On leave from work. Did some digging of holes in the garden, some scuba diving, and some coding. I also helped someone move to Newcastle, but that is too traumatic to speak of. I didn’t get nearly as much coding in as I wanted to. Thursday was CLUG, which included a visit from people such as Alan Cox, who has recently fallen off a horse.
I am currently working on a bug in Panda 0.4 which means that blank pages cause the library to segmentation fault.
Back from LCA. The world in Canberra hasn’t changed much. People don’t seem to be rushing forward to offer assistance for the running of next years LCA in Canberra. I guess I’ll just wait and see who comes out of the wood work this week before I make a decision either way.
Worked a lot on Panda today — it pretty much has a full set of line drawing options now (these will be released in 0.4 sometime real soon now). I just have to work out what the story is with the fills, which don’t seem to work at the moment. I also spent a small amount of time looking at the man pages for zlib, because they have been on the TODO list for way too long.
Dug a few holes in the back garden as well. I have a space out the back that I want to put a pergola on, but before I can do that I need to lay some concrete, and before I can do that I need to run the pipes for the inground watering once and for all. So I spent a lot of today digging trenches for the plumbing.
Tridge seemed real keen on FreeCIV, so I have downloaded it and will give it a try. Another good idea I got from LCA is inline docbook documentation for Panda, which I have also started playing with.
Overall, a fairly productive day.
Tridge talked about his latest hacks for the TiVo. See tivo.samba.org for more details.
At the dinner John ‘Maddog’ Hall gave a very interesting talk regarding the problems his parents have with using technology. It reinforced some of the issues about user interface design and ease of use that face Linux today. To a large extent I am grateful that I don’t currently develop anything with a user interface, so this isn’t my problem. The Mac when it first came out is a good example of how to do things right.
Hugh Blemming of LinuxCare OzLabs fame gave a talk comparing the difficulties of developing drivers for a device where the manufacturer was supportive, with developing drivers for a device where the manufacturer was not supportive. The talk also discussed reverse engineering, and some of the serial protocol techniques that Hugh used. overall a very good talk.
Raph is the new maintainer of ghostscript, and is also the developer of libart. He gave a very interesting talk (especially given my PDF bent) about 2D graphics, the next version of PDF, and things like that. He also talked about the directions he saw ghostscript moving in the future. I also had an interesting talk at the end of the talk with Raph about PDF 1.4 (I now have enough information to start implementing it in Panda), and with an employee of an airline which uses a product called Open Skies to send itinery faxes out to customers. They are now looking into using Panda to generate the faxes before they are sent out — it should be cool because it will only take about a dozen lines of PandaScript code.
I need one of these machines…
Rusty gave a very funny talk about kernel hacking and how it makes you more attractive to women. Pretty much the key note you have when you are not having a keynote.
Ho hum. I wonder if these things will ever actually be released. There are interesting features here, like multiple page sizes, and variable page sizes, i386 binary support and PA-RISC binary affinity (binary conversion at run time). They will be cool if they ever arrive, but I think I am of the opinion that the world is much better off going with one of the risc architectures like PPC or Sparc.
David is easily the best speaker that has been at the conference so far. He took the relatively complex issues of scattered gather ip packet sending, and network card checksum generation / verification, and made it interesting and relevant. He also had many interesting stories about things like IBM’s reaction to the Linux dominance of the current SpecWeb99 statistics and stuff like that…
The most interesting comment he made was that there are some people who want to design constantly, and sit down as talk about how to make something the best it can be, whilst he would rather people went home and coded, and then showed the world the code… The basic gist was that something is better than nothing, and that code can always be improved once something actually exists.