Working with Linux Mint as much as I have been lately, I’ve grown accustomed to the color-coded terminal prompts: green when you’re a user, red when you’re logged in as root.  That’s a nice touch.  The red lets you know you’re working without a net, and the color in general shows you at a glance that all your output is finished and you’re back at the command prompt.

I’ve been using Mint’s /etc/bash.bashrc file as a template to add the color prompts to other distros, like Debian, and it works fine.  But now I’m using Xubuntu on my main desktop, and Mint’s bash.bashrc settings were mysteriously not giving me colors when applied to Xubuntu.  It took me about half an hour of screwing around with it to track down the issue, which turned out to be this setting and block of explanatory text in the user’s local .bashrc file:

# uncomment for a colored prompt, if the terminal has the capability; turned
# off by default to not distract the user: the focus in a terminal window
# should be on the output of commands, not on the prompt
#force_color_prompt=yes

Well, isn’t that nice.  Someone at the Xubuntu project, or maybe at Ubuntu proper, is lecturing me about where my attention should be when I have a terminal window open.  Apparently I shouldn’t be paying attention to the prompt at all.  There must be something wrong with me if I find the colors useful.

Am I alone in thinking this is a weird place to stumble across self-righteousness?

Update:  Apparently this exists in Debian, too, so Ubuntu probably got it from upstream.  But no matter where it came from, it’s still presumptuous.

 

Well, I’m surprised to find myself using an Ubuntu distro again.

I’ve been using Linux Mint for almost everything these days, but like everything else in the Ubuntu/Debian family, Mint has been pulled in multiple directions by the unfortunate decision of the Gnome developers to go batshit insane, and the concurrent decision by Ubuntu to force a mobile device interface onto the 24″ dual monitors of bemused and disappointed desktop users all over the world.  So now the main branch of Mint has three different desktop managers: Gnome 3 (with or without the Mint extensions), MATE (a fork of Gnome 2), and Cinnamon (Mint’s own spin on making Gnome 3 usable).  I read the Mint forums regularly, and the one thing these three desktop environments have in common is that they are all apparently full of frustrating bugs.  Gnome 3 is not fully baked, and the bolt-on attempts to customize it back into something half as usable as its predecessor are even less baked; and MATE is also a work in progress.  Everything available in Mint 12′s flagship release seems like beta software.

I’m sure Mint (and other Gnome-based distros) will work all these problems out in time, but right now it’s just chaos, and I don’t want to climb down into that snake pit until it settles down.  So that leaves me with two options: either  1) stick with Mint 10 or Mint 11 until the problems are resolved, or 2) migrate to a different desktop environment, at least for a while.

I really don’t like the idea of sticking with an aging release, missing kernel updates and key app changes, for the year or two it might take for Gnome (and the distros that rely on it) to stabilize.  I might, if I thought it would clear up in that time, but right now I don’t trust them to do it that fast.  I think it’s technically feasible that they might solve their problems in that time, but it would require them to listen to what their users want, and so far they’ve shown little inclination to do that.  So I’ve been sort of pushed toward changing to a different desktop.  I’ve been threatening to do that for a year or more, and now it’s time to actually try to make the jump.

I like to stay in Debian/Ubuntu/Mint distributions, rather than migrating to Red Hat or SUSE based distros, partly because I like Debian’s apt package management better than Red Hat’s rpm or SUSE’s YAST, and also because the Ubuntu branch has always had the best font rendering of any Linux distros, which is a big deal to me.  Aside from Gnome, there are four practical choices for desktop managers in the Debian/Ubuntu/Mint distros: Enlightenment, LXDE, KDE, and XFCE.

Mint 12 KDE has recently been released, but it’s based on Kubuntu, which I have never had good luck with. I find it overcomplex and unstable.  Enlightenment is beautiful (see Bodhi Linux if you want to have a look) but I have yet to find any 64-bit distro built on it.  This leaves LXDE and XFCE, and there’s really no reason I can see to pick LXDE when XFCE is clearly the more flexible environment and both are about equal in resource requirements.  So XFCE it is.

Linux Mint has an XFCE version, but it’s based directly on Debian.  I like it a lot, and I think it is Mint’s future, but for now I prefer to stick with a distribution that gives me access to the Ubuntu and Medibuntu repositories.  So that ruled out Mint, and the next logical step was to try Xubuntu.  There is now a 64-bit version of Xubuntu 11.10 available, so I downloaded that and installed it to a spare hard drive on my main system to test it.  Note that I did not set it up on a test box; I used my main system.  Partly this is because I “borrowed” the test box’s monitor and desk space for work purposes, but mostly it’s because I’ve done enough experimenting with XFCE to know generally what it’s capable of, and this time I needed to see specifically how it would perform on my actual primary hardware – a quad-core Intel box with 4GB of RAM.

