May 2007 Archives

When I did my Xubuntu install (the same is true for Ubuntu), I immediately started Firefox and went to my first Web page with embedded Flash.

And then it happened.

Firefox asked me if I wanted to download and install the Flash plugin. Why say no? So I said yes.

Problem: Dailynews.com -- the Web site from the newspaper I work for -- is covered in Flash Addsads, content come-ons, etc., and it was dragging my old systems to a crawl. On a modern 3 GHz system, it's OK to run a ton of flash. Not so on a 1 GHz converted thin client with questionable video and audio throughput.

But how to get rid of Flash? It's not so easy. Mozilla's help pages offer instructions on how to expunge Flash from Windows and Mac OS X, but nothing on getting rid of it in Linux. It's not an installed package, so Synaptic doesn't even know it's there. Apt-get also knows nothing. Why? Because it's hidden.

Finally, after a few sessions of Googling for an answer, this Adobe page provides the answer:

Removal instructions

Manual removal (for users who installed the plug-in via Install script):

Delete libflashplayer.so binary and flashplayer.xpt file in directory /home//.mozilla/plugins/

And for those with RPM package systems:

RPM removal:

As root, enter in terminal: # rpm -e flash-plugin Click Enter and follow prompts

On a related note, a commenter said I should try the Firefox extension Stop Autoplay. It didn't work.

But the other method does. I'm Flash-less -- and on my low-powered system, that's the way it's going to stay.

Update: After a half-hour, I can say definitively that my system is running much, much better. Flash is one of those things that's great if you've got the power but a CPU-hogging nightmare if you don't.

Bottom line, choosing applications that match your hardware (and needs) is the way to go. No matter the window manager (or "desktop environment"), some apps just need a lot of juice.

Consider the MIL-spec Panasonic Toughbook 30:

Out in the field or down in the dirt, the durable Panasonic Toughbook 30 is built to take a beating. MIL-STD-810F-tested for ruggedness, this brawny workhorse is encased in magnesium alloy, with durability designed into every seal, hinge and connector. Plus, as the industry's fastest fully-rugged mobile PC, it's built for lightening-quick processing and wireless connectivity. Communicate in real time from remote areas, access databases online and run sophisticated software applications even in the harshest environments.

Or get a really old one for $169.99.

Put together this primo system, The "Money No Object" desktop PC.

But for those with less money, ZDNet's Hardware 2.0 blogger Adrian Kingsley-Hughes also spec'd out a $500 budget PC.

Scientific Linux has quite a pedigree, according to DesktopLinux.com. Developed by Fermilab, CERN and other heavy-science types, the distro is based on Red Hat Enterprise Linux recompiled from source (which must be available according to the GPL license under which Linux is released) and is said to be mostly compatible with RHEL.

I've wanted to try CentOS, the other free "remix" of RHEL, but I would rather try a live CD first. Scientific Linux has it, and I'm downloading the ISO now. I'll report later on how well it works.

By selling at retail, Dell will reach a different kind of consumer than it currently does with its direct-only channel.

While my current Dell box at the office (Optiplex GX520 with 3 GHz Pentium 4 and 512 MB RAM) was part of a big corporate order that numbered in the hundreds of units, it isn't my first Dell.

Way back in the early '90s, pre-Web, we bought a Dell at Price Club, the warehouse store now known as Costco. It was a 386sx 25 MHz model with something like 4 MB of RAM and an 80 MB drive. It shipped with Windows 3.1, which was barely usable at the time. We mostly ran DOS (I think it was at 5.5 or 6), and that box got us on a bunch of local BBSes, plus the GEnie, Prodigy and AOL online services. Never mind that this PC couldn't run much of everything today, but in its day, it was well-built and ultra-reliable. It gave me a good impression of Dell.

Today I'd be more inclined to assemble my own system, if only to facilitate easy upgrades of the various components, from motherboard to optical drives to video, sound and networking cards. While most of us don't do all the upgrading we say we're going to do, it's nice to have the option. I still plan to replace the motherboard, drives and even power supply in the now-10-year-old This Old PC, if only to a) prove to myself that it can be done, and b) from an environmental and "simplicity" standpoint to save the case, keeping it from going into a landfill and eliminating the need for a new one. And I'm cheap.

But back to Dell. Selling through a mass-market retailer and offering customizable systems online are two very different businesses. To compete with HP/Compaq, Dell needs to be out there, side by side with its competitors.

A smarter bet for Dell would be its own mall-located, branded stores, like the Apple Store, and unlike the current Dell mall kiosks in that they'd have actual store space and actual inventory that customers could purchase and carry home. It didn't work for Gateway, but it could work for Dell (or for HP).

Maroon 5 broke an iTunes record by selling more than 101,000 digital copies of new album "It Won't Be Soon Before Long" on the Apple music service during its first week of release, with many lured to pre-order the album before its May 22 release date by the promise of two extra tracks and the ability to get ahead in the virtual Ticketmaster line for concert tickets.

There were 50,000 preorders of the 13-track album, ArsTechnica reported.

Add to that 243,500 digital sales of the single "Makes Me Wonder" from iTunes and other download sites.

It almost brings the glamour back to the music business.

After a few weeks of wrangling with e-mail at the console, I now have a working setup that allows me to send and receive mail from two accounts, one personal, the other for work, on my Debian Etch box.

(For those who haven't been following that closely, my hardware is a Maxspeed Maxterm thin client with hard drive and CD-R drive connected to the single IDE header. The motherboard is by ECS, it's sort-of mini-ITX sized, the processor is a VIA C3 Samuel at 1 GHz with VIA's PC133 chipset, and memory is maxed out at 256 MB.)

I'm using mutt as the mail client and msmtp as the SMTP client for outgoing mail. And while real geeks use vi, I've become partial to nano of late, and that's been my default editor for e-mail ... and everything else.

My personal account is with freelinuxemail.com, which looks like, acts like and is provided by fastmail.fm, but which includes free use of their SMTP server. Linspire sponsors the service, and I think they're picking up the tab. (Thanks, Linspire!) The service is IMAP only, which I prefer anyway, because I need to access my mail from a variety of places, programs and PCs.

However, my work account is on a POP server, and I really don't want to pull all the mail off the server and onto a single PC -- and I could never get the company's outgoing SMTP server to work. Instead, I'm using the SMTP server from my broadband provider, DSL Extreme (who's tech support department declined to help me with this project, since, as they told me, "You know we don't support Linux." I figured it out anyway, and I will keep trying to figure out my company's SMTP server, even though everything is working just fine like it is.

The main sites that helped me get all of this working were the main mutt site (mutt.org), the msmtp site, an article from O'Reilly's MacDevCenter.com titled How to Use mutt, FastMail, and Mail.app Together on Your Mac, even though I'm not using a Mac, and My First Mutt, and all are required reading before you embark on this yourself. That and about a hundred Google searches for sample configuration files and advice specific to these programs and my mail providers.

And before I continue, yes, I tried Pine, but I could never really get it to work, especially for the outgoing mail.

The key here is that mutt is configured to read e-mail from my IMAP account. To read my POP mail, I hit lower-case "c" to navigate to other folders, and instead of an IMAP or local folder, I go to my POP site and then log in. The e-mail comes up from my POP account, and I can read, delete and reply to messages, but what I don't delete remains on the POP server, accessible by my Web-based client, Outlook on my Windows box and Thunderbird (or SeaMonkey) on my Linux boxes. (I got the lower-case "c" trick from My First Mutt. Works like buttah.

Here's how I do it:

(type lower case letter) c
pop://pop.mypopaddress.com (note: use pops:// if it's a secure server)
(at prompt for login, enter my e-mail address)
(at prompt for password, enter my password, then the mail appears)

As I said, I haven't yet gotten msmtp to work with my company's SMTP server yet, but I did get it to work with DSL Extreme's (substituting my work e-mail address in the "From" category).

I can choose to send mail from three different accounts, freelinuxemail.com, my work account and dslextreme.com with key bindings I created with help from msmtp's manual (the section on using msmtp with mutt).

Things I'm going to work on in the near future are shortcuts for frequently-used e-mail addresses and some kind of system for organizing saved mail in my /home directory.

Here are my config files:

.muttrc

set use_from=yes
set editor=nano
set imap_user=myaddress@freelinuxemail.com
set imap_pass=mypassword
set spoolfile=imaps://mail.messagingengine.com/INBOX
set folder=imaps://mail.messagingengine.com/INBOX
set record=+"Sent Items"
set postponed=+"Drafts"
set certificate_file=~mutt_certificates
set envelope_from=yes
set sendmail="/usr/bin/msmtp"
set realname="Steven Rosenberg"
macro generic "<esc>1" ":set
from=myaddress@freelinuxemail.com"
macro generic "<esc>2" ":set
from=myaddress@dailynews.com"

.msmtprc (make sure this is set to read/write ONLY by you. The command is: $ chmod go= .msmtprc )

account linux
host mail.messagingengine.com
from myaddress@freelinuxemail.com
auth on
tls on
user myaddress@freelinuxemail.com
password mypassword

account dsl
host smtp.dslextreme.com
from myaddress@dailynews.com
#for "user" in DSL Extreme, the @ and everything after it are not needed
user mydslextremelogin
password mypassword
#auth login is the only authentication that worked (I went through every one that msmtp offers before I tried this one)
auth login

It's probably not a great idea to have passwords in the config files, but it does speed things up, and once I pull the CD-ROM drive, it'll be pretty secure, especially because it doesn't boot from USB either. That's assuming anybody around here a) wants to read my terribly exciting e-mail and b) knows what Linux even is, let alone how to hack into it.

One of the things I'm going to try in the future is to remove the passwords and maybe even the user names and see how the setup works then. For now, it really speeds things up to have user names and passwords entered automatically.

Another must: getting the e-mails to display with the newest on top (they're now on the bottom).

It's taken a couple of weeks to get to this point, and I'd like to thank all the readers of this blog who wrote comments and told me about their console-based e-mail setups.

While on this project, I also set up all of my mail accounts in Thunderbird, but where's the fun in that?

Again, the greatest thing about e-mail at the Linux console is the blinding speed with which e-mail can be read, replied to and deleted.

I've been keeping mutt running on tty1, with a text-based Web browser -- either lynx or elinks -- on tty2. Both browser have their good points -- elinks has better CSS support, but sometimes looking at a page with no CSS in lynx is better. I'm sure I can turn off CSS in elinks, but it's just as easy to start the other app.
I should probably start using a third virtual console with Midnight Commander running, but I haven't been using mc that much lately. Still, if you're using the Linux command line and not using at least two or three virtual consoles, switching between them with alt-F1/F2/F3/etc., you're not really living.

