks. > Vertical typesetting is quite a challenge. Consider Old Mongolian. It also writes vertically from top to bottom but unlike Chinese and Japanese, from left to right.
> However, I'm not at all a TeX expert so I don't know which options are > available. > > Best, > Adam > > Sent from my mobile phone. > > On 07.12.2012, at 00:03, Gerrit <[email protected]> wrote: > > Hello Adam, > > thanks for you answer. I didn't know that the @ thing is a Windows feature. > Well then I guess it does not work. > > I just wondered if it might be an easy and actually good way to create > vertical Japanese texts (not just a paragraph or a text box, but the entire > document): Everything like columns, page break, sections etc. would work > flawlessly . Incorporating Western text in the text would also work without > any problems. > > Basically I just thought about using the @ font, rotating the entire page > 90° clockwise (so that the text is vertically and the alignment is correct), > flipping the width and height size of the paper, so that a portrait paper > stays a portrait paper, and then the text would work. > Horizontally written picture boxes/captions or tables etc. could be done > with the normal rotating package (i.e., rotating them back). The only > problem would maybe be the head and foot, because the rotated page would > then have the head and foot on the right and left side, resp. But I thought > about tackling that issue afterwords. Either way, I just wanted to try it > out. > > Rotating every glyph independently (like it is done in the xetex manual) > does not seem to be that suitable for longer texts, and you would have to > cope with many many many packages and other problems. > > As far as I understood the vert feature, it works for rotating stuff like > the colon (ten), the full stop (maru), brackets etc. A normal character > would not have to be rotated. This is then necessary if you actually do it > for rotating every single glyph, but not if the entire text becomes rotated > and you basically just rotate the page backwards. > > I actually really think that something like the @ thing would be the easiest > way to implement vertical typeset into Xetex. > > Gerrit > > Am 06.12.2012 23:48, schrieb Adam Twardoch (List): > > Gerrit, > > this is a custom functionality of the Windows API, a "poor man's" method to > get vertical typesetting in "normal" applications which cannot deal with > real vertical typesetting. The "vert" feature is different: it provides > additional 90 degree rotation for those glyphs which are read better in a > horizontal arrangement rotated by 90 degrees. I.e. you use the "vert" > feature in a *real* vertcal typesetting context where CJK glyphs occur one > under the other, but e.g. for Latin glyphs it makes sense to set them so > that the reader has to turn his head to the right. > > So "vert" is completely independent of what you're asking. If XeTeX cannot > do "proper" vertical typesetting then perhaps indeed there should be a font > selection function that just rotates everything set in that font. I'd rather > have such a mechanism exposed than to rely on a non-cross-platform "@" > prefix "OS hack" (a hack actually provided by the OS). I don't know whether > such mechanism already exists in XeTeX though. Perhaps it does? > > Either way, you'd still want to apply the "vert" feature to do additional 90 > degree rotation for certain glyphs, or -- if used in the scenario you're > proposing -- to actually *un-rotate* them, so they bacome horizontal again. > > A. > > > > > -------------------------------------------------- > Subscriptions, Archive, and List information, etc.: > http://tug.org/mailman/listinfo/xetex > > > > > -------------------------------------------------- > Subscriptions, Archive, and List information, etc.: > http://tug.org/mailman/listinfo/xetex > -- Zdeněk Wagner http://hroch486.icpf.cas.cz/wagner/ http://icebearsoft.euweb.cz -------------------------------------------------- Subscriptions, Archive, and List information, etc.: http://tug.org/mailman/listinfo/xetex