I did a tremendous lot of messing around with that first installation.  I wanted to see how close I could get to reproducing my familiar Gnome desktop, and after a lot of trial and error I discovered several key differences.  Some were resolvable and some not; some were actually better in XFCE, others just seem like pointless difficulties.

  • The desktop does not allow free placement of icons.  It’s a grid, and it’s pretty rigid about keeping you to that wide-spaced grid.  I’ve found several workarounds by googling the issue, but so far none of them have worked for me.  It’s not a deal-breaker, and I’ll continue to work on it, but I haven’t found a solution to it so far.
  • On the other hand, window decoration can be customized far more easily than in Gnome.  There’s a handy GUI tool for placing the buttons and window text just as you like it.
  • I struggled with trying to get Compiz working in Xubuntu for some time and finally gave it up.  I got it going, but I couldn’t use Metacity themes with it; only Gnome 3 themes, unless I wanted to replace that with Emerald, which is a bag of trouble all by itself.  Finally I had to ask myself what I was investing all that time for.  I use Compiz in Gnome for three things: wobbly windows, centered window placement, and compositing for Docky.  Of the three, only the wobbly windows were not available with XFCE’s own built-in compositor.  I decided I could live without them.
  • I miss the Mint Menu.  It was so easily customizable, for one thing, whereas the XFCE main menu is not customizable at all.  There apparently used to be a GUI for editing the XFCE main menu, but no longer.  There’s an LXDE menu tool that also works on XFCE menus, but it’s not very intuitive, and after I used it I found my XFCE installation seemed to be using parts of the Gnome 3 interface.  I could no longer change wallpaper using the XFCE tool, for example, but the Gnome 3 tool was there and let me pick Gnome 3 wallpapers and themes.  Coincidence?  Maybe.  I don’t know whether I could have fixed it or not; I just let it go because I knew I’d be reinstalling after my experimenting was done.  What changes I needed to make to the menus, I accomplished by editing their xml, but that should be easier.
  • Docky doesn’t integrate as well with XFCE as with Gnome.  The XFCE terminal emulator and file manager both put up generic dock icons when running, even though they have launcher entries.  I solved the first half of this problem by switching to gnome-terminal, and I’ll probably have a go at PCManFM for a file manager and see if it solves the other half of the problem.  But for now I can live with it.
  • Speaking of the file manager, the default Thunar file manager has an annoying inability to tell when the user wants to grab and drag a file or select a group of files.  In Gnome, if you button down on a file and start moving the mouse, it assumes you want to drag the file somewhere.  In Thunar, you have to click the file once to select it before you click again to drag it; otherwise it just starts drawing a multiple-selection box.  I hope this is configurable, or I hope PCManFM doesn’t do it.
  • This one is probably unique to me, because I don’t know anyone else who does this.  In Gnome, I found that a desktop launcher that points to a shell script will pass a path and filename to the shell script if you drag and drop the file onto the launcher.  I used this to create drag-n-drop launchers to burn DVD isos to disc.  This does not work with XFCE launchers.  However, I’ve found that if you have the actual script file on the desktop, it does work, so instead of launchers I use short script files that spawn a new gnome-terminal process and kick off the destination script.  It works, and I’m happy.
  • The default appearance for XFCE desktop icons is to have a solid color box behind the text.  I think that looks horrid.  Fortunately there is a workaround for that; it involves creating a .gtkrc-2.0 file and adding a few lines of configuration to it.  Now they look like Gnome and just about every other modern desktop environment.
  • Panel management is more logical in XFCE than in Gnome.  You lock down the panel’s configuration with a single selection, rather than item-by-item as in Gnome, and choosing to edit the panel preferences gives you an interface that controls them all, making it easy to sync them up.  There are a lot more options in there too, like setting the panel to automatically adjust its size within optional limits.

I know this sounds like a laundry list of complaints, but it’s not, really.  Most of them are more like observations about things that are different, or that I just don’t understand well enough yet to change.  The desktop itself is beautiful now that I have it set up the way I like it, which really didn’t take long at all.  It works fine with the Nvidia driver and multiple monitors in TwinView mode.  Docky works great, except as noted above.  Xubuntu came with the Droid fonts already set up as default, which was nice for me, because I hate the over-styled Ubuntu font.  And once I got done experimenting and started over with what would be the final installation, I put it on an SSD instead of a standard hard drive.  And you know what?  Running a low-resource desktop environment on a quad-core with an SSD is screamingly, frighteningly fast.  Gnome 2.32 was never slow or sluggish on this box, but XFCE on the same hardware is so fast that most things happen pretty much instantaneously.

Things I have left to do:

  • Try to change the file manager.
  • Figure out how to disable the desktop grid.
  • Change away from Evolution mail.  This has nothing to do with Xubuntu or XFCE, it’s just that the new version doesn’t handle previews the way I’m used to and ignores the font settings.

