Well, today was my last day as a member of the rPath team. I will be moving all my blogging activities to http://blog.gafton.net (they weren't that many to begin with, but now with some newfound time on my hands, who knows...). It is unlikely I will be syndicated as part of Planet Conary much longer, so I am looking where to attach my new blog site to...

American Airlines has decided to charge $15 for the first checked in bag. As a not infrequent air traveler, this strikes me as the ultimately boneheaded move. Most of the flights I am on are fully oversold already, and one of the more annoying issues is that the overhead storage always gets filled with luggage halfway through boarding. People just don't like checking in their luggage - overstuffing a carryon (or carrying a bigger one onboard) is much more convenient.

If capitalism has taught anybody anything, it is that usually businesses get ahead by offering convenience; they make profits by taxing convenience. People will pay for something that is perceived as having value, and will resent a price increase on a bad service. Most of my fellow business travelers resent having to check in a bag because of the additional time required to retrieve it, as well as an increased risk of losing the bag in transit - and the US airlines' track records on lost luggage are particularly appalling. The delays caused during the boarding process by too many carryons again translate in more money lost for the airlines, as pilots have to make up the lost time by flying faster and burning more fuel. I doubt that a $15 tax on the first checked bag - something people avoid doing already - will recover any of that cost.

If somebody at American Airlines would be smart, they'd put a tax on the large carryons. The perception of paying for convenience would be kept intact, business travelers would be happier because (maybe) the overhead bins would be able to accommodate their luggage, and airplanes would manage to leave the gate ontime in maybe a higher percentage, letting pilots fly at a more leisurely pace. I'm just wondering what sort of McKinsey alumni Harvard grad it took AA to come up with their current proposal...

One of the more annoying pieces of crapware I have to remove from the manufacturer's Windows preloads is Norton Antivirus. That's not because I hate it on principle - that is because I hate it based on my own usage experience. I (have to) use Windows for maybe one hour every couple of months or so. Even with that limited usage pattern I managed to develop a deep resentment to this worthless antivirus "solution" - for the way it immediately hogged all my available bandwidth on bootup to download enormous updates, for the way it didn't let me cancel that, for the way it stopped my work until the updates were installed, and for the way it required a reboot or two just before continuing. Ultimately, for the way it used to make "boot into Windows" a 30 minute affair before the damn thing got out of my way.

I thought I was alone wondering in amazement how people can charge money for it, until I read somebody else's perfect description of how I feel about it:

If ever a class of software deserved to be cast into the lower reaches of Hell and run on Satan's own desktop, it is this. Performance- sapping, space-hogging, noisy, irritating and prone to inducing just as many problems as they purport to solve, these horrible, ineffective, expensive lumps of digital thuggery keep entire platoons of support engineers in business and home users in tears.

That's part of a good article about how software should not work. Actually, the software itself is mostly fine, it is the surrounding packaging and integration into the operating system that sucks by being invasive, obnoxious and/or just plain annoying. In the world of Linux distributions and free software we have been spared the whole mess created by having every single application handle its own install, update/upgrade, dependencies and uninstall. Can you imagine the horror?

Like so many of my source code jockey brethren, I am one of those saddened by Lenovo's sudden decision to stop selling laptops equipped with "normal" aspect ratio screens. This widescreen all-the-way fad is bugging the hell out of me. At around 5lbs, a T-series laptop with a 14.1" screen and 1400x1050 resolution was a rather nice machine for a traveling programmer. It seems that it's either too few of us buying normal aspect ratio screens or nobody at Lenovo can possibly think of why anyone would value vertical resolution above and beyond the wide-screen idiocy.

Anyway, I was in the market for a new T series laptop, and as I was pricing out a new one online, I noticed a rather great deal they're offering their Linux users:

Yepp, that's right - one can upgrade from SUSE Linux Enterprise Desktop to a full-blown Thinkpad PC DOS License for the low, low price of $8,473.65. I think I should upgrade the CPU to the fastest they offer, memory to the max and get the biggest, baddest hard drive they offer to be able to run that PC DOS beast...

Thinkpads have been my laptops of choice for many years. I have personally owned/changed/worked through 9 of them so far. The Thinkpad line always used to have a 14.1" regular aspect ratio screen in their lineup. As a "businessy" user, I hate change. I hate having to change my coding habits to deal with widescreens. A source code terminal would only utilize half the screen estate, which would bug me to no end. I hate having to purchase a new laptop bag and/or travelcase to accomodate the new screen dimensions. I hate wasting time researching which bag would work best - my current one has not been produced for years now, and keeping up with new developments in the luggage industry has not been one of my curiosities, ever. I hate that these new Thinkpads only come with a 6-cell battery instead of the awesome 9-cells they used to have before. I hate that the new fully-Lenovo designed models are bulkier, thicker, boxier and much heavier than the ones they replace.

I hate that people who feared that Lenovo would end up trashing the prized Thinkpad line appear to have been right.

At long last, my loyalty to the Thinkpad series is forced to an end by Lenovo's dumb decision to stop selling models that I would like to use. The part that sucks the most about this is the fact that I have to waste time looking at other brands of laptops - something I have not done in a long time.

A while ago I have started looking for a more energy efficient way of storing my electronic waste. I do have a big server tower now with a bunch of SATA drives in it, and over the winter its warming effects on the temperature in my office have been appreciated. However, current hard drive prices have made it so that a pair of 1TB drives in a RAID1 configuration can provide me with the needed storage space at a cost that will be easily recuperated from the savings created by turning off this 450W energy hog.

The default option would be to get a USB 2.0 enclosure, connect it to an old laptop, and use the laptop to serve my NFS and samba needs. However, USB 2.0 is slow and I have managed to stress out the usb-storage driver in more than one occasion. This solution is not much on confidence, not much on speed, I'll try to avoid it for as long as possible. (also, e-sata and laptops don't quite mix...)

The next option is a dedicated dual drive NAS. The two contenders here seem to be the Linksys NAS200 and the D-Link DNS-323. Looking over the interwebs it seems that the general consensus is the Linksys device is rather slow. So off I went and I purchased the D-Link one.

That turned