As far as the e-mail setup goes, there are probably better ways to do it, and like all highly configurable applications at the console, it has the potential to grow, change and morph depending on my needs and skills.

While I cited many of the Web pages that helped me, I'd also like to acknowledge the books that have taught me much about the Linux command line. My main book, my bible if you will, is "The Linux Cookbook, Second Edition" by Michael Stutz (here's the free online version of the first edition, but it's well worth buying and having at hand). Stutz's book is so well-written and complete that it's an enjoyable read as well as a valuable reference. Even though it was published in 2004, it remains extremely useful, since the command line isn't subject to the faddishness of the GUI ... not that a new edition wouldn't be welcome.

I've also been using "Linux in a Nutshell," by Ellen Siever, Stephen Figgins and Aaron Webber, another great O'Reilly book, and I've just gotten the other Linux cookbook, titled "Linux Cookbook," by Carla Schroder, also from O'Reilly ... also published in 2004 (and still plenty useful, although again, a new edition would be great.

Since this is Day 27 of A Month on the Command Line, here are the things I haven't gotten done yet:

-- Printing: I tried the trusty Apsfilter, which has worked for me on occasion in Damn Small Linux, but I couldn't get it to find a network printer. I will try CUPS before this is all done, and I think that will work. But why didn't Debian ask me about printers during the install?

-- vi/Vim: I never got that good at running vi or Vim. I read somewhere that even when you're running vi in Debian, it's really Vim. I'll have to look into that. But while I type fast and have to write stuff really fast, I didn't crave the ease with which I could move around the page in vi. I grew attached very quickly to nano, which, with it's F3 keystroke to save and F9 to delete a line, quickly became my favorite command-line app.

-- Images without X: I didn't do much with Imagemagick. Since I couldn't actually view an image in that suite without X, I never got to the point of sizing a photo at the command line. Zgv was already installed in Debian, but it died while trying to display a photo. I checked the config file, and it should've worked but didn't.

-- Blogging from the command line: I never did try out the program that allows direct posting to Movable Type blogs. It's something I'd still like to try, but once I had e-mail working, the will to continue wasn't quite as strong. I did post from elinks to the blog, but it wasn't easy (and for some reason the posts didn't go live without "intervention" from a GUI browser).

But I did a lot of Web browsing with lynx and elinks -- and both are extremely fast and enjoyable ways to read text on Web pages. I highly recommend trying them at the command line.

What separates Linux from Windows and Mac is the fact that Linux from the command line is made up of apps that are up to date (they have to be, given that they're running production servers at very high levels in the enterprise) and extremely configurable. The fact that you can run the Linux command line on hardware that's extremely old and underpowered, yet get all the modern security features that Linux has to offer, is a powerful incentive to use it, especially for older hardware. I've blogged extensively on how to keep older Macs running, on which browsers work, which e-mail programs support today's servers (answers are few and Netscape Mail) and how a decade-old, unsupported OS can function in the modern computing world. But with Linux, all that fades away. Window managers go from full-featured and heavy to extremely light ... and there's always the shell. Linux developers care about users -- and not about rendering old hardware obsolete so you'll buy whatever's new just so you'll have apps that run.

And while the books I've mentioned above have been very, very helpful, I'd love to see a command-line book that walked the newbie through all the apps and setup of a Linux system, with specific instructions on how to set up e-mail, network connections, with even more information on all the current command-line apps. Kind of like what the "Cookbooks" do, but with more specific information on going from nothing to a full command line system. I guess there's not quite the market that there is for Ubuntu/SUSE/Fedora books, or for books aimed at sysadmins, but it would be nice to see. The fact that you can easily set up even Ubuntu as a command-line-only system -- and not just as a server -- makes this something that I think more people would do, if they only knew how.

Back in the days before the Thin Puppy's CF card died, I was running my Maxspeed Maxterm thin client with flash memory instead of a conventional hard drive. And since Puppy Linux takes great care NOT to write to flash very often, the media is supposed to last virtually forever. Why mine died is a mystery, but it wasn't due to wear (more likely I killed it with static electricity).

Now that I'm running Puppy 2.16 (new to me this week!), I've been thinking about going back to flash for this thin-client box -- I'm booting from CD and also have a regular-sized 14 GB hard drive connected outside the box (yes, I truly am thinking outside the box -- or my PC is).

But the conventional wisdom is that for "normal" operating systems that don't use RAM disks, you'll kill flash quickly with the constant writes required by the OS.

But today on Low End Mac, that question didn't come up for these guys who are running their Mac laptops from flash memory. That link was to the letters about this original article, which, in turn, refers to this article about doing it with a Powerbook 1400 (one of which I have ... but which is too frustrating at this point to even contemplate using for my everyday computing for reasons that have nothing to do with flash). This final article -- filled with woe about flash cards that didn't work with the 1400 -- does address longevity of the flash media, saying it should work for "years and years."

I'm going to try it again (maybe even with Debian), but I'm also going to back up all my data ...

And again, if you want to boot from flash but are nervous, give the new Puppy 2.16 a try. I'm in my first full day of use, but so far all is going very, very well.

... it's just a subset of the overall Dell forum, but it has a Linux logo, complete with Tux (that's it above). And here's the Dell Linux Engineering Web Site, where the project's Wiki lives.

Today's debut of Dell PCs with Linux preinstalled threatened to be eclipsed by another Dell bombshell -- the Round Rock, Texas, PC giant will supplement it's direct-to-you sales method with a heaping helping of Middle American retail through Wal-Mart.

Dell told the Associated Press that June 10 will mark the debut of two Dimension E521 desktop models at 3,000 Wal-Mart stores in the U.S., Canada and Puerto Rico.

The deal "represents our first step" into global retail, Dell spokesman Dwayne Cox said, according to the AP report.

No prices were announced, but the cheapest Dimension E521 goes out the door for $359 at Dell.com.

It doesn't look like Dell will offer Linux on its Wal-Mart boxes, and a check of Wal-Mart's Web site shows that the company -- today anyway -- no longer offers the Linspire-based Linux desktops it had been quietly selling for some time.

Hmmmm ... maybe it's Dell and Wal-Mart's way of giving Microsoft an open-mouthed kiss with extra tongue.

But back to Dell's Linux offerings -- prices for a single laptop and two desktops preloaded with Ubuntu 7.04 begin at $599 and are slightly lower than when equipped with Windows Vista. Oh ... and Dell's Ubuntu won't be able to play .mp3's, DVDs, Flash and a whole bunch of other stuff without the user going through the usual hoops, a kiss of death for the Dell-Ubuntu experiment, according to Adrian Kingsley-Hughes of ZDNet's Hardware 2.0 blog.

And finding Dell's Linux-equipped PCs at Dell.com isn't easy. What's more apparent is Dell begging its customers to stay with come-ons for Windows XP, which they really, really want you to know they're still selling.

There are no links on the Dell home page, and the first thing you see is the ubiquitous "Dell recommends Windows Vista Home Premium." Of course it does.

But in the dropdown menus for Desktops and Notebooks, there are links for "Open-Source PCs." The headline is "Ubuntu has arrived by popular demand." And there's a handy link, right there in the middle, for Windows users who somehow navigated to the page as if by evil sorcery:

The main thing to note is that when you choose open source you don’t get a Windows® operating system. If you’re here by mistake and you are looking for a Dell PC with Windows, please use the following link.

But for those who do want an open-source box, you can get one from Dell with Ubuntu ... or FreeDOS. Yep, nobody mentioned in this whole Ubuntu-Dell lovefest that the Texas computer giant will ship you a box with FreeDOS on it. What the hell? I'm sure there's a reason for it ...

Here's how Ubuntu describes its FreeDOS offerings:

FreeDOS is a completely stripped down PC operating system for experts and people interested in working in a DOS environment. The main things that set FreeDOS apart from Windows and Ubuntu is that it doesn’t have a graphical interface (i.e., it’s typing only, no mouse) and it only supports DOS applications, such as classic DOS games, business software written for DOS and embedded DOS systems, such as retail cash registers. Generally, users who want FreeDOS know what it is and what they want to do with it. Other users should look to our Ubuntu and Windows platforms.

Embedded DOS systems? If they're out there, I guess it makes sense ... but again, what the hell?

Anyway, when you click through on the Ubuntu link, here's what you get: A mid-level desktop, low- to -mid-level laptop and then a more upscale desktop

Not bad at all. As an exercise for this entry, I "built" a primo XPS 410 system with as many upgrades as made sense (i.e. nothing stupid), no extra warranties or support, and it came in at a hefty $1,964 with Intel® Core™2 Duo Processor E6700 (4MB L2 Cache,2.66GHz,1066 FSB), 4GB Dual Channel DDR2 SDRAM at 667MHz, 500GB Serial ATA 3Gb/s Hard Drive (7200RPM) w/DataBurst Cache™, Dual Drives: 48x Combo + 16x DVD+/-RW w/ dbl layer write capable, 19 inch Ultrasharp™ SP1908FP Digital Flat Panel, 256MB nVidia Geforce 7300LE TurboCache, Dell A525 30 Watt 2.1 Stereo Speakers with Subwoofer, Dell USB Enhanced Multimedia Keyboard and 13 in 1 Media Card Reader.

So you CAN spend a ton on a Linux box at Dell.

But here's the kicker for Canonical. Although this disclaimer appears --" Dell provides hardware support only. Software support is available through Canonical and Linux Community" -- there are options available with each system for "Starter Support" ($65), "Basic Support" ($125) or "Standard Support" ($275). That's above and beyond any extra money you want to give Dell for "In-Home Service, Parts and Labor."

Who's providing this non-free support? Certainly the "Linux Community" means Ubuntuforums.org ... but the others must be Canonical, which I hope is ready for what could be an onslaught. It's an experiment, all right, in more ways than three, and it will be very exciting (and I hope not disheartening) to see how it plays out in the months ahead.

My task today was a simple one: Pull four Quicktime videos from an SD card and burn them on a CD.

I booted Xubuntu Feisty, and after a few minutes found xfburn, which came standard with 7.10. I had a couple of Thunar windows open, that was it, but xfburn pretty much crashed on launch. I went into the process manager, killed it, then closed my Thunar windows and restarted xfburn.

(The hardware: VIA C3 Samuel 1GHz processor, 256 MB RAM, TDK 32x CD-RW drive, and yes, I have a swap file.)

Selecting the files and burning the CD was easy and went smoothly. I had my CD.