Other than that, it’s a nice, fast, pretty, usable desktop, and under the hood it’s the same Linux I’m used to.  I’ve already migrated my documents and such to it, and plan to stick with it for at least long enough to make sure it’s usable long term.

I think it is.  I think XFCE is ready for prime time.  It’s not just for low-resource systems anymore.

 

Here’s a story.

In 1979, I was young and overconfident, out on my own, married too young, living in a basement apartment for $190 per month rent.  We didn’t have any assets to speak of, really, just a couple of beat-up cars and some furniture rescued from a curbside.  My clothes were threadbare.  We got by, but we weren’t exactly living at the Ritz or anything.

One night we were on our way back home to our little apartment, in my wife’s car.  I was driving, and we were on a frontage road.  For non-Midwesterners, a frontage road is a small local road that runs parallel to a major highway, to keep traffic from snarling on the main road.  As we approached a side street, a car coming the other way suddenly turned left in front of us, and we collided.  Typical minor fender bender.  My wife and I got out and looked at the cars; the other driver did the same.  Her name was Mary Lou Tiger and she was only a few years older than we were.  I can’t say with authority that she was drunk, so I’m not going to make that statement.  I will say that I wondered whether she was.  But other than that, she seemed fine, if a little shaken up.

The police arrived in a little while, and we all described what happened, and were given accident reports to fill out.  Mary Lou Tiger was issued a citation for making a left turn in front of me, and we left.  Our car was not seriously damaged.

Some weeks later I found myself called to testify in court.  Apparently Mary Lou Tiger was contesting the ticket.  I arrived in court to find Mary Lou Tiger had acquired a slick-looking attorney, one E. Paul Lanpher.  She had also acquired a neck brace, which I believed then (and still believe now) was stage dressing.  The attorney asked me and the police officer questions about the accident and where it had occurred, then triumphantly noted that the police officer had written down the wrong name for the cross street.  They didn’t contest that the accident had occurred, just that the ticket was written wrong.  The assistant DA representing the city tried to have it amended, but the attorney prevailed and the ticket was thrown out.  Mary Lou Tiger walked out scot-free.

That might have been the end of it, except that not long afterwards I got some legal papers delivered to me.  Mr. E. Paul Lanpher had filed suit against me on Mary Lou Tiger’s behalf.  Suddenly this left turn she pulled in front of me, this accident she caused, was somehow my responsibility, and she wanted money.

Fortunately, I did have insurance, and with a reputable company.  I went to see my agent with the papers, and he grinned and said, yeah, they’re playing the game now.  No problem, we’ll take care of everything.  Give me those papers and don’t worry about it any more.  And that’s what I did, and I never heard anything about it again.

But it has stayed with me all these 33 years.  It was my first experience with someone abusing the legal system to try to screw me over.  I believe she knew it was her fault, and so did the lawyer, and yet they came after me – or rather, after my insurance – hoping to turn a shady profit.  I don’t even know if they succeeded, or whether State Farm Insurance told them to go pound sand.  It hardly matters for the purpose of this blog.  What matters is that they tried.

One day Mary Lou Tiger will Google herself, or one of her friends will Google her, and this blog post will come to her attention.  Hi, Mary Lou.  I’m still mad at you, you bitch.

 

In case you have not heard of this yet:

Jessica Ahlquist, a 16-year-old high school student in Rhode Island, has just won her lawsuit against her school over a prayer banner hung in the school.  The ACLU took the school to court, with Miss Ahlquist as the nominal plaintiff, after the school board refused to take down the banner.  Here is what the banner said:

Our Heavenly Father.

Grant us each day the desire to do our best.
To grow mentally and morally as well as physically.
To be kind and helpful to our classmates and teachers.
To be honest with ourselves as well as with others.
Help us to be good sports and smile when we lose as well as when we win.
Teach us the value of true friendship.
Help us always to conduct ourselves so as to bring credit to Cranston High School West.

Amen.

Now, that’s pretty obviously a christian prayer, but after several acrimonious school board meetings where person after person stood up defending the banner on religious grounds, the board tried to claim the banner was not religious but merely historical.  The judge saw the tapes of the meetings, didn’t buy it, and ruled that the banner is a prayer, thus it violates the establishment clause and must come down.

