... Eudora¹

For a number of reasons, I'm trying to move my email to Evolution, but my perl script for automatically converting my many email filters to Evolution filters is not yet working. One of the reasons I want to leave Eudora is that I find it very disruptive that every single email filter insists on opening its folder when new mail for it arrives. (It is time-consuming to close all those windows.) As a result, I am moving my infrequently searched email folders to gmail, leaving only active ``work folders'' in Eudora. Thank you Google for gmail!

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... player)²

Winamp is the only sound player I have found for Windows that will play a sound file without moving a window to the foreground. This is important when clicking on PDF ``run:soundfile'' links in presentation. (I use PDF files generated by L^ATEX and powerdot instead of PowerPoint.)

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... file.³

I had checked ``ignore bad blocks'', but nothing worked after many hours of trying different modes of operation (booting from DVD, etc.). It just kept saying it couldn't copy the drive, giving an obscure error message regarding some cluster not being allocated [the restore was said to fail due to a ``bad sequence'']. According to a Web search, all one can do in this situation is keep rerunning scandisk and retrying. Beyond Compare did the obviously correct thing of simply logging the non-copied file and continuing. I simply renamed the non-copied file BAD-BLOCK at its source and restored the file from a back-up mirror. Norton Ghost ``gave up the ghost'' in total failure, and wasted a ton of my time in the process. This may be an example of a program making itself useless because it assumes the user can't do anything but click a few buttons in the UI, and at the same time they can't afford to provide technical support when things go wrong. In the Linux world, I always get useful answers from a Web search, and anything can be fixed, and usually already is, because the whole world has access to the source code.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...⁴

I copied my Eudora email from my Windows computer to the Eudora mail folder on the Mac, and none of it was recognized at all! I could probably fight through this by changing directory and file names until it looked right, but it's just a lot easier to continue using the Windows version in my VMware virtual machine. (Also, I've got 100GB to burn there, so I might as well use it for something.) Since Eudora is no longer supported by Qualcomm, I tried importing the Windows-format Eudora email folders into the Mac Mail app -- it started off reading happily, and my hopes were raised, but then the Mail app crashed (I sent a crash report to Apple -- the one and only instance of a crash I've ever encountered on the Mac Pro). Another nice thing about Windows virtual machines is that it is easy to keep old versions of Windows lying around if that's what it takes to keep Eudora running (of course all the current mail protocols such as POP could go away, but I think that will take a while).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... systems.⁵

Actually, I plan to use the ext2 file system, which is said to be compatible with ext3 (just change the filesystem type in the mount command from ext3 to ext2). I prefer ext2 because files can be undeleted more easily. Apparently, ext3 file deletions clear out more inode info making undelete infeasible (according to what I've read on the Web). I understand the main feature of ext3 over ext2 to be journaling, which avoids long fscks after a sudden power-down, etc. However, I think undelete is more valuable than journaling, unless someone can convince me otherwise.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... items]⁶

When I did this in 6/2005, it was necessary to edit and enable /etc/yum.repos.d/fedora-extras-devel.repo. However, this caused later updates (via ``yum update'') to fail due to missing dependency libc.so.6(GLIBC_2.4) which was called for by ginac, which in turn is used by Octave. Disabling fedoras-extra-devel allowed yum update to continue. Now (9/2005), it appears octave-forge can be updated without enabling fedoras-extra-devel. It's odd that all updates are aborted in yum when even one package hits a dependency problem.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... gv⁷

Fedora Extras include gv as of Oct. 27, 2005. Before that, it was necessary to deal with the source tarball: gv-3.5.8.tar.gz from http://wwwthep.physik.uni-mainz.de/~plass/gv/.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... reboot.⁸

On my system, it was also necessary to type

/usr/sbin/alsactl restore

after each reboot. I do this in my own sound initialization script, where I also say qjackctl -s. This is supposed to happen in the ALSA sound startup script.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...lm_sensors⁹

To set up lm_sensors, run sensors-detect. Then run /usr/bin/sensors to print out the current readings. You can also select ``System / Administration / Services / lm_sensors,'' and click ``Restart'' to update the numbers.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... Fan¹⁰

When I first built the system, I did not use a thermal conducting cream (not included or even mentioned in the directions) between the Pentium and its fan. As a result (I think), the computer experienced frequent lock-ups for months until I noticed this and added some cream. I would say I lost several days to trouble-shooting lock-ups, trying various work-arounds.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... life.¹¹

This could be said to be my fifth PC, and never did I receive a motherboard manual with any of those machines. Nowadays you can probably always get the motherboard manual on the Web, but that's only if you can figure out the manufacturer. My oldest PC, from ``PCW Computers'', according to a sticker on its case, seems to have no identifying marks whatsoever on its motherboard, and the PCW Computers websites I found seem to contain no documentation for what I have. My very first PC, an ancient AT&T 6300 ``DOS-Unix merge'' PC, had incredible amounts of documentation, reminiscent of the old mainframes. Maybe the motherboard was documented in there somewhere, but the computer was so old (pre 'AT' class -- perhaps an 'XT' style) that it could not be upgraded at all when I looked into that. The AT&T PC DOS-UNIX merge idea was a great concept, but the software execution was poor, in my opinion, and the performance was abysmal. With my OmniView switcher, and Samba, I have implemented a ``Windows-UNIX'' merge using two machines, and I think it's a great solution.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... wasn't).¹²

This may have started when I installed FastNet99, a tool which updates LMHOSTS to avoid DNS lookups. I uninstalled it, and I turned off using LMHOSTS in the Network Control Panel, but the problem persisted (for all browsers). There was probably something mucked up in the Windows Registry. The strangest behavior of all was that if I accessed www.redhat.com, I got a website talking about Red Hat Release 5.2, when the current release was 6.2! I deleted all my browser caching, etc., so I have no idea where the old site could possibly have been coming from, unless it was actually an old host still running at Red Hat Software serving up an obsolete website. Suffice it to say that I upgraded to W2K as soon as I possibly could (on Feb. 14, 2000).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.