Then I went back into Thunar to delete the .mov files from the SD chip. Xubuntu wouldn't let me do it. I can't remember the exact error message, but it was something about there not being enough room in the trash. In any event, I couldn't complete the task.

So I rebooted into Puppy 2.14 from the live CD. One good thing about Puppy, when you have a Linux swap file on your hard drive, the system finds it and makes use of it. Instead of a sketchy 48 or so MB of "free RAM," I have 496 MB at my disposal. All that means is that Puppy won't start dying if I download a lot of files.

I quickly mount the SD card with Puppy's MUT (Mounting Utility Tool), open the directory in the Rox Filer, select the four videos and delete them.

And ... while I'm at it, Puppy handles the multiple Flash windows full of continuous advertising crap on Dailynews.com with SeaMonkey better than Xubuntu does it with Firefox. Sure, SeaMonkey is pegging my CPU at 80 to 90 percent, but the system is still running. I couldn't open a second SeaMonkey window on my second desktop, but I was able to open two additional tabs (one with Lxer, the other with BoingBoing), and then start Abiword on the other desktop.

As an aside, I'm not alone in the opinion that Dailynews.com's continuously running Flash content is obnoxious. I've had the discussion with the powers that be -- at least the powers that know I exist, and those Flash ads are paying the bills, so nothing's changing. My solution: read Dailynews.com in the Dillo browser, which doesn't do Flash and leaves the annoying, continuously downloading videos that I'm not watching.

Flash abuse -- it's real, and it's a pain in the ass.

...

Back to today's computing challenge; once again, Puppy is my best friend. I can't wait to try Puppy 2.16 and the MiniPup I read about on Desktoplinux.com.

If it's a laptop you want, bring money -- lots of it.

According to my new favorite Microsoft site, a $1,500 laptop won't get 'er done.

Joe Wilcox of Microsoft-Watch says you need to bring $1,999 to the bar to get adequate graphics performance for Windows Vista:

I surveyed computer after computer, even several desktop PCs, with Windows Experience Index ratings falling between 2.1 and 3.3. In every case, graphics hardware dragged down the scores. Again, I contend graphics is the wrong place to shave margins.

In my Vista testing, there is noticeable performance benefits on computers with 256MB dedicated graphics compared to those with 64MB or 128MB discreet graphics and shared memory for the remainder. More importantly, at some time in the future there will be more applications tapping into Aero and Windows Presentation Foundation—and that's when customers are going to feel the graphics pinch.

While pricey, Best Buy had one notebook that rated above 4.0—the Pavilion v9260nr, for $1,999 at the store in Bowie, Md. It was the first Vista computer I've seen for sale where the graphics accelerator wasn't the lowest-rated component. The hard drive brought the score down to 4.5 (the NVIDIA GeForce Go 7600 graphics rated a 4.6). The HP notebook, with 17-inch display, 2GB SDRM, 200GB hard drive and TV tuner, came with Vista Ultimate 64-bit. It's one really honking portable.

Consumers shouldn't have to spend two grand to get the ultimate Vista machine. But they should expect—and get—more from a $1,500 or $1,600 notebook. That Sony model, VAIO VGN-AR320E, looks loaded for bear with Best Buy advertised graphics of 335MB. But how many consumers or small business owners are going to know to look—as I did—and see that only 64MB is dedicated graphics?

Wow. And I'm complaining that $400 laptops are too pricey.

For another interesting perspective from Joe Wilcox, check out Is Vista One Step Ahead? In this article, Wilcox looks at Bill Gates view of Microsoft being "ahead" of the curve, and how that generally means software that won't run on current hardware. He cites Gates' 2003 unveiling of Longhorn (which would become Vista), which would need the following:

4-6GHz processor

2GB+ memory

1TB hard drive

Graphics processor 3X today's performance

1GB Ethernet, 54Mbps wireless networking

And while Wilcox suggests that hardware makers step up and give Vista users what they need to really experience Microsoft's new OS (and give application developers the incentive to produce software that goes beyond what XP can do), I wonder if its a case of the tail wagging the dog ... albeit profitably for Microsoft and the PC industry throughout the years.

In Gates' (and Wilcox's) defense, most of these hardware benchmarks are here -- especially with the latest Intel and AMD CPUs, the unveiling of the first consumer-oriented TB hard drive, the proliferation of 1 GB Ethernet hardware (even if corporate America's existing cabling won't support it) and the rise of 802.11n.

Here's Wilcox's assessment:

In my experience, most Vista problems are a result of insufficient computing power or incompatible software applications or drivers. I resolved the former by moving to the lovely Lenovo ThinkPad T60p. Vista still is a dog getting out of the gate, meaning slow bootup or wakeup, but she's a gallant race horse once moving.

Are you ready to saddle up?

I'm not, but I know that in five years, it'll all be old hat.

With the help of My First Mutt, I figured out how to import POP mail into mutt:

To browse a POP3 mailbox, just hit c to change mailboxes. Then, instead of typing the name of a local mailbox, you can enter the location of your POP3 server. This is much like typing the URL for a page in a web-browser:

pop://username@mail.example.com/

Mutt will ask you for your password and you can browse, delete, edit and save the messages as normal. If your POP3 mail account supports SSL encryption, then you should use pops:// instead of pop:// to enable it.

And it works for outgoing mail, too! I guess it's a lot to type, and there's probably a way to handle it through .muttrc, but getting ANYTHING to work is a giant step for me, and I'll just bask in the glow until tomorrow.

Again, I don't know how this works, but I can seemingly read the mail in mutt while it remains on the POP server, which is exactly how I want it for the time being. A total accident -- but freakin' brilliant.

I will post configuration files at some point in the near future. (Trust me, there can never be too many configuration files on the Web, even from one as inexperienced as myself).

To recap the programs I'm using, mutt brings in the mail from an IMAP server (and now from a separate POP server), and that mail is sent out via msmtp. I had some success with fetchmail and pine for the POP account, but I still can't get pine to send any mail. At some point I'll probably need to throw procmail into the mix.

The IMAP service I'm using is freelinuxemail.com (from the company that also offers fastmail.fm, except that freelinuxemail.com includes free use of an outgoing SMTP server. And for those who care about it, the identical Web interfaces for freelinuxemail.com and fastmail.fm are extremely quick to navigate and also very light on resources (with the option of turning CSS stylesheets off). And fastmail.fm was the first free e-mail service I've seen that promotes the use of IMAP, where the mail remains on the server and can be accessed by any number of mail clients from multiple locations -- instead of POP, which generally downloads all messages to the client computer, leaving you out of luck if you want to change mail programs or locations. For me, I'm reading e-mail in three different places on up to five different computers, so IMAP fits me a whole lot better.

For the past few years, though, I've been using Web-based mail services (chiefly Yahoo! Mail and our company's Web mail interface (which, ironically, is like having IMAP service, although for external mail clients we only have a POP server). But for the speed alone, it's nice to get away from Web interfaces and back to standalone mail clients.

Again, and for the fifth or so time, IT SHOULDN'T BE THIS HARD.

I've had some success with my IMAP mail from freelinuxemail.com. Mutt and msmtp are handling things very well. I'd love to beef up my .muttrc file to sort things (maybe I need procmail) and I'd like to figure out how to create additional folders on the IMAP server, but for now just getting e-mail and being able to send it is good enough.

As I said in a comment on the last A Month at the Command Line post, I tried pine again.

Well, first I tried to piggyback a POP account on my current mutt/msmtp setup, but that wasn't working at all. So I tried to configure pine as a POP mail client. After I tried the settings in about a half-dozen Web tutorials, I was able to download all my POP mail (which unfortunately meant that those e-mails are no longer on the POP server and are no longer accessible by the company's Web-mail interface ... and no, there doesn't seem to be IMAP capability).

I've got to tell you, in this day and age of always-on Internet, IMAP is the way to go -- you can read and write e-mail from a variety of locations and computers and not have mail stored on one (unless you want to save them locally, that is). It makes for a lot more flexibility.

But back to pine. I used fetchmail to get the mail into the box and put it wherever mail is supposed to go (into my Linux user account) so mutt can find it. I'm not rock solid on the particulars, but it works. Sending the mail -- that's where I have a problem. In mutt, I needed msmtp (or esmtp) to send mail. Pine doesn't seem to need an external program to send mail, but I just can't get the server to accept my login and password. There might be authentication, SSL or TLS that I need to address. I've got to look into it again and then figure out how to do it in pine.

But I'd rather not use pine at all and do everything through mutt ...

I did spend a little time in Xubuntu 7.04 today ... during which Firefox crashed ... then IE 6 crashed under Wine, so I rebooted into Debian's console and began using lynx and elinks. I don't know if it's frames or java, but elinks seems to handle things better -- at least at dailynews.com. It makes reading text on the Web a whole lot quicker -- and might be worth using even in a GUI.

Anyway, while reading the Daily News story by Eugene Tong about a lady who's been to 177 straight "Tonight Show" tapings, I clicked on an HTML "mailto" link for Eugene and was immediately in mutt, ready to compose an e-mail. It all happened so fast and worked so well.

Yep, I'm in Day 2 of having mutt and msmtp doing my e-mail, and it's been really great.

And while I'm on the topic of e-mail at the Linux command line, I got a nice comment from console user kotnick on how he handles e-mail:

Well, wouldn't this be easier (this is my own setup):

1) fetchmail to get mail from all accounts which provide pop3/imap (I get mails from 6 different accounts)

2) procmail to sort it out and filter out spam with spamassasin

3) mutt to read mail (although I use claws-mail)

4) sendmail to send mail (you can use it on every smtp server that provides relaying, as gmail does, as my local isp does)

4a) if you find sendmail too cryptic, you can try postfix

And from my previous e-mail entry, reader jason does this:

I use Fetchmail, Mutt, Procmail and Postfix to handle my mail, and have also changed my mind several times over the last few years over which anti-spam tool to use.

Here's a brief summary of how it works:

Fetchmail retrieves incoming messages from my accounts on IMAP servers, and delivers them via SMTP (port 25) to Postfix. Fetchmail configuration involves creating a ~/.fetchmailrc file.

Postfix queues the messages and invokes Procmail to deliver them . mailbox_command = procmail -a "$EXTENSION" in /etc/postfix/main.cf

If there is a ~/.procmailrc file, Procmail applies the rules in it to filter the incoming messages into various folders. This implies that you don't need a ~/.procmailrc file for this setup to work, but by having one you can filter out spam and duplicate messages, filter various mailing lists into different folders, etc.