Miss Ahlquist had already been the target of online and real-life hate campaigns before the judge’s ruling, but that’s nothing compared to what followed.  Let’s look at some of the things her schoolmates said about her afterward:

  • May that little, evil athiest teenage girl and that judge BURN IN HELL!
  • Hope you’re happy #stupidbitch
  • If this banner comes down, hell I hope the school burns down with it!
  • U little brainless idiot, hope u will be punished, you have not win sh..t!  Stupid little brainless skunk!
  • Your a puke and a disgrace to the human race.
  • Shes not human shes garbage
  • Fuck Jessica alquist I’ll drop anchor on her face
  • definelty laying it down on this athiest tommorow anyone else?
  • Jessica Ahlquist may have won her case, but she’s going straight to hell.
  • I hope there’s lots of banners in hell when your rotting in there you atheist fuck #TeamJesus
  • When I take over the world I’m going to do a holocaust to all the atheists
  • like i hope you go to hell i honestly do youre a shitty person
  • @jessicaahquist your home address posted online i cant wait to hear about you getting curb stomped you fucking worthless cunt
  • gods going to fuck your ass with that banner you scumbag
  • wen the atheist dies, they believe they will become a tree. so we shld chop her down, turn her into paper then PRINT THE BIBLE ON HER.
  • what a little bitch lol I wanna snuff her
  • lol I wanna stick that bitch lol
  • nail her to a cross
  • Brb ima go drown that atheist in holy water
  • We can make so many jokes about this dumb bitch , but who cares #thatbitchisgointohell and Satan is gonna rape her.

Wow, there’s christianity in full flower, huh?

With examples like that, we don’t even need the establishment clause to know that banner should come down.  It should come down because it clearly, unquestionably doesn’t work.

That banner’s been up there for over 50 years, and it’s had a lot of time to do its work.  Clearly every one of the people opposing Miss Ahlquist believes in the power of prayer and endorses this particular prayer.  Well, does it look like it worked for them?  Do they appear to have “grown mentally and morally”?  (Not if their spelling and language is any indication.)  Do they seem “kind and helpful” to their classmate?  Are they “good sports who smile when they lose as well as when they win”?  Have they “conducted themselves so as to bring credit to Cranston High School West”?

The point of that prayer was to encourage the students to be decent, upstanding citizens and generally good people.  Do those horrific threats, curses, and insults sound like they came from good and decent people?  No, they sound like they came from vicious barbarians, who are demonstrating clearly for the whole world to see that this prayer has completely failed to have any effect on them.  They are not filled with fairness, decency, and compassion for their fellows as the prayer requests.  They’re filled with hatred, violence, and selfish spite.

This particular prayer has just had a very public test, and it has failed spectacularly.  No surprise to me, of course.  And it’s also no surprise that these ignorant cretins don’t see the colossal hypocrisy of defending a prayer for kindness, decency, and morality with threats of beatings, rape, and murder.

How very christian of them.

 

I know I keep coming back to this topic, but I just can’t get my head around religious claims that we are the creations of a perfect god who is unhappy with us for what we are.

It just doesn’t add up.  If your perfect god can do anything, how is it that he created something that disappoints him, something defective and broken?  If we are not what god wanted, why did he create us this way?  And why punish us for his failure to create what he wanted to create?  If you had a retarded child, would you beat it every day for being retarded?  Well, if you were a sadistic cretin, you might, but not if you were “enlightened” and “perfect.”

Here’s a new one to throw back at the evangelists and fundies.  If god created us, and made us defective, and through our defects we cause harm, then how is it our fault?  I’d think it would be the creator’s fault.  If your bottle of Coke turns out to be full of rat poison, do you sue the bottle?  Do you sue the drink?  No, you sue the Coca-Cola company, who made it and were responsible for the defect and its resulting damage.  Well, your heirs sue them, anyway.

Religions, christianity especially, are very fond of telling us we are fundamentally flawed.  They love to do that.  You are fatally flawed, yet you are the creation of a perfect being who can do anything, and he loves you, but he’s going to punish you if you don’t constantly apologize for being what he created you to be.  Sounds crazy, right?  And if you’re wondering whether I mean the god or the story is crazy, the answer is yes.

I think it’s time for a consumer product safety action against this “god” fellow.  Clearly he’s not building his products with any reasonable safety standards in mind.  They’re prone to self-destruction, they fall apart without warning, they operate inconsistently, and they’re poorly designed for many of their basic functions.  So far all we’ve heard from him on the quality issue is a series of self-contradictory press releases which deplore the faults yet mostly blame them on operator error, along with some vague promises about what will happen to the defective units after they’re taken out of service, which is too late to do any good.

If this “god” is really capable of building a superior product, we haven’t seen it yet.  I think some regulatory action is in order.  Maybe a discovery injunction.  Either prove he’s actually running the factory, or close down his PR machine and run him out of business.

Now if we could just figure out where to send the process server.

 

Steven J. Vaughan-Nichols over at ZDNet wrote in a recent article that the open-source web server nginx has taken over second place for overall web server popularity, in some metrics anyway. This puts it far behind front-runner Apache – nothing’s going to catch Apache any time soon – but a hair ahead of Microsoft IIS.

Think about what that means. Nginx – the project of a lone programmer in Russia, released as free software – is whipping Microsoft’s enterprise product, which costs thousands of dollars and has a huge marketing and development organization behind it. Facebook and WordPress.com use nginx. That’s amazing, and really impressive. It impressed me enough to finally overcome my inertia and try nginx for myself.

