Older blog entries for karlberry (starting at number 202)

I saw a question in another blog (sorry, I neglected to write down which one) about pdfTeX's \pdfnormaldeviate primitive (also exists in MetaPost). Since I feel semi-responsible for the pdfTeX manual these days, I looked into the sources, talked to Thanh and Taco, refreshed my memory of normal distributions via wikipedia :), and came up with this:


Generate a normally-distributed random integer with a mean of~0 and standard deviation 65\,536. That is, about 68\% of the time, the result will be between $-65536$ and $65536$ (one standard deviation away from the mean). About 95\% of results will be within two standard deviations, and 99.7\% within three. This primitive expands to a list of tokens. \introduced{1.30.0}

Feel free to send me any questions/comments about this or other bits in the pdftex manual. Simplest to email me directly.

Not sure if it will matter to anyone reading this, but I guess I'll mention that the early bird discount on TUG membership goes away after March 31. Please join or renew if you haven't already, and thanks.

in response to will and randomdeterminism who replied on their blogs to my post. (btw, randomdeterminism, who are you, if you don't mind telling me? just curious.)

1. it's not the state of affairs of tables in latex is so bad. please try to avoid such flame bait :). it's a matter of what i know in my head -- i know \halign and i don't have anything but the simplest table constructs in latex and context in my head. i didn't doubt that both latex and context can do the job; anything can do the job. as always, my goal was to get it done as quickly as possible so i can go on to some other overdue project :).

2. as for pdfpages, it (specifically, \includepdfmerge) seemed like the natural choice; perhaps that is just my ignorance.. the real case is not one image, it is 33 different images. 33 different \useexternalfigure commands doesn't sound fun.

3. you finessed the question of scaling by explicitly specifying arbitrary sizes. part of my problem was that my sizes were not obvious, the scaling is not the same in all cases, and i didn't want to compute a bunch of five-decimal place numbers by hand.

4. the real figure will appear in the next tugboat, we can debate it further then :).

I wanted to typeset a 6x6 grid, with each cell framed, and a PDF image in each cell -- and with the center 2x2 square replaced with another image. Except for the center "window", it is trivial with pdfpages. With the window, I could see no good way to eradicate the frame lines in those cells, or put them back after the fact.

I also tried doing it in four blocks (above the window, to the left/right of the window, below the window). The main problem there was the scaling. I saw no easy way to scale the 2x2 left/right blocks the same as the 6x2 top/bottom blocks, since the scale factor was not predetermined and not known. (I think I'll ask the pdfpages author to write out the calculated numbers in the .log.)

I ended up using plain \halign, just the way the TeXbook says to do ruled tables :). The entire template was &#&\vrule width.5pt #\cr (the leading & repeats it as often as needed). Turned out to be way easier to have complete control in this case.

Posted something of a planned schedule for TeX Live 2011. Slippage of months would not be unusual.

Posted registration page and more information for TUG 2011, to be held in Kerala, India this October.

Think I got the new 0.8.3 release of biber in TeX Live's Build and Master trees. So many special cases, oh well ...

Been enjoying James Gleick's new book, Information. A quote from Ada, on trying to (what we would now call) debug her programs for Babbage's (theoretical) Analytical Engine:

"I am in much dismay at having got into so amazing a quagmire & botheration."

More info on TUG'11 in India. Registration form, bursary applications, still coming. If questions or ideas, email tug2011 (at) tug (dot) org.

Finally installed biber in TeX Live, that is, the precompiled binaries made by the authors (with Par::Packer, which I hadn't known about before; very cool). We'll see how many other TL builders want to try ...

193 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!