Keep Moving Forward
Posted by Trixter on June 13, 2007
We are machines. Biological in nature — chemical, actually — but all the same are we: Machines. And machines have quirks, and design flaws, and patterned behavior.
I can only assume that my Serotonin valves get stuck from time to time, because for the last few days I’ve been slowly getting back up on the horse. Well… a different horse. I’m doing stuff; just not the stuff I’m “supposed” to be doing. In the last few days, I have accomplished the following:
- Built new machine for Max out of spare parts, connected it to the Internet (scary for a 7-yr-old (and his parents!), but he wanted it), made it stream TV shows from the ReplayTV unit wirelessly through two floors
- Cut the back lawn, which had grass a foot high (not exaggerating). Hopefully my crazy neighbor will stop calling the police now.
- Returned the email of a few old friends — they had emailed me in 2004 (!!!!) and I only NOW returned the mail, but hey, I always answer my email. Eventually.
- Figured out a new way to get more performance out of the 8088 Corruption player and started work on the first phase of implementing it (hint: A/V chunks now need to be sector-aligned). Also brainstormed SpeakerTracker implementation ideas.
- Cleaned up the DVD-quality version of the MobyGames Classic Gaming Expo 2004 retrospective video and made it available online
- Submitted a backlog of about 150 cover scans to MobyGames that I had scanned in 2005 but never edited/uploaded
- Backed up my 180GB network share to six DLT tapes. Last backup was… 2002? (Thank you, RAID-5!)
Not too shabby, I guess. None of it is probably what I’m “supposed” to be working on, but progress is progress. The hardest obstacle of all is when you realize your main issue (lack of direction, aimlessness, a smidgen of depression, etc.) is completely without merit compared to the real, legitimate problems that others have — death, poverty, etc. — and that realization makes you feel bad about yourself, feeding the causality loop. Argh!
Chemical machines have another nice habit: They try to fix other chemical machines when they’re in need of repair. Honest thanks to everyone who wrote me, blogosphere or otherwise, with a kind word or helpful suggestion. I read them all, and they did help.
We keep moving forward, opening new doors, and doing new things, because we’re curious and curiosity keeps leading us down new paths. — Walt Disney
Keep moving forward.
Oldskooler Ramblings 
Tomer Gabel said
Good to hear. Ups and downs are natural for everyone (damn, but I should know!), the trick is learning to keep the downs in check :-)
T Zenios said
I think you have it wrong about your depression (which is what it sounds like to me) not being a legitimate problem compared to death, poverty etc in the world. If this has been going on for years (sounds like it), maybe you should see a doctor about it. Not like its going to kill you to get a 2nd opinion, depression is a major problem in the world and it can hit anybody (rich and poor), my brother is suffering from it now. There are some great drugs that will get you through the fog and into your old self again , if that is all you need.
. said
> (hint: A/V chunks now need to be sector-aligned)
…and pulling 4 sectors per frame?
Dumped the DOS FS for your own sector interleave?
Install another HD for this?
RP
Trixter said
Not anything quite as exotic; the change is being investigated because someone clued me in to the fact that, when you read (parts of) a file in DOS, DOS issues BIOS calls to read the data from disk into a TEMPORARY buffer, then COPIES that buffer to your buffer. I can’t believe I didn’t realize this before, but it makes perfect sense given that the target buffer in userland doesn’t need to be DMA page-aligned, but BIOS sector reads always do.
So the hope is that, if I parse FAT16 myself to get a list of sectors to read, and read them myself into memory, I can avoid the redundant copying and get some more throughput, making 60Hz possible without rebuffering. This will be offset, of course, by physically reading more data than necessary due to padding, and the annoying task of making sure every memory area I allocate is aligned to a DMA page boundary.
Trixter said
T Z: I appreciate the advice. If it lasts more than a week (so far I’m coming out of it), then I will definitely think about it since we have new insurance and my copay is only $25 a visit now :-)
. said
> So the hope is that, if I parse FAT16 myself to get a list of sectors to read, and read them myself into memory,
Will DOS reorder or move HD sectors of a file _after_ the file is written?
At least for floppies, DOS didn’t reorder or move file sectors.
If not, then you could trap the BIOS interrupts and record the BIOS sector list when the file is saved – since it’s location would be “permanent” until deleted.
RP
Trixter said
DOS won’t reorder a file’s data unless you specifically request a defrag. And although it is intriguing to trap BIOS sector reads to build the list, you’d also catch the FAT16 reads as well, and directory reads, so you’d get some junk in there. It’s not hard to parse FAT16; the only hard part is making sure that I allocate memory aligned to DMA page boundaries. Why is that hard? Because if you take too much time to do it, you’ve eaten up the gain we get from sector-aligned reads.
cgm said
Hopefully you will post all those soundclips from the Apple IIgs version of Silpheed I sent you back in oh….. 1998. Mind you since then I’ve gone to college and now have a full time job :-P. But at least I still got that IIgs and I’m
-cgm
Trixter said
I still have the clips, but obviously the Silpheed article has been put on indefinite hold :-) But never say never…