The big draw of nginx (pronounced “Engine X”) is speed and low footprint. I’ve noticed that my Apache server tends to open lots of tasks. Nginx is supposed to be faster and more efficient than Apache, so to test this claim, I installed nginx on the lowest-powered machine on my network – a tiny Asus eeePC netbook with an Intel Atom processor and 2GB of RAM. I figured if it could run effectively on that platform, it’s pretty darn efficient.

The eeePC was already running Linux Mint 10 and had Apache installed, just to serve up a single placeholder page. I installed nginx from the repositories, disabled Apache by setting its /etc/init.d entry not executable, and rebooted just to get a clean launch, then navigated to a domain served by that machine. Nginx came up just fine and served the placeholder page just as Apache would have. The only way I even knew it was nginx was by navigating to a bad url at that domain, where I saw nginx’s 404 instead of Apache’s. So far so good. But I had bigger tests in mind.  I noticed nginx from the repos was a fairly old version, so I added a ppa to get the latest version and upgraded it before continuing.

Next step was to set up WordPress. I followed the WordPress 5-minute installation process, which took somewhat more than 5 minutes, because I also had to install and configure mysql, php5, and the php5-mysql and php5-cgi packages, and alter the sites-enabled entry for the domain to handle php for WordPress. Fortunately there are abundant guides on how to do that last part, available for easy googling, which is a good thing because nginx’s configuration language is significantly unlike Apache’s. But it didn’t take too long before I had a working WordPress installation on one domain while another one was correctly serving a single static page.

Up to this point things had been going perfectly. The next step got me into trouble. I decided to migrate dwasifar.com over to the test server temporarily and see how it worked under actual load. Transferring the files went okay, and the additional configuration seemed to load fine. But no matter how I altered my local DNS, I could not get local machines to go to the test server for dwasifar.com; it always seemed to go to the live server. Finally I decided to disable that part of the configuration and move back a step so I could figure out what was going on, and at that point I discovered that nginx was no longer serving any pages to the outside world, even the ones that had been working previously.

Naturally, since I had been tinkering with DNS and DHCP on the local network, I looked there for the problem, but nothing I did on the name server or the eeePC made any difference; nginx was simply not there as far as the outside world was concerned. I did notice that every time I rebooted the eeePC I would lose my resolv.conf settings, so I had been trying to prevent that from happening, and eventually I just decided I had messed things up enough on the eeePC and the best thing to do would be to start fresh ( which turned out not to be needed, in the end, but we’ll get there).

I decided to go with a real server version of Linux this time and not a desktop distro with things stripped out. There is no server version of Mint, so I went back to Ubuntu. I downloaded Ubuntu Server 11.10 and installed it from a USB stick. It seemed to install correctly, but hung at boot. I tried Ubuntu Server 10.10, on the theory that if Mint 10 worked, so would Ubuntu 10.10. It would not install from the USB stick; it wanted a CD. I tried Ubuntu Server 11.04, and got the same results as 11.10 – apparently installed fine, but hung at boot. So I said to myself, screw this, and I went to Debian Stable, currently version 6. Debian took a lot longer to install, but when it was done, the system booted up perfectly, and I started to configure it.

Because my AT&T Uverse service requires it to assign an IP by DHCP for any machine that has a public IP, I had to set a local IP as well on this machine, so the first step was to run ifconfig to check the current IP setup. Imagine my surprise to find the machine had the intended local IP already set up as eth0, with no sign of the public static IP. I played around with /etc/network/interfaces and restarted networking a couple of times; no change. Finally I went and looked in the Uverse gateway, and swore mightily when I discovered that the gateway, all on its own, had dropped the public IP assignment I had set up for this machine and was now serving it a local IP which was supposed to have come from the dynamic pool but mysteriously was not in the pool range.

It all made sense. This was why the machine had suddenly stopped responding to outside requests; because the static IP was going nowhere. I probably wouldn’t have needed to rebuild this machine if I had known that earlier, although it’s probably still a good thing that it’s now running on a known good, clean, new install of a real server environment that never had a GUI. Resolv.conf was still being overwritten on boot, and I finally found out it was because of the dhclient process, so a couple of changes to dhclient.conf straightened that out. I transferred all of dwasifar.com over to the new machine, again had problems getting local DNS to resolve it consistently, and finally just set it up to resolve to one of the other domains instead. This worked, and I finally had a working clone of dwasifar.com on the test machine.

So after all that, the big test. Would nginx on a puny netbook be able to keep up with Apache on bigger hardware? I tested it myself, and to me over a local connection it looked like the live site still had the advantage; I saw a brief pause before each new page loaded. But it wasn’t a fair test from inside, because the connection between my local desktop and the web server is gigabit, whereas the netbook only has a 10/100 ethernet port. So I asked several friends to try the two systems, and compare them, and report back to me if they detected one was faster or more responsive than the other.