Mutt accesses the local mail folders (/var/mail/username as the primary inbox, and folders under ~/mail) to read the mail. Of course, mail is delivered and filtered entirely in the background.

For outgoing messages, you can configure Postfix with the relayhost command, as in relayhost=mail.isp.domain to deliver outbound e-mail via your ISP's SMTP server. Alternatively, you can just let Postfix deliver mail to the destination directly, without any additional configuration required. Some spam filtering services may filter out your messages if you do this. In my case, I can't tell the difference between the two alternatives, so I just let Postfix handle outgoing messages itself (without relaying via my ISP).

The point to emphasize here is that all of this isn't too difficult to set up (a certain investment in reading manual pages and setting up configuration files is needed, especially for Procmail and to a lesser extent Fetchmail), but once set up, it works really well - year after year - with only incremental improvements. That's why some of us are so addicted to this way of working. I can't remember how long it took to set up the first time, but given how well it has worked ever since, it was well worth it.

Stephen J. Vaughn-Nichols of Linux-Watch lets this one out at the end of his story on Microsoft's allegation of patent infringement in open-source software (get ready fo a big chunk of quoted material, all of which is necessary to understand the point):

(Begin quote)

So, while Microsoft's latest claims may sound terrible to the layman, any attorney worth his or her salt will know that these are old and basically bogus statements. So, why is Microsoft trotting them back out again?

I believe it serves two purposes. One is to spread more FUD about Linux and open-source. For the first time, a major computer vendor, Dell, has committed to a consumer desktop Linux. More states, like California, are considering making laws that require the use of the ODF (Open Document Format). From Microsoft's point of view, it was time to get people worried about open-source again.

The other purpose is to try to get leverage against the upcoming GNU GPLv3 (General Public License, version 3). The latest draft includes patent language that will make it much harder to make patent deals, such as the November 2006 Microsoft and Novell partnership.

Why should Microsoft care? Because, Microsoft, by distributing SLES (SUSE Linux Enterprise Server) certificates to customers such as Dell, as part of the Novell/Microsoft partnership, may have just placed any IP they might or might not have in Linux, under the GPL.

No, that's not just open-source fanboy talk. Prominent open-source lawyers, like Eben Moglen, the executive director of the Software Freedom Law Center, believe that by distributing the SLES certificates, Microsoft has become a Linux distributor, and therefore subject to the GPL.

For Microsoft, being subject to the GPL in any way, shape, or form would be a nightmare scenario. If they can get some leverage in their fight to get away from the GPL by getting people frightened of open source, they will.

(End quote)

It's pretty clear that Microsoft will do whatever it needs to do in order to remain at the top of the software heap. But there are cracks in the armor. When MS was woefully behind on the Internet, it succeeded in crushing Netscape and promoting its own IE browser. And while Netscape did, indeed, fade away, so did Microsoft's attention paid to its own browser, leaving room for Firefox (and with it Google search) to take away a large chunk of that market that can't be reclaimed, even with IE 7.

At another level, the threat of lawsuits will only embolden Google to further develop its Google Apps stable of online-based productivity programs -- and possibly to enter the operating system market with a product that, like Linux, doesn't need Windows to function.

And while some -- or even many -- makers of open-source software might be turned back by Microsoft's legal charge, there will also be some who will fight any potential litigation. And actually getting into court is something Microsoft really does not want. Remember the last time Bill Gates had to do a depo?

Microsoft is ramping up its fight against open-source software -- including the Linux operating system and other free applications -- with threats to sue over what the software giant says are over 200 violations of Microsoft patents.

The reason the story gained so much traction over the weekend was this Fortune article posted Monday on CNN Money.

The conflict pits Microsoft and its dogged CEO, Steve Ballmer, against the "free world" -- people who believe software is pure knowledge. The leader of that faction is Richard Matthew Stallman, a computer visionary with the look and the intransigence of an Old Testament prophet.

Caught in the middle are big corporate Linux users like Wal-Mart, AIG and Goldman Sachs. Free-worlders say that if Microsoft prevails, the whole quirky ecosystem that produced Linux and other free and open-source software (FOSS) will be undermined.

Microsoft counters that it is a matter of principle. "We live in a world where we honor, and support the honoring of, intellectual property," says Ballmer in an interview. FOSS patrons are going to have to "play by the same rules as the rest of the business," he insists. "What's fair is fair."

Here is how Brad Smith, Microsoft senior vice president and general counsel, sees it:

... he does break down the total number allegedly violated -- 235 -- into categories. He says that the Linux kernel -- the deepest layer of the free operating system, which interacts most directly with the computer hardware -- violates 42 Microsoft patents. The Linux graphical user interfaces -- essentially, the way design elements like menus and toolbars are set up -- run afoul of another 65, he claims. The Open Office suite of programs, which is analogous to Microsoft Office, infringes 45 more. E-mail programs infringe 15, while other assorted FOSS programs allegedly transgress 68.

The story isn't all Microsoft. There's a lot of information about Stallman, who created the GNU Manifesto and the software movement around it (GNU meaning "GNU's not UNIX").

The Fortune story covers the history -- and current state -- of Microsoft's negotiations and deals (or no deal) with Linux end-users, as well as Red Hat and Novell (who together control the commercial Linux server market) , and it tries to explain why MS made a deal with Novell that results in the Redmond giant paying the much-smaller Linux packager, instead of the other way around.

There's a lot of commentary and reportage out there on this issue (thanks to Lxer for most of these links):

The Fortune article: Microsoft takes on the free world

Linux-Watch: Microsoft reignites its war on Linux

eWeek: Microsoft Claims Open-Source Technology Violates 235 of Its Patents

O'ReillyNet: Looks like Microsoft is ramping up for Son of SCO

Linux Journal: Meeting Microsoft's Patent Threat

Tux500 blog: Intentions Clarified - Battle Lines Drawn

itWire: Microsoft to sue Sun over OpenOffice.org?

CNET: Report: Microsoft says open source violates 235 patents

Microsoft-Watch: Microsoft's Open Letter to Open Source

After days and days of being able to receive IMAP mail but not send it with mutt, I finally cracked the problem. What took me a little while to understand -- that mutt needed a separate SMTP client to send Internet e-mail -- took a lot longer to actually get working.

First things first: The IMAP mail service I'm using is freelinuxemail.com, which I believe is sponsored by Linspire. The mail service is run by fastmail.fm, but what makes freelinuxemail.com different is that you get to use their outgoing server for free; it costs extra with the plain fastmail.fm.

At the freelinuxemail.com/fastmail.fm site, there are configuration instructions for pine. (I tried them, they didn't work, deleted pine, days later reinstalled pine, tried again, still didn't work). But nothing for mutt. Some of those who replied to my previous post are using up to four separate applications to handle their e-mail, and I was getting mighty dizzy from the whole thing. I tinkered with my configuration files for days, but still nothing).

I thought the problem was esmtp, the SMTP client I was using.

While pursuing other solutions to the outgoing-mail problem, I looked around for other SMTP clients and found msmtp.

I immediately knew what I was doing wrong.

From the instructions for using msmtp with mutt (bolding for emphasis added by me):

2. Configure msmtp:

Create the file .msmtprc in your home directory, with no more
permissions than user read/write (0600, -rw-------).

I had the wrong permissions on .esmtprc all along. How did I miss that?

To change the permissions on the .msmtprc configuration file, I turned to Michael Stutz's "The Linux Cookbook, 2nd Edition," Page 169, for the command:

$ chmod go= .msmtprc

Now .msmtprc was set to private (read/write only by me), and I was in business.

I'm pretty sure resetting the permissions on .esmtprc would have yielded the same result, but for now, I'm sticking with msmtp, and I'll try to figure out how to use this combination (mutt and msmtp) to access a POP account, and, hopefully, multiple IMAP and POP accounts simultaneously.

By the way, while it did lead me astray at times, the instructions I used for initial configuration of mutt and esmtp were from O'Reilly's MacDevCenter (although the path to esmtp was different, I didn't need to match anything up with OS X's Mail.app since this is Linux, and I guess on OS X the read-write permissions of .esmtprc were not an issue).

Another good mutt resource on the Web is My First Mutt, which I'm just beginning to explore.

I will post my config files for mutt, msmtp and esmtp as soon as I get the time.

And thanks to all who helped and encouraged me to continue with e-mail on the command line.

When was the last time you installed Windows NT 4.0? If your answer is "never," I believe you. If you've done it countless times, do I have your sympathy? I need it.

My most recent major Windows upgrade (chronicled on my This Old PC blog) was taking a Win 98se box to Windows 2000. For those who think Windows has some kind of compatibility advantage over Linux, let me recount how in Windows 98 I didn't have a prayer of getting my cheap Airlink 101 wireless card to work, USB was spotty, and the thing was painfully slow to boot and to run.

Once I did the upgrade to Windows 2000 and added all the service packs and upgrades to that old 333 MHz PC, I was able to get wireless to work, but it didn't last long. Soon enough, the PC stopped recognizing the card (even though Puppy Linux always recognized it but could never make it work ... but that's another story).

So in the interests of comparing Windows performance with Linux on the same hardware -- my test machine, a Maxspeed Maxterm thin client (1 GHz VIA Samuel processor, 256 MB RAM, mini-ITX motherboard (which I now know is called a "micro-mini-ITX, measuring 17 x 19 centimeters) with single IDE header, single RAM socket, built-in networking, graphics and sound) -- I will endeavor to install a Windows system, download enough open-source applications to make it work and see how Windows does in my various desktop tests.

My Windows 2000 Professional disc is labeled as an upgrade, so I get out the only working Windows full-install disc I have -- Windows NT Workstation 4.0, circa 1996. I have some trouble formatting the 14.4 GB IDE drive -- the installer keeps saying the drive or partition is too big. So I keep making it smaller until I can continue. I end up with a 4 GB partition for the C: drive. I have the choice of a FAT or NTFS filesystem, and I opt for NTFS. The partition is initially formatted as FAT, but I decide to continue on, and it is converted to NTFS a little later in the install.

The process goes pretty well, except for my network controller, which isn't detected -- and I have no idea which driver to choose. Once I do select an alternate, I have to go on the install disc and look for it ... nothing automatic or helpful there.

So I play around with Windows NT 4.0 for a few minutes. Among its stellar attributes are Internet Explorer for NT version 2.0 (and remember, I had no networking, so I couldn't really try it out). And then there are the usual suspects -- WordPad, NotePad, the calculator -- all the exciting things a bare Windows system is known for.

