Work-In-Progress Reports for July 2001

Density Compression Routine Ported to Amiga: Ported density compression routine to the Amiga (after much struggling with the stupid Amiga compiler we use). So we removed a few features that made it very slow on the Amiga, consequently getting us slightly less compression ratios. However the time it takes is now only about 1 sec/track, instead of about 45 sec/track. (more)

Density Compression: * Wrote density table decompression routine. * Implemented loads of integrity checks, to ensure proper data migration between Amiga and PC, and proper compression. * Fixed some minor bugs in compression, found through extensive testing. * Tested on a lot of track images, now both compression and decompression work as expected. (more)

Timing Table Compression: The timing table compression was written. It is about 2...4 Kb/track, which is a bit higher cost than we expected. Still, not bad compared to uncompressable 50 Kb. The compression algorithm became quite complex though, and since there would be too chances for error when converting it to 68000 assembly, so it stays in C. Not worth the effort at the moment. Analysed the compressor output and made a few improvements. (more)

CTA Graphs: The following are timing graphs for a Copylock track and a "long" track. Pretty cool. {{ :track:old_zoom_copylock.png |Copylock Track}}\\ {{ :track:old_zoom_long.png |Long Track}} It is now known that the time sampling works 100%, so the work on compression can start. (more)

CTA Work Started: Not much happened during the weekend, mostly due to the weather being incredibly hot. Work started on the final CTA (CT Analyser). Made a prototype. The density graph will get some development time so we can actually see a few tracks. If they "look" good, we can make the timing table compressor (see older WIP). This will mostly bring down the timing table from 50 Kb to around 1 Kb. (more)

Setup: Setting up development environments, etc.: * Set up an recently bought A1200 system. * Migrated old sources and tools from A500. * CT version 2.0 testing - does indeed work perfectly on new hardware. * Checked "RN", the Copylock copying tool. The FDC reads the data as it was completely valid. Seems Commodore improved it quite a lot. (more)