I have heard back from three out of four people. Two say they seem no different. The other says the netbook running nginx has a slight edge.

This is probably reason enough to migrate dwasifar.com to nginx eventually. I still have to work out how to set up the ssl part of the site, for webmail and proxying; and I probably want to get a little more familiar with the configuration language before I make the jump. But if nginx on a netbook can compete with Apache on a real server, well, that’s very tempting.

And of course, I am sick of this damn Uverse gateway. It is absolutely inexcusable for it to simply forget a static IP assignment. I never touched the damn thing’s configuration. One of my friends runs Comcast business class internet and loves it. I’m really tempted. Wouldn’t you be?

 

There’s an interesting article over at Forbes explaining why Best Buy is headed for bankruptcy and failure – in one analyst’s opinion, anyway.  He takes a recent visit to Best Buy with a friend as an example, and extends it out to become a general critique of Best Buy’s business practices and philosophy, comparing it to competition from Amazon.  Here is the key point:

[Amazon] simply does what consumers want.  Best Buy does what would be most convenient for the company for consumers to want but don’t, then crosses its fingers and prays.

That is as neat a summing up of what is wrong with American brick-and-mortar retail as I have ever seen.  Best Buy is the issue here, but it’s not just them.  Sears/Kmart has this problem too, for example, and so do many, many other businesses that should know better.

Case in point.  The Forbes article describes the writer and his friend’s experience with a Best Buy staffer, who didn’t know any answers to questions about the products the pair were shopping for, but was very persistently and awkwardly trying to sell them an unrelated video streaming service, and kept pushing the conversation back to that, to the point where the writer was forced to be rude to him to chase him off.  He identified this as “anti-service,” and it is.  The staffer was not only not helping, he was actively impeding the attempted purchase.

I have had this experience myself.  Shopping for a laptop in a Staples store, I was approached by a sales girl – and I do mean “girl,” as she could not have been more than 18 –  who asked if I needed any help.  I proceeded to ask a few technical questions about the laptop.  She didn’t know any of the answers, but instead wrenched every question into an opportunity to try to sell the Staples extended warranty program.  This became frustrating very quickly; eventually I realized I was not going to get any actual help from her and attempted to politely chase her away.  But she stuck with it, still trying to sell me an extended warranty on a laptop she had given me no reason to buy, until I finally just excused myself and left the store without buying so much as a ballpoint pen.

The reason for this is obvious, and well-explained in the Forbes article.  These floor staff are not given training about the product; instead, they’re just told to approach customers about the upsell du jour, and their performance ratings are based on that, not on how many customers they made happy that day.  Yes, it would be very convenient for Best Buy and Staples if customers really did want to be sold the Upsell du Jour instead of what they actually came in for.  But that’s not what we want.  It’s very unlikely I will ever shop at Staples for a laptop again, because they’ve shown me they’re more concerned about maximizing their add-on profit than they are about making sure I get the laptop I need.  At the very least they could put the full specs out on display with the product, but they don’t even do that, so there’s no way for me to know what’s in the box if the display doesn’t tell me and the staff can’t talk about anything but the extended warranty.  And it’s the same all over; Office Max and Office Depot are exactly the same.  And when Circuit City was in business, they behaved the same as Best Buy does now.

Another example.  In Kmart’s checkout lines, an automated survey about the shopping experience interrupts the checkout process by displaying itself on the customer POS terminal.  You can’t swipe your credit card until you make a selection.  I wrote to Kmart pointing out that the data they receive that way is worthless, because most people will just push any button to make it go away and complete the transaction, and in any case it doesn’t make sense to ask people what they thought of the experience when it isn’t over yet.  They did not reply.  Since then I’ve noticed the clerks have started to just reach over and push any button to make the survey go away.  If the data was bad when the customers were making random selections, it’s worse now.  And here’s that same attitude again.  It would be convenient for Kmart if this were what customers wanted, but they don’t.  The survey is not a customer service tool, it’s a customer anti-service impediment, and any money they spent deploying it and analyzing the results is money worse than wasted.  I would rather they put that money into dental insurance for their checkout clerks, so I don’t have to look at someone’s horrifying rotted teeth and be reminded that these people barely earn a living wage.

The Forbes analyst correctly points out that people go to Amazon not because Amazon has technological advantages, but because they are interested in providing what the consumer actually wants, rather than in pretending the consumer wants what a merchant wants them to want.  Netflix learned that lesson the hard way in 2011.  Sears/Kmart is going to learn it this year, I’m guessing, in bankruptcy.  And maybe the Forbes analyst is right that Best Buy will learn it soon too.  But it’ll be too late to do any of them any good.