My only chance is the upgrade to Windows 2000. In my unschooled opinion, Windows really only became useful in the world of wireless networking and USB with Windows 2000, and I've had pretty good luck with it on my Pentium II box, my wireless problem notwithstanding. And I've said many times how much I like Windows XP (for which I don't have a disc, or I would install that).

And since I could only create a 4 GB partition for Windows, there's plenty of room to dual-boot with the Linux of my choice.

I get a couple "spoolsv.exe has generated errors and will be closed by Windows. You will need to restart the program. An error log is being created" messages, but the "successfully installed" message does appear, and the system reboots. However, upon logging in, the message reappeared.

But I am able to continue.

The good news: Windows 2000 finds my network adapter.

The not-so-bad, but not-great: It doesn't remember my static IP settings, so I re-enter them.

The nice thing about Windows 2000, as opposed to Windows 98, is that when you make a change to the network parameters, there's no need to reboot.

The graphics look like total crap. For some reason, the highest resolution I can select is 800 by 600 with 16 colors (not 16-bit, but 16 COLORS). I haven't seen graphics this bad since ... we had Windows 98 (not even the "se" edition) here at the Daily News. Ah ... fond memories of crap hardware. And I'm back in the wonderful world of Internet Explorer 5. I'm looking at the Dailynews.com home page, which is lousy with Flash, and now the browser is asking me whether or not I want to install the Adobe Flash Player. I accept, even though I've got to get IE 6 and Firefox on this box. Not to mention Open Office, the GIMP (for which I'll need to also get the GTK+ runtime libraries), AbiWord, Irfanview, EditPad and so much more.

During the Flash install, IE crashes. Just like old times. Now I can't run IE at all -- it crashes upon launch every time. Thanks MS and Adobe! I reboot and regroup.

IE is still spotty, I can't figure out how to remove Flash, but I've managed to keep it running. Now I'm going to do a Windows Update, which should take plenty of time. Right now it's hanging -- but that's typical with Windows Update, something I do have experience with.

Windows Update hangs, IE crashes ... I manage to find the Macromedia folder and move it to My Documents. I still can't delete it -- one of the files is "in use." Remember files you can't delete? It's one of the many charms of Windows.

My "Windows Update" is hanging ... literally ... but now that I can actually use IE, I'm downloading all the open-source software I'll need for my test. Compared to the average Linux distribution's package management, this is brutal ... a hanging Web page. I've encountered this in the past, and what I do is start the Windows Update process and then walk away from the computer for as many hours as it takes. I'm not all that confident at the moment.

AbiWord, the GIMP and the GTK+ runtime libraries, and Firefox all download in minutes. Open Office still has half an hour to go (much of this is due to a slow mirror). I begin to download the free Avast antivirus package and discover something: There's a version for Linux. I'm no expert on viruses or antivirus programs, and I've heard that viruses are rare in Linux, even rarer for Mac's OS X ... but it's nice to know that antivirus protection is available.

Back to the Windows install ... now the system isn't recognizing my USB flash drive ... so much for transferring files that way. (Later it comes back in Windows 2000.)

After the Windows Update goes nowhere, I get an idea. Microsoft, in its infinite wisdom, probably changed the Web address. I go to Microsoft.com and navigate to the Windows 2000 home page. Now Windows Update starts working. You'd think MS would offer a "redirect," but clearly they're not paying much attention to the pre-XP crowd.

First I have to install Service Pack 4. Do I have to get IE 6 separately? I think so.

Application aside: I'm writing this entry in AbiWord for Windows. I love an application that's small, loads fast and does what I want. Give me the option of "typographical" (aka "smart") quotes, and I'll go to the mat for you as best word-processor ever.

Installing the GIMP: The GTK+ runtime environment is a ZIP file ... and with Windows, you can't unzip without installing an unzipping program. So I have to download PKZIP.

PKZIP for Windows 9 asks for a license key. I opt for the "30 day trial." After a few months of Linux, I forgot that Windows doesn't even allow you to unzip a file without paying extra.

I get GIMP installed, and now I'm installing Open Office. My 4 GB of disk space is rapidly filling. I have space left to create a new partition ... but can I make this partition bigger? Working with Windows disks is far from my strong point. And no, I don't have Norton. It's great that users of Windows need to keep paying for essential utilities.

1.8 GB of my 3.9 GB C: drive is full now.

Once I have Open Office installed, I can close it and relaunch in under 5 seconds in this Windows 2000 environment. I know that it only launches so quickly because most of it is preloaded in RAM, whether I'm running it or not -- but it remains impressive, and if OO was that important to me, I'd appreciate the quick loading and tolerate the drain on system resources. That said, I'd like the preloading to be user-selectable, both in Windows and Linux (where OO doesn't sit in RAM waiting to be launched, I believe).

Even in Windows 2000, my graphics remain at 800 x 600 at 16 colors, and everything looks terrible? On this monitor and with the built-in VGA in the thin client, I can do 1024 x 768 in Linux at 16-bit color, not just 16 colors.

AbiWord launches as quickly as it does in Linux.

I try to start GIMP. It won't run. Too bad I dumped the install files. The much-smaller Irfanview does launch. The one thing I don't like about Irfanview is that when I shrink a photo to, say, 200 pixels wide, and then add a 1-pixel black border, the photo is now 202 pixels wide. The GIMP (and most other programs) cut in on the photo with the border and retain the original dimensions. It's something I can live with. But I wonder why the GIMP won't run. I did install the GTK+ environment.

After another install, GIMP still won't run. So I go down a version to 2.0.5. During the install of that version, I am told my GTK+ is too old ... so I install THAT again, and finally get GIMP 2.0.5 to run. Now my graphics still look like hell -- I can't really see what any images really look like -- it all looks like pixilated pop art -- but I do have the GIMP. And it loads in 20 seconds. I had load times between 20 and 30 seconds with GIMP in Debian 4.0, and 60 seconds in Xubuntu on this same computer. But as far as selecting and editing photos, Win 2000 is useless with the current state of the display. I think I need a Super VGA driver. Hunting down video drivers is not something I've had to deal with in a long time.

Now that GIMP 2.0.5 did install, I retry 2.2, and it goes through. The new GIMP starts in a little over 20 seconds -- same as the older version.

Firefox starts in about 10 seconds -- pretty good. It's a lot slower on my 333 MHz system.

The next day, I decided to try a clean install from the Windows 2000 disc. The disc detected my previous Windows install but still let me start over -- and this time I could make the disc partitions of my choosing. So I made a 6 GB C: drive and a 2 GB D: drive, with with the rest left for future Linux partitions.

On this second install, I hope my VGA/Super VGA problem will be taken care of. Hopes dashed, I still have 800 x 600 with 16 colors.

Now I have the base Windows 2000 installed, so it's time to install Service Pack 4 and then do all the updates -- 55 of them.

And while these updates were downloading and now installing, I decided to lift the cover on the thin client's box and once again try to identify the motherboard. I knew it wasn't a VIA, even though that's the maker of the CPU and chipset. I see the letters EVEM written in the middle, and do a Google search for that. Turns out it is an EVEm motherboard from ECS, with a PDF manual here and a page of BIOS and driver updates here.

The manual confirmed what I already knew from Maxspeed -- the memory maxes out at 256 MB. I wish it weren't so, because 512 MB would make things so much easier. The manual's somewhat detailed info on BIOS settings will be helpful. And there is a video driver, which just might be very helpful in Windows, bringing my resolution past 800 x 600 (no driver needed for any Linux I've tried, of course).

After the Windows updates install, I download PKZIP again (all the downloads from yesterday's install died with it), then download the ECS video driver for Windows 2000. I do the install, reboot and have 1024 x 768 video in 16-bit color. The 15-inch monitor looks better in 800 x 600, but most Web pages I'm using won't fit comfortably in that space, so I'll keep it cranked up for now. At least the colors look good.

I didn't have to wipe the drive and do the Win 2K reinstall today, but I was able to create a bigger C: drive than with the NT 4.0 base, and I had to see if Windows 2000 as a base would take care of my driver problem (it didn't). And in the meanwhile, I learned exactly what kind of motherboard I have and now possess a 40-page PDF manual describing it in some detail. But the big coup in learning where the motherboard came from was finding the video driver that enabled the display to look normal.

So I download the PDF manual on the thin client (I found it on my Dell box while the updates were rolling). ... and I can't open it. Windows 2000 doesn't come with a PDF reader. ... Gotta go to Adobe and download one. I've gotten used to not needing to do that -- once again, every Linux distro I've tried has built-in PDF support, all for reading PDFs and most for creating them, too.

After the Windows 2000 install is done, there is plenty of space left. I have a Ubuntu 6.06 LTS disc, so I install that -- and all goes perfectly. Ubuntu installs GRUB, detects my Windows partition and allows for dual-booting, all without heavy geeking of any kind. Why 6.06 and not 7.04? Since Windows 2000 is old and "stable," I figured I'd put the most stable Ubuntu around on it.

Overall the Windows 2000 install process went OK. It was no harder than installing the many Linux distros I've tried (the ones that worked, anyway) and certainly no easier. And for the hardware I have, it is generally detected better in Linux than by Windows.

The biggest difference between Windows and Linux is that Microsoft's OS comes to the desktop with very bare bones. There's IE5, Wordpad, Notepad ... and not much else. But the average Linux distribution contains dozens of utilities, full office suites, powerful graphics programs, multiple text editors,mail and FTP clients and much more. And all of that software is FOSS (free, open-source software) not shareware that compels you to pay if you want to keep using the program beyond a short trial period.

Since I already had the Windows 2000 disc, it cost me nothing to install it. But if I didn't have it, I'd be hard pressed to fork over a couple hundred bucks for the privilege of using Windows. And the greatest thing about Linux, with its many distros, desktops and applications is that you can try and never have to buy. It's a powerful thing, indeed.

As if we didn't already know, Google has its sights set on Microsoft's cash machine -- the applications market:

(Via Red Herring)

Google today officially confirmed what many analysts have been saying for months: the Mountain View, California-based search leader is going after the software market in direct competition with Microsoft.

Reflecting the company’s explosive growth, Google’s new tagline will be “Search, Ads and Apps,” said CEO Eric Schmidt in a speech at its annual shareholders’ meeting.

Last year Google launched a package of online applications such as browser-based word processing and spreadsheet software that look much like Microsoft’s bread and butter Word and Excel programs.

And the Googlers did talk about the free, open-source OS that is nibbling on Microsoft's other moneymaker:

In a question and answer session with shareholders, one woman asked Mr. Schmidt and Google co-founder Larry Page how they would avoid becoming known as “the new Microsoft, and not in flattering terms.”

