2013-11-12 Yesterday, I've talked about Performance at ChaosSeminar of CCC Ulm. [0] It's been a nice talk without much ``action'', which I appreciated. I think this talk is one of the better ones, although the beginning wasn't smooth enough. I was able to transfer some major points of my view. In the afterthought, it occured to me that a part of the audience needed quite some time to realize that I'm not really talking about what is usually meant with ``performance'' and it's optimi- zation. This was quite a surprise for me, but, of course, if they don't know me well enough yet and the talk is titled ``On Perfor- mance'', what else should they expect? The talk is a collection of various thoughts, brought together to become one topic, one talk. I'm satisfied with the result, although I needed to rearrange things until just before the talk. Other talks had straight forward story lines, this one is clearly a collection. I think, I like both kinds. The next one is likely to be a straight forward topic again. That means less work and less uncertainty. The recordings are already online, thanks to Bjoern's incredible work. [1] For the first time, the announcement poster [2] included the CCC Ulm logo. I've included it into with the Heirloom -mpictures troff macros: Further more, I've generalized the poster template. The next step would be to add a frontend to it in order to allow others to create posters too. When I've formated my (troff) slides, I needed to manually add a BoundingBox to the PostScript sources, to crop the A4/Letter pages to the screen format. In the troff pipeline, just before I write the PS to disk: awk 'NR==2 {print "%%BoundingBox: 0 0 600 450"} 1' Ghostview (gv) displayed the slides then per- fectly, but xpdf didn't respect it. In conse- quence, I've switched to gv for the presenta- tions. This was quite okay, because it displays a small slide list on the left. Nonetheless, I wanted to have 4:3-formated PDFs. Today, I was able to solve the problem. The BoundingBox was only part of the necessary addition, without a correkt EPS header line, the BoundingBox was ignored because the header told that the file is plain PS. A PS header (first line) looks like this: %!PS-Adobe-3.0 The EPS variant looks like this: %!PS-Adobe-3.0 EPSF-0.0 [3] Hence, the new BoundingBox-insertion command is this one: sed '1s,$$, EPSF-0.0\n%%BoundingBox: 0 0 600 450,' (The `$$' expands to a single `$' in makefiles. Maybe I should rather use: sed '1s,.*,& EPSF-0.0\n%%BoundingBox: 0 0 600 450,' to remove the need to care for the dollar sign. ) I've also switched to sed(1), because it felt more appropriate. The makefile action reads: foo.pdf: foo.ps ps2pdf -dEPSCrop -dPDFSETTINGS=/prepress $< [0] http://marmaro.de/docs/chaosseminar/on-performance/ [1] http://ulm.ccc.de/ChaosSeminar/2013/11_performance [2] http://marmaro.de/docs/chaosseminar/on-performance/poster/ [3] https://en.wikipedia.org/wiki/Encapsulated_PostScript http://marmaro.de/lue/ markus schnalke