Monday, September 21, 2009

Why I consider myself atheist now

Religious stories involving the supernatural seem to be mythological, and literally believing them seems brainwashed or even crazy. However, scientific understanding has such gaping holes that it's impossible to rule out that something like what religions talk about is part of reality. So, for a long time I considered myself agnostic.

Recently I realized that calling myself agnostic can imply that for all that mythological and crazy stuff, I'm saying "I don't know, it could be true". It doesn't feel right, and being unsure about delusions can lead to more delusions. To see the problem, imagine being unsure about something which is generally recognized as a delusion: the idea that there are satellites which read your thoughts and put thoughts in your head. This is kind of like the idea of God hearing your prayers and providing guidance.

I don't deny that religion can be a useful psychological tool. I don't deny that it can offer moral and ethical guidance. However, this doesn't mean that religion is fundamentally true, and I don't think that it's necessary for these purposes.

Thursday, September 03, 2009

In Vista, the "Previous Versions" tab requires administrative shares

No matter what I did, the "Previous Versions" tab for every file said "There are no previous versions available". System Restore was turned on, and both the GUI and vssadmin said that there were restore points. I was able to access old versions of files using syntax like "\\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy1\config.sys". I was even able to access old versions via the "Previous Versions" tab when I accessed files via shared folders (eg. "\\computername\public").

The solution was enabling administrative shares (such as C$). In the past I set the dword at "HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\services\LanmanServer\Parameters\AutoShareWks" to 0, which disabled them. I simply deleted that registry entry and restarted. (Note that even though Vista is a workstation OS, the registry entry is under LanmanServer.)

I wasted way too much time on this. I hate when something fails and there is no indication what went wrong. (I imagine something was trying to connect to C$ and it was failing, but that error was never reported.) This is actually my main complaint about Windows.

All this also applies to Windows 7.

Wednesday, September 02, 2009

JMicron JMB36X vs. Maxtor 92040U6

While backing up a partition to a Maxtor 92040U6 IDE (PATA) drive, Linux occasionally tells me stuff like:
ata9.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x6 frozen
ata9.00: cmd ca/00:00:ff:02:c9/00:00:00:00:00/e0 tag 0 dma 131072 out
res 40/00:03:00:00:08/00:00:00:00:00/b0 Emask 0x4 (timeout)
ata9:00: status: { DRDY }
ata9: soft resetting link
ata9:00: configured for UDMA/66
ata9:01: configured for UDMA/33
So, there a "write DMA" command timed out and the drive is ready. I'm not sure what else one can learn from that. I wonder what's at fault. The motherboard is Gigabyte GA-P35-DS3R. The PATA port is handled by a JMicron JMB36X chip, and the hard drive is a 10 year old Maxtor 92040U6 drive with MA540RR0 firmware. I've in the past I've had problems with the Memorex 3202 3292 PATA DVD+-RW drive which make me suspect the JMB36X. However, I also suspect the hard drive.

I'm curious because Windows 98 SE used to lock up when I was using the drive, and when I tried Windows 7, it would occasionally hang for a while with the hard drive light lit. In both cases there were no error messages, so I couldn't be sure what was happening.

If Windows 7 doesn't log error messages when there is an ATA timeout on a hard drive, then the people who wrote the ATA driver are idiots. Meanwhile, Firefox fails to properly display the Linux source file which produced the error message, and searching through a lot of text in Internet Explorer 8.0 is insanely slow, as if some evil genius invented a search algorithm that's even slower than a totally dumb algorithm. Sometimes, software sucks way too much. But hey, in this case at least it gives me something to do during the backup, and I don't have to routinely use PATA and deal with these errors.

Do all OSes still use a dumb LRU algorithm for caching?

I am backing up a partition from SystemRescueCd using dd and gzip. Occasionally I run ls to see the size of the output file. If I haven't used ls in a bit, the DVD drive spins up. So, it seems like Linux is using a dumb "least recently used" algorithm for caching, and the partition backup I/O is constantly removing ls from the cache. Do all operating systems still use such a dumb algorithm?