Mr. Page added, “It’s natural to be concerned about it… We’re not the same kind of company” as others out there, Mr. Page said. “Our engineers all run Linux. It’s free. You can hack on it. It’s deep in our DNA”

DBR commented on my Blogging without a GUI post and told me about mtsend.py - A Command Line Tool for Movable Type.

Here's a description of what it does:

mtsend.py is a command line tool that utilise Movable Type’s XML-RPC interface. It allows its users to edit/post/view/list post entries on a Movable Type site. It takes input from the standard input, and sends output to the stdout, just like all the other well-behaved command line applications. It uses a file format similar to Movable Type’s import export format.

Updated: It appears that mtsend.py also works with WordPress, as WordPress has implemented a compatible XML-RPC interface. In fact I am using this same script everyday since I switched from MT to WP.

...

With a command line tool that can work with pipes and standard IO, it makes posting much easier for me. I can now write up a blog entry in Vi, save the text file, run it through aspell check, and then direct it to mtsend.py to have it posted to my blog site. All under your favourite Unix shell!

And on this page -- and included with the source code -- is clear information on how to write the configuration file for mtsend.py.

As you may be able to tell from the .py part of the name, the app is written in Python. There's this note: "You need to install 'xmlrpclib' if you are using Python 2.1."

And as an aside, Scott Yang, who wrote the app, uses the w3m text Web browser. I'll have to try that one.

Here's Scott's reasoning behind the development of mtsend.py:

I found myself spending most of my day in the front of text console/command prompt. My typical Windows desktop consists 2 of VIM windows and 2 SSH session to other Linux boxes, and half the website I browse everyday is in the excellent text-mode browser w3m. I don’t like to fire up Mozilla and open up the Movable Type site just to post an blog entry. Moreover, I have very bad spelling problem, and posting through the web interface does not allow me to check my spellings. Of course I can edit my post in a separate editor, run that external file through a spell checker, and then paste the content into the Movable Type’s edit entry page. However, since Movable Type has already exported its interface via XML-RPC, why don’t I just write a simple script to post that text file after I’ve run it through the spell checker? The result becomes mtsend.py.

If I can figure it out, It'll be amazing.

There's no end to the tips and tricks in Linux.com's CLI Magic series of articles.

Among the cool ones:

Using youtube-dl at the command line to download YouTube videos.

An overview of GNU find.

The watch command.

There are over 100 of these CLI stories in there -- thanks, Linux.com!

In keeping with its frantic pace of development and improvement, today the Puppy Linux team has released the beta of Puppy 2.16. The good news, for me anyway, is the "minimalist look" (see above), and the "Save" icon on the desktop for USB users. Also welcome are and improved Pmount drive-mounter -- an alternative to the MUT (Mounting Utility Tool). Here are the full release notes.

Also of note:

The GUI SFS Boot Manager -- to easily load and manage SFS add-on modules (for things such as Open Office)

A change in filesystems from Unionfs to Aufs, which allows for "true flushing of RAM" on boot, meaning less of a chance of running out of RAM (it's not pretty when it happens, trust me).

The (experimental, at this stage) ability to encrypt your pup_save file (while Puppy is already super-secure due to the way your files are saved in pup_save, this will theoretically keep people from accessing your files if they are sitting at your box and happen to boot Puppy).

This YouTube video from Vermont librarian Jessamyn West chronicles her installation of Ubuntu Linux on a donated PC -- on which Linux is a necessity since she didn't have a copy of Windows for the box. She also blogs about it:

My install process went like this: download and burn the Ubuntu disk image to a CD. Turn on the computer with the Ubuntu CD in the CD drive. The computer boots Ubuntu from the CD. You have the option to run it this way or install it to the hard drive. You have the option to install it on a partition (and keep Windows also) or just erase the drive and install Ubuntu as the only operating system. You restart the machine and it runs Ubuntu and it Just Works.

Remember all that controversy about MySpace blocking videos from Photobucket because they had commercials in them that provided no revenue to the News Corp.-owned social-networking site?

Well now MySpace has "solved" the problem ... by buying Photobucket. Valleywag broke the news, and offers this analysis:

What does the deal mean for other startups which have piggybacked on Myspace? On the one hand, it's heartening. For a venture so dependent upon it, Myspace paid a hefty multiple. But Photobucket is the largest of the "widget" makers, ventures which depend on a share of the real estate of larger sites, rather than drawing visitors to their own properties . If even it could not develop an independent advertising business, the prospects of other widget makers are dim.

Photobucket's deal should be cause for celebration for Alex Welch and his partners, and for Trinity Ventures, which invested $10.5m in the company last year. But, for other widget makers, it marks something of a ceiling on their likely exit.

The story was confirmed by TechCrunch, and given a price: $250 million.

From TechCrunch, here are the financials for Photobucket:

Photobucket generated $6.3 million in revenue last year and planned on hitting $25 million or more this year. They have 40 million registered users and add another 85,000 per day. ... They’ve raised $15 million over two rounds of financing.

You'd think that e-mail and command-line Linux would be like chocolate-chip cookies and milk.

Not so. Getting e-mail from a server and into this Linux box has proven to be more than a little frustrating. Clearly I need a comprehensive reference to get this right.

I don't want to set up my own e-mail server, although at this point I'm thinking that just might be easier than the path I've been down thus far.

After extensive Googling, I decided to use Mutt as my Internet mail client, and after more extensive Googling, I figured out how to set up a rudimentary .muttrc file to get mail into this computer. So far, so good.

I wanted to use IMAP as my protocol (the account I'm using supports both IMAP and POP), and I eventually found out that Mutt will NOT send IMAP mail on its own. I don't exactly know the terminology (MTA ... MUT ...), but I decided, after more Googling, to use esmtp. I haven't gotten that to work yet. In fact, I'm not sure of the path to esmtp, which I need to put in my .muttrc file to make the whole thing work. That's how much I need to learn about Linux in general and Debian in specific.

Bottom line, I need to find a reference that really walks me through configuring multiple POP and IMAP e-mail accounts on a command-line Linux box. Any suggestions?

Blogging from the command line without a GUI -- and no e-mail gateway -- can be done ... almost.

Using the command-line, text-only Elinks browser to post blog entries -- or to complete Web forms in general -- is a lesson in trial, error and frustration. But disciplined use of keystrokes -- and a little dumb luck -- allows these browsers to post to Movable Type blogs such as this one.

The biggest obstacle to blogging in Movable Type with Elinks is that the Save button doesn't work. That's a big problem.

To get around the dead Save button, just go to the Authored On box and enter an appropriate time for your entry to have as its time stamp. Then press Enter.

When the browser asks whether or not you want to post the form, respond in the affirmative.

At that point, you may need to log in again. If so, do it, and your entry will be saved on the site.

But ...

For some reason, the entries from Elinks are marked "published," but they don't appear on the live site until, at some future point, other work on the blog is done. AND ... the picking of a catetory for the post doesn't seem to be sticking. The latter I can live with, the former remains puzzling.

Note: Testing was done in Elinks. This doesn't work in Lynx.

What if you have one PC, one monitor, two users (two keyboards and two mice)?

According to this report from Techtonic, via Discovery News, you can run Windows with the screen split down the middle:

New technology from Microsoft Research India in Bangalore could end the waiting game in offices with limited computers. Researchers are developing software that splits a computer screen in two halves, each side with its own operating system, desktop, applications, cursor and keyboard.

The technology could help small businesses and schools in developing countries reduce their computer costs and could even work in homes where more than one family member needs access. "At the most basic level, we are allowing two users to work completely independently on the same machine, sharing both the processor and monitor," said Udai Singh Pawar, assistant researcher and project leader.

(The image above is a screen shot from the Microsoft Research India screen-splitting project.)

Here's another very good vi cheatsheet, "Wow, I didn't know you could do that in vi," by Patricia Bender.

It was a bit of a mystery. Once I closed Vim, I couldn't figure out where it was saving my files in Windows XP. It wouldn't be a problem if I could remember the name of each and every file I created, but since I can't remember my own name half the time, it is a big deal, especially since this isn't a command-line system on which I can just type ls at the prompt. And it gets worse: For some reason my "search file" icon doesn't work either -- never has. (Thanks, again, Windows!)

Vim help was no help, but I did a "save as" in Vim, and it clued me in:

By default, Vim saves it files in the User directory, which is in the Documents and Settings folder under the C: drive.

So the path is C:/Documents and Settings/user/filename

Another mystery solved.

The vi coffee mug -- it's been suggested, but instead I downloaded and printed a couple of reference sheets for vi, the best of which is Donald J. Bindner's single-page Vi Reference Card. The best thing about it is that it keeps everything on one page (and doesn't bleed off the page like the Debian Reference Card.)

I also have Using the VI Editor (here's the PDF version), which is also pretty good but on two pages. I'll have to tape them together, because I need to see EVERYTHING at once.

I probably should get the O'Reilly book on vi, but that would be overkill at this point.

You might ask whether or not I'm coding the HTML by hand in these entries. Yes, I am.

The hardest thing for me in vi and Vim is going from insert mode to command mode via the esc key. It's so far away.

While vi is good about using the home row of keys to navigate while in command mode, going into that mode in the first place requires me to take my left-hand fingers off the keyboard to hit esc. Getting back to insert mode with a or i is extremely easy, but esc remains so far away.

I wonder if any enterprising geek type ever came up with an esc footswitch to make getting into command mode possible without the fingers leaving the home row?

I guess hitting the esc key is no worse than using the arrow keys or the mouse, so I'm goint to stop bitching for now and get these Vim-created files on the blog.

Much as I'd like to spend the entire day working on my Linux box, the reality isthat I have to spend a lot of time doing my real job, which entails running the Unisys Hermes publishing software on my company-supplied Dell PC.

So how do I improve my vi skills in the world of Windows?

The answer is Vim -- the expanded version of vi that's already standard in a number of Linux distributions, including Debian. There are also versions of Vim that run in Windows and on Mac OS.

The Windows version of Vim downloads and installs like any other program. There are command-line options, but since I'm locked out of my command line (accessible by a password I don't possess), I'm just using it as a window in XP).

Aside: I can't run a command line in Windows (it hasn't been helpful because I can't run .BAT files by clicking them), but I can install programs, just not with the Windows Add-Del program utility. If I download and install a program, I have to go to the directory and use the app's own "de-installer" to get rid of it. Thanks, Windows!

