January 11th, 2009

Apollo 4 on column of fire

A headline seen on BuzzFlash.com

Message to Barack: You are Carrying Things a Bit Too Far with This "Bi-Partisan" Nonsense. The Repubs are Just Going to Betray You, So Start Punching and Stop Trying to be "Liked" by Everyone. "Obama to honor McCain on inauguration eve." Yuch!

Message to BuzzFlash: You, hypocrite, have a beam in your eye: you refuse to accept Barack Obama’s overtures. I can’t help you get that beam out, unfortunately, because I must deal with my own: I find it hard to bear people who refuse such overtures. At a brunch, I would get along more easily with Rick Warren.

Apollo 4 on column of fire

Urge to kill RISING ...

It is unbelievable to me how braindead the handling of OpenType is in the ‘mainstream’ free software community—by ‘mainstream’ I mean outside of the world of TeX and it’s relatives. Cairo, the graphical back-end of ‘choice’ these days, is writing useless ToUnicode maps into PDFs (because some document that looks like it is meant only for Far Eastern scripts ‘recommends’ the useless maps), and is performing needless gymnastics with ‘glyph clusters’ that become necessary only because of lack of a useful ToUnicode map. The ‘glyph clusters’, meanwhile, show up also in HarfBuzz, the OpenType engine of ‘choice’—a piece of software that not does lookups in the wrong order, whose source contains nonsensical comments about ‘feature order’, whose bug reporting system contains screaming from people who don’t want the code associated with the nonsensical comments fixed, and which as a result of all this is unbelievably complicated and difficult to use.

No one in the TeX world seems willing to touch this stuff. It’s radioactive. Xdvipdfmx still doesn’t write the useful ToUnicode maps, I think, but that can easily be remedied. I just haven’t gotten around to it. Adding meaning ToUnicode support to ConTeXt Mark IV was pretty trivial, and I haven’t bothered to supply my patch mainly because I expect the developers to get around to it themselves, once the design has settled down some. But the free GUI-oriented stuff that is out there is hard to stomach, and it irritates me even to try to help improve matters.

Seriously, there is much really wonderful development in the free software world, such as the operating system I’m using, which makes Microsoft Windows look like a pile of dog crap, but the ‘mainstream’ software for production of text is incapable and bad. Hang onto your Adobe CS.

Late edit: Mind you, I suspect that the ‘glyph clusters’ may actually serve a purpose in some non-latin scripts, like maybe arabic with all its incredible fanciness, and they may also be useful in some unusual cases with Latin scripts, but I haven’t found such things necessary when making e-books.

Late late edit: I don’t want to be un-nice, and much of the free software works pretty well for me, at least with my patched-up fontconfig and if I don’t try to do anything fancy, but I feel released to rant like this because Adam Twardoch, a name well known in the font community, already had a similar rant a few weeks ago in a FontForge mailing list. He was particularly annoyed about Scribus. (His annoyance with FontForge was FF’s crashiness, although that at least is a documented misfeature. :) )
Apollo 4 on column of fire

Obsessive-compulsive disorder

This is how it goes. If I get myself involved in the software community in a way that involves a lot of uncertainty, I have difficulty getting the problems out of my head. I like certainty; if I write a patch for FontForge, it almost surely will be accepted. On the other hand, even though I am on perfectly good terms with the Unicon programming language community—I’m officially some kind of baronet or the like in that community—if I write a non-trivial patch it will end up in a years-long queue for consideration, while the target moves and eventually I abandon my own patch. And it never ends; my mind touches on all these things countless times throughout the day, and in an anxious way that is mixed in with thoughts about war carnage and the fat content of foods.