 

I maintain a web and mail server, and there are times when I want to access the mail on that server via webmail rather than a regular IMAP client like Thunderbird or Evolution.  This requires a webmail client running on the web server that can connect to the mailboxes via POP or IMAP.

When I first set up the server I followed a howto that included setup for Squirrelmail, which is a very popular webmail client, but not as fast as I would have liked.  So I switched to Roundcube, and I’ve been using that for some years now.  But the newer versions of Roundcube are slower, and they have an annoying IMAP authentication bug that I’ve had to code around when I try those versions, so I had stuck to using an older version of Roundcube.  It worked fine, really, but because I can’t leave well enough alone, I went looking for something else to try, and I wound up downloading and installing a PHP webmail client from AfterLogic, which they call Webmail Lite.

AfterLogic Webmail Lite is the free version of their paid product, Webmail Pro, and it’s got enough features to be perfectly functional as a pure mail client.  The Pro version adds other features for users, such as an appointment calendar, and for administrators, like the ability to configure the product to multiple mail servers (the Lite version limits you to one).

The installation instructions on the AfterLogic website are easy to follow, and actually go into a bit of overkill about setting file and folder permissions.  It only took about five minutes to download and install it, and another few minutes to do the initial setup.   There is a setup page that walks you through the steps of pointing the configuration at the right address, choosing IMAP or POP, SSL or not, and so forth; when that’s done you delete the installation directory and log in as a user.  No specific user setup is required; Webmail Lite automatically sets up each new user the first time they log in, which is a nice touch.  All the user needs is the correct password for their existing mailbox on the system.

I logged on as a normal user and got logged in immediately.  It loaded up in what has now become the standard vertical column configuration, wherein you get one column of inbox and another vertical rectangle for preview.  I’ve never liked that, but fortunately this is a user configurable setting, and easy to switch to horizontal panels – inbox on top, preview pane underneath.

AfterLogic Webmail Lite is fast.  I mean, really fast, and really responsive.  Roundcube was a lot faster than SquirrelMail, and AfterLogic is that much faster than Roundcube.  When it loads, you have your inbox list immediately, and when you click on an item in the inbox, the preview is there immediately too. It’s also a lot more visually polished than Roundcube, with a nice layout, clearly understandable buttons, and a choice of five themes, two of which  - AfterLogic and AfterLogic Dark – are actually tasteful.  (The other three – Alice, Barbie, and Ya-Ya – look like they were designed by the colorblind for the just plain blind.)  I settled on AfterLogic Dark, which is a pleasing and subtle theme in shades of dark dusty blue.

Unfortunately the appearance choices are also where I ran into the first thing I didn’t like, which was that the inbox font was inconsistent from browser to browser, showing up with a serif font in Chrome and a sans-serif font in Firefox.  I made a mental note of it and continued.

After checking out the mail-reading options, which worked flawlessly and rapidly, I went to send an outbound mail, and here is where the biggest problem came in.  It simply did not work.

I opened a forum issue on it and continued trying to find the solution on my own, and through examining logs (which are copious and detailed, if you have them turned on) and experimenting, I determined that there was a serious bug in how Webmail Lite handles outbound mail if the credentials for sending it are different from the inbound mail server’s login.  If you have configured it to point to one server for IMAP and a different server for SMTP, it connects to the proper SMTP server but tries to use the user’s IMAP server credentials to authenticate to the SMTP server.  This fails, obviously.  I reported my findings in the forum thread, and got a reasonably prompt reply from someone named Igor who confirmed the issue and promised to take it to the developers.  In the meantime I picked apart the PHP code, found where the SMTP authentication takes place, and worked around the problem for the moment by replacing the variables holding the login and password with literals containing the actual login values for my outbound relay.  Kludgy, but it works, and will get me through until AfterLogic solves the problem properly.  This was not the first code change I wound up making.

After that, everything was more or less working, but there were still a couple of things that didn’t quite suit me:

1) The aforementioned font inconsistency problem.  I went looking at the .css files in the themes (called “skins” in Webmail Lite) and found that they contained only Apple and Microsoft fonts.  So I edited the .css to include Droid Sans as the first choice, which made all the fonts consistent across browsers on my Linux machines.

2) The login screen contained a checkbox to make future logins automatic by saving the login info in a cookie.  Since I use more than one mailbox, I didn’t want that option, and edited the code to hide it by setting style=visibility:hidden in its <div>.

3) The favicon – a tiny red envelope – has some cuteness attached to it whereby it flashes in the Chrome or Firefox tab if there is any unread mail.  A nice touch, but annoying to me, since I often don’t open every mail that arrives; a lot of the time the subject tells me what I need to know.  So I suppressed that too, by changing a file called favicon-blink.ico to be the same as favicon.ico.  I guess it’s still blinking, but alternating between two identical icons now, so you don’t see it.