With Windows Vim, it's easy enough to use the menus if I get stuck, and most of the time the relevant vi/VIM command-mode equivalent is listed. It's a great learning tool, and a pretty good editor besides. I'm really enjoying having multiple documents open on a single screen instead of each being in a separate window (the norm for most Windows editors, including EditPad.

And if you really want to know Vim, here's 500 PDF pages of info.

While I did get my flash drive mounted without difficulty, I spent considerable time trying to post to this Movable Type-powered blog with Lynx and Elinks. Both let me do everything, except actually SAVING the entry. I've had similar trouble with the Dillo browser, and that's a full-fledged X application, so there's something squirrely about the Save button in Movabble type.

But for general browsing of Web sites, including this blog, the text-only browsers are more than adequate -- and quite sophisticated in what they can do.

I did get a few entries through as tests, and I think it was Elinks that produced them. I hit something, somewhere that enabled the posts to save, but I couldn't replicate my actions for the real blog post, which sits on the hard drive on my Debian box, awaiting the beginning of the week ahead.

Thanks to all who left comments on A month on the command line, Day 1.

Mounting my USB flash drive from the command line was easy.

First I made a directory:

$ mkdir flash

Then I became the "super user" (entering the root password when prompted):

$ su

Then I mounted the drive to it:

# mount /dev/sda /home/steven/flash

Then exiting as su:

# exit

Now a check of what's there:

$ ls flash

And I see the directories and files on the flash drive. Thanks again, everybody.

I'm sure all real geeks know about this already, but there are vi-instruction-emblazoned mugs, shirts, bibs and aprons available via Cafe Press.

I'm sure I'll get my vi sea legs eventually -- if I don't abandon all hope before then.

I'm tired of debating the relative merits of Linux's various desktop
environments and distributions. Tired of reading about them, tired of
writing about them.

So today an experiment begins:

My month at the command line.

I'm going back. To the $ prompt. No GNOME. No KDE. Not even Xfce, IceWM or
Fluxbox. No X at all. Just the console.

I could use one of my standard Linux systems and either not start X, or kill X and drop down to the command line. But I wanted a command-line system, one built that way.

Being partial to Puppy Linux, I looked into OneBone Puppy -- a special mix of the live-CD distro
without a GUI. And it had Elinks, a text-only browser. I burned a CD and booted it.

I still like the idea of running a live CD and keeping the filesystem in RAM. But more than that, I needed a system that would allow me to install the maximum number of packages -- and do it easily -- so I could get as much done on it as possible.

I went Debian.

I grabbed a hard drive from my stack (yes, I have a stack -- what's it to you?), hooked it up to the Maxspeed Maxterm thin client, put the Debian 4.0 Net Install disc into the CD drive and began. This particular hard drive already had Xubuntu 7.04 on it, so I shrunk that partition, made a new 2 GB space for my command-line-only Debian syste and continued forward. When the time came to select the type of Debian install I wanted, I chose a mail server. I could've left all the boxes blank, but I figure I might as well put something on there. I'm not hurting for disk space in this setup, and I just might want to learn a little about what makes a mail server tick. But before running there's walking and crawling.

The reason I chose Debian? Apt-get. I know apt-get works with the Debian-derived Ubuntu. Damn Small Linux can also be installed on the hard drive and made to function as a Debian system. But Debian has worked so well as a traditional desktop system for me, I wanted to keep the core in my command-line system as well.

Once the Debian installer finished, it was time to apt-get as many things as I thought I needed. I started with the Elinks text-based Web browser and then added a bunch of stuff I read about in Michael Stutz's
"The Linux Cookbook."

Here's what I put on so far (with descriptions, because I can't remember
what half of it is):

alsa-utils -- sound configuration
cdtool -- CD player (doesn't work because I don't have the CD audio hooked up to the motherboard)
ee -- allegedly easy text editor
elinks -- text-based Web browser
emacs -- big daddy of text editors
fetchmail -- gets POP mail
fetchmailconf -- configures fetchmail (but doesn't work without X)
imagemagick -- edits images from the command line (this I've got to see to believe
joe -- easier text editor
lynx -- a more popular text-based Web browser
mc -- Midnight Commander file manager
mpg321 -- console-based mp3 player
sox -- sound program
splay -- another console-based mp3 player
wv -- MS Word file translator, now part of AbiWord and KWord
zgv -- Console-based image viewer

And in case you're wondering, I'm writing this in joe because I'm so rusty
in vi that I can barely get through two lines of text before the box starts
beeping uncontrollably and my co-workers begin complaining about the noise. Joe's not bad, but it's no vi. In any case, I have plenty of text editors to choose from. I didn't even mention vim, the juiced-up vi that's already included in the base Debian.

This isn't the shock to the system (my brain, not the computing system) that it might have been. I've been here before.

When I was an undergraduate at UC Santa Cruz in the mid- to late '80s, IBM-compatible PCs and Macs were expensive, and the sole Mac lab was clear across a large, wooded campus. And who wanted to print on Imagewriters, even in 1986?

Unlike most universities, where timeshared computer use was meant for and
restricted to those in computer science, engineering and the like, UCSC had a policy and a mission to
bring the rest of the world into the universe of networked computing.

The thin, photocopied instruction manual, "Unix for Luddites," by Scott Brookie,
cost about $2 at the campus' Bay Tree Bookstore, and it gave away the secrets of
Unix.

First secret: Any student could get an account on a machine named Unix
B. Yes, there was a Unix A, and a C through probably G. Unix B was the only
one of those systems not connected to the Internet -- so mail only went to others in the system, and there was no access to Usenet (and nobody knew what it was, so we didn't miss it).

But Unix B did accommodate about 25 or so users at a time via terminals in 24-hour labs at every one of the campus' eight geographically dispersed colleges.

The terminals we hacked away on were mostly of the adm3a and VT100 variety -- and even then the adm3a was considered a dinosaur. Some labs had highly prized Wyse terminals that
weren't so fuzzy and had great-feeling keyboards -- and those went fast
during prime paper-writing season. And did I mention the giant line printers kicking out that giant green and blue paper? Those were for drafts only -- UCSC and Unix B had something else with which to entice the average essay-writing student:

A real laser printer.

The main point of "Unix for Luddites," and of the UCSC Unix push, was to get
regular students to use Unix to write papers (in vi), format them (in
nroff), and then pick up the printouts -- MADE ON AN HONEST-TO-GOD LASER PRINTER
-- at a computer center buried in a little-traveled area of the campus.

A laser printer! At the time, Apple had just begun selling its first 300-dpi LaserWriter for $6,995. Having access to a laser printer was quite a lure; that and having a teacher's assistant in my Shakespeare class berate me for sloppy typing (I had a MANUAL typewriter and a lot of WiteOut, for crying out loud) was enough to plunge me into the mysterious world of Unix.

And there were a lot of people doing the same thing. In the days before the
Web, before chat rooms, before enough people had PCs and modems and could
connect to PC-based computer bulletin boards, typing the who command on Unix B let you know the logins of the other two dozen users on the system (and it was often so well-used,
you had to wait to log on until somebody else logged off). You could send
mail and use the talk program -- the latter being as much of a time-soak back then as instant-messaging is today.

But back to the command line. I want to see just how much I can do without a
GUI. I don't think I can blog directly from here -- those Web interfaces are
just too complex for a command-line, text-based browser. I guess I should've
put X on here so I could use Firefox, but that would be defeating the
purpose. First of all, I need to figure out how I can get POP or IMAP mail at the command line. Then I'll look into getting an e-mail bridge to Movable Type like the one that makes Blogger.com so useful.

So far, my biggest discovery (thanks, "Linux Cookbook"!) is Imagemagick,
which can supposedly resize photos from the command line. Can't
wait to try that one. And if I can somehow get this entry out of the Debian
box and into one with a modern browser, you'll see it on the Daily News
Click blog. (Note: I couldn't figure out how to mount my USB flash drive, so I rebooted into Xubuntu and grabbed this entry from the Debian partition.)

Meanwhile, I've got to make a decision: vi or emacs? I need to buckle down and learn one of them. I can't use joe forever.

Even geeks need to laugh.

Steven J. Vaughan-Nichols of Desktoplinux.com has all the reaction to Dell's announcement that it will load Ubuntu Linux onto select desktop and laptop models. It's a good way to find out the mood of those behind the other Linux distros, including Red Hat, Novell, Xandros, Linspire plus a bunch of analysts. It's important reading, for this week anyway.

Here's the part I keyed in on:

Looking ahead, (Raven Zachary, the 451 Group's open-source research director) thinks, "We'll see some wait-and-see from the other hardware vendors: 'Let's see how Dell does with this deal.' Also, you could see Red Hat, Novell, and others swarm in to the other hardware vendors to avoid an Ubuntu sweep."

So can we expect a lot of similar deals in the coming weeks/months? Who's going to bring HP/Compaq, Lenovo and the other hardware makers to the mat?

Caitlyn Martin of O'Reilly's LinuxDevCenter blog, today writes about the Ten Tec RX-320D, a shortwave radio receiver controlled via your PC. The default software runs on Windows, but there's also a Linux package that Martin herself helps maintain for both Ubuntu and Vector Linux:

Why a PC radio? First, I can maintain a nice database with the rx320 software that lets me go directly to a large variety of broadcasts with frequencies and times, all recorded based on my experience of what I can receive clearly at my location. I just right click on the virtual radio on my screen, click frequency database in the menu, choose my station listing, and pick what I want. There’s no meaningful limit on the number of stations I can record information about. Much like a spreadsheet I can sort my stations anyway that’s convenient at the moment with a click or two on a column header.

Part of my geekiness is that I've done a bit of shortwave listening over the years (I sold my Sony ICF-2010 radio on eBay last year -- the money was too good), and I'm also a licensed, yet inactive amateur radio operator -- call letters KC6FYL -- unless my ticket has expired. I can tell you that at $349, this Ten-Tec radio it is quite a bargain, even though it looks like a plain black box -- the power is in your PC, and it's better to have a PC controlling your radio than a bunch of knobs, switches and onboard computer components that can fail.

And TenTec is an American company that makes this stuff in the Great Smoky Mountains

Also on the O'Reilly Linux blog: Julia Kemp talks briefly about sound problems in Debian. Yep, I've had them too, and one of her suggestions has worked very well for me: running alsamixer at the command line and PUMPING UP THE VOLUME. Works every time.

I've been trying to get my Palm Tungsten E to work with Xubuntu -- trying being the operative word.

I can sync via GNOME-Pilot, and I see all my Palm files in a directory, but what do I do after that? I downloaded J-Pilot, and that is not working at all -- the Palm syncs, but it has nothing to do with J-Pilot, and none of the handheld's data is flowing into that program. And every time I reboot, the whole thing goes to hell again.

Somewhere along the way I had to download pilot-link -- if I needed it that bad, it should be one of the dependencies for the other apps when they're installed through Synaptic.

To get the syncing working in the first place, I followed the advice in the Ubuntu forums to do:

$ sudo modprobe visor

(visor? are they kidding?)

and to add a line to fstab:

sudo mousepad /etc/fstab

(substitute YOUR text editor for mousepad, then add the following:)

/proc/bus/usb/ /proc/bus/usb/ usbfs none 0 0

The Web sites for all of these applications are out of date and extremely poor. Yeah, I know it's free software, but it's pretty much useless. If you have to totally geek it to get things working, there should be detailed, comprehensive information somewhere. It's nice to have 18,000 Ubuntu packages, but how many of those apps have real value? That's another question.

On a somewhat related note, I got a nice comment from a guy who works on HPLIP -- the HP printing utility for Linux. I'm still impressed that the program works so well -- and that a big hardware company is investing time and money to make things work better in Linux. Hell, Palm isn't investing in its own brand and on the platforms it already "supports," so why should they invest in Linux?

Getting the Palm Desktop -- an ancient program on both platforms -- running in Windows and Mac is blissfully easy. Getting things working in Linux are just way, way too hard. I'm not saying I won't figure it out but it's frustrating. It doesn't need to be this difficult.

But getting back to Palm and Linux. I tried KPilot in Ubuntu. I did get a sync, but what do I do with it? I was told that it would only dump data into KOffice programs I don't have installed. Then it wouldn't sync ... what did I do wrong? Guess I broke it.

I tried to install the real Palm Desktop software under Wine. Miserable failure.

My best hope is JPilot, which actually looks like it has the facility to handle the main components of the Palm world. I couldn't get it working in Xubuntu ... maybe in Ubuntu. Remember ... if the Web pages for the programs are sparse and haven't been updated in the past couple of years, chances are the project has been abandoned and the apps just might not work.

I know that the PDA is dying, that it only now exists as a value-added accessory to a cell phone. I also knowthat Palm's commitment to the stand-alone, non-phone PDA product is extremely shaky. But Palms are cheap, and I'd love to see a simple-to-configure, bread-and-butter Palm Desktop application that works.

In All Roads Lead to Debian, my account of a network-install of the venerable Linux distribution on my hacked thin client, I talked about the mixed-bag nature of getting printing to work. I could print to many printers on our network without any configuration at all, but I was having one hell of a time printing to a certain printer that wasn't showing up automatically:

The next test: configuring a network printer. I went to Applications-System Tools-Printers, and just like in Ubuntu, a bunch of the printers on the network were detected automatically. I easily select one and successfully print with it. But ... I try to add my "favorite" printer (i.e. the closest one to my desk) ... and while the instructions are clear -- and there are more of them than with Xubuntu -- I am unable to make it happen. I go through all the steps, and at the end, after I click "Apply," nothing is added, nor are there any messages telling me what went wrong. So I am able to print, just not exactly where I want.

I also try the utility under Desktop-Administration-Printing. While different, it still doesn't work. But since I can print somewhere, I'm good for now.

I booted into Debian today and decided to try the HP Linux Imaging and Printing utility -- HPLIP (do you say "hip-lip"?). Worked like a charm. Since my printer wasn't auto-configured, all I needed to do was select Setup New Device, then select the proper kind of connection (network for me) and then enter the IP address of the printer. HPLIP did the rest. Easy as m-f'ing pie.

HPLIP is as easy to use as anything out there -- including Puppy Linux's excellent printer-setup utility, and it's miles ahead of the main printer-configuration utilities in Debian and Ubuntu, both as far as ease of use AND actually working.

It's another feather in the cap of Debian (not that this same utility isn't available on who knows how many other distros -- this is just the first time I've used it). Question: If HPLIP works so well, why are so many other printer-config utilities installed by default in these distros? Choice is good, but only if all the choices are good.

Desktoplinux.com, the site where I first learned about the Dell-Ubuntu deal, is among those confirming it.

In the item, Steven J. Vaughan-Nichols goes into more depth about how Dell currently offers Red Hat on Dell Precisions, and that the company's "business" desktops are "certified" for Novell's SUSE Enterprise Desktop Linux but not sold loaded with the OS.

And then there's this tidbit:

While neither Red Hat nor Novell commented in time for this story, a Dell partner told DesktopLinux that they expect to see a multi-boot Linux system from the PC vendor that will offer users a choice of booting into Red Hat's community Linux, Fedora, Novell's community Linux, openSUSE, and Ubuntu. The official said, "They don't want to upset their business partners Novell and Red Hat. Also, those are the 'big three' for the enthusiast market."

Now that would be cool -- you could test the top three distros, all GRUBed together, without going through the installs.

So I try to run Synaptic in Debian 4.0. I select my packages to download and install, and then Synaptic prompts me to instert the Debian net-install CD into the drive. Problem: I disconnected the CD drive. I need a freakin' CD in the drive to download packages? Hopefully this will be a one-time thing.

But it only did that for AbiWord. I just installed Mousepad with no difficulty (and no CD).

So I try AbiWord again. I get the same message about inserting the CD. I just click OK, even though I don't have the CD drive connected. The message continues to reappear; I continue to click OK. Eventually nothing happens when I click OK, so I click cancel. It looks like nothing's going to happen, but Abiword appears to be installing.

I probably could've avoided this by connecting the damn CD drive, but then I'd have to reboot and plug it in. I'll have to try again, because AbiWord did not install.

I think this is some kind of quirk due to the network install I did. But asking for the CD on a net-connected computer? That's crazy.

Dell and Canonical have made it official -- the PC-making behemoth will install the No. 1 desktop distro, Ubuntu, on select models. The full interview on video with Mr. Ubuntu, Mark Shuttleworth, is on the Direct2Dell site, which recently asked users what they wanted in their next Dell (most said Linux, and most of those said Ubuntu). But if you don't want to watch the video and would rather read it, go to Ed Burnette's ZDNet blog.

Here's one question and answer:

How did this relationship between Canonical and Dell come about?

I think both organizations have been eyeballing each other for some time. Ubuntu has grown very rapidly as a desktop platform, and many of our users are running Ubuntu on Dell computers. On the Dell side, I think folks have noticed they were hearing about Ubuntu more and more. Michael Dell picked up on the trend of adoption. So, over the last couple years we've slowly been engaging. With the results of the Dell IdeaStorm there was sufficient critical mass for the idea, effectively, to take another step and turn some of those discussions into a project. The team has been working fast and furiously on that for the last couple months.

George Ou, who writes the Real World IT blog for ZDNet, did a lengthy analysis of Microsoft Office 2007 vs. the free Open Office 2.2 suite. Not surprisingly, Open Office continues to be a resource hog, although the situation is improving over previous releases. Go to the entry for all the numbers, but here are some of his findings:

Office 2007 base memory consumption went up significantly compared to the Office 2003 I measured last year, but it's still significantly less than OpenOffice.org 2.2. Some of the OpenOffice.org applications, like Base, require Java to run, and the memory consumption spikes over 70 megabytes as soon as you start navigating in the interface. However, the difference between Microsoft and OpenOffice.org base resource consumption has gotten smaller.

... we can see that the OpenOffice.org ODF XML parser (while vastly improved) is still about 5 times slower than Microsoft's OOXML parser. OpenOffice.org also seems to consume nearly 4 times the amount of RAM to hold the same data. While OpenOffice.org continues to have fewer features than Microsoft Office, it continues to consume far more resources than Microsoft.

... It would appear that OpenOffice.org 2.2 has gotten significantly better than version 2.0, but it still has a lot to work on. ... So while I may still consider OpenOffice.org a resource pig, the pig has definitely lost some weight.

Since this is an open-source vs. MS issue (and, to some extent, a Linux vs. Windows issue, even though OO has both Linux and Windows versions), there are dozens of comments in various states of support and anger. At least one points out that once you open one Open Office app, it's quicker to open another one.

It's hard not to notice that Microsoft Office apps open extremely quickly in Windows (and, of course, they don't open at all in Linux, unless you're doing so under Wine). I have MS Office 6 on my old Mac Powerbook 1400, and that version is a real, honest-to-God dog, it's so slow. But on a modern Windows box, MS Office is, if anything, fast as hell.

On my Windows box (which DOESN'T have MS Office), the Open Office "Quickstarter" is always sitting in RAM, allowing a fairly quick start of the program. I don't quite know how I feel about it in terms of resources. I don't really use OO that much -- I'm mostly running our paper's publishing system (Unisys Hermes) and for blog writing, when I'm not working directly in Movable Type, I use AbiWord or EditPad. And I open about one spreadsheet a month (I'm a total Excel-phobe) in OO.

I use OO so little on the Windows box, I'm still on version 1.1.4. I have version 2 downloaded; I just have to get around to installing it. We have MS Office on the iBook at home, but I'm not all that comfortable using it (I've gottten rusty in Word over the years). Of course, I have OO on most of my Linux systems, but I'm mostly using AbiWord and a variety of text editors at this point. My love affair with KWrite pretty much stalled when the only time I could get "typograpical quotes" to work was in MepisLite, a distro that Mepis pretty much abandoned. In both Slax and Kubuntu, the "smart" quotes don't work. So it's AbiWord for the moment (and I'm hoping for my two most-wanted AbiWord features -- "smart" quotes and the ability to change the case of letters from the keyboard -- to be added, though I am not holding my breath).

But in the larger world of open source and Linux, Open Office is VERY important. The fact that it's free is a powerful incentive to use it -- and since it covers most of MS Office's bases, it's essential for many who might consider switching to Linux for desktop use. To "sell" open source apps and operating systems to the unbelieving public, you've gotta be able to deal with MS file formats, and while AbiWord and Gnumeric suffice, OO is better, albeit way slower. But if you're spending your whole day in, say, OO Writer, you load it once and keep it running -- start times for the program aren't such an issue.

For instance, I use the GIMP a whole lot, and while it's slow to load, as long as I've got the memory to run it, I just start it when I first need it and leave it running. Not something I'd recommend with less than 512 MB (and something I'd definitely recommend with 1 GB). But for low-resource Linux systems, mtPaint does what I need -- and it loads in a few seconds. The same is true in Windows: Irfanview isn't as powerful as the GIMP, and the former program is kind of quirky at times, but it does a pretty good job of editing images.

Of course, the best thing to do is get a PC with tons of CPU power and memory and just be blissfully unaware of all this.