These last three things should be configuration options, and they aren’t.  I tweaked them on my own, but most people won’t want to do that, or can’t.

Aside from those issues, I do really like AfterLogic Webmail Lite.  Now that I’ve done the tweaking, it’s fast and attractive.  I’m waiting to see how long it takes Igor to come back with something from the developers.  I imagine I’ll continue to use it; I can always switch back to Roundcube if I change my mind.

 

Two days ago Michele Bachmann confidently predicted an Iowa caucus win, in defiance of polls saying she was going to get her ass handed to her at the ballot box.  Her reasoning: god wanted her to win.  Apparently he told her so, and was preparing a “miracle” to see to it, because after all, he’s her personal friend:  “We’re believing in a miracle because we know, I know, the one who gives miracles.”

Yesterday she got her ass handed to her, as the polls predicted.  Sixth place.  5% of the vote.

So, I’m guessing she’ll shut up about having a direct line to god now, right?  She’ll admit she was mistaken about that, and maybe the voice in her head she mistook for god was just her own wishful thinking?  She’ll acknowledge that either god doesn’t want her to be president after all, or else he has no power to influence elections?

Ha, ha!  Oh, yeah, that’ll happen.

But it will be great fun to see what she does come up with to explain why god lied to her and broke his promise to deliver a “miracle.”  Though I have to say, divine intervention to make Bachmann president would be a “miracle” in the same sense that insurance companies describe natural disasters as “acts of god.”  It’d be a pretty good indication that god hates us.

 

Right now with the Republicans gabbling like a bunch of mindless geese in Iowa, there’s a lot of attention to the Tea Party/Libertarian concept of “free market,” which in their eyes means no regulation, no control, everyone in the market just does as they please.  ”Laissez-faire.”  There was a time in my life when I believed that would work just fine, as the Tea Partiers and Paulites believe today.  But here’s the flaw: if an unregulated market for [x] eventually leads to the consolidation of the [x] market into the hands of a few giant, powerful concerns, then the competition that these people claim to treasure is squeezed out; and without the possibility of competition it cannot really be considered a free market.  If the barriers to entry to the [x] market are so high that no new player can reasonably compete unless they already have huge assets, it becomes a catch-22; you can’t get the assets unless you’re a player, and you can’t be a player until you have the assets.

What made me think of this was the price of milk.  A few years ago we had a dairy oversupply, resulting in low prices for milk, butter, cheese, and so forth.  The dairy industry’s response to this was to sell a lot of dairy cows to slaughterhouses.  This resulted in a brief period of low beef prices and an extended period of higher dairy prices.  This makes sense for the dairy industry, if they can manage it – who wouldn’t want to do less work and make more money?  But the only way it can be managed is if the “dairy industry” is consolidated enough to act monolithically. That wasn’t possible a generation or two ago, but it is today, because today huge corporate interests control most American farming activity.  You couldn’t count on a nationful of small farmers to each independently sell Bossy and Elsie to be made into hamburger. But if there are only a few big players in the industry, who all see the advantage in doing that, it’s going to happen, and the result is more profits for them and higher prices for us.  Joe Farmer, Elsie, and Bossy don’t matter any more; they’re not competition for Dairy Inc., and will never be, and if Joe Farmer hasn’t sold out to Dairy Inc. yet, he will soon.

This is not competition.  This is hegemony, and if you’re out there campaigning to elect a president who promises to knock down what few barriers remain to complete megacorporate dominance of every market, expecting it to bring you cheaper and better products and services, look around.  We’ve been bulldozing those barriers for a generation now, and corporate juggernauts run almost everything.  Laissez-faire paradise should be knocking at our doors.  Instead it’s a process server with your foreclosure papers.

Laissez-faire “free markets” don’t lead to unrestrained competition.  They lead to monopoly – de facto if not de jure, because if a small number of huge “competitors” move in lockstep instead of competing, there’s no actual choice between them for the consumer.  Maybe it’s collusion, maybe it isn’t.  It hardly matters; what matters is that the market becomes monolithic, either by nominally independent action, or through concerted efforts by trade groups.  The RIAA and MPAA are perfect examples of this.  The MPAA is a functional monopoly, representing an industry that is acting as one unified entity to skew the legal system, the better to capture and exploit its market.  Same for the RIAA.  It’s pointless to argue that Warner competes with Sony competes with Disney; the pertinent unanswered question is who competes with the MPAA?  And the answer is, no one.  And no one can.  It’s a monopoly, and it was allowed to attain that status by buying special privilege from government.

Is this what we want, really?  Because it’s what the Republican candidates are selling.  Come, vote for us, and we’ll double down on the economic policies that stole your jobs and retirements.  If we just give the corporate interests even more of what they want, which is ownership of everything, sooner or later the benefits will trickle down to us.

Any day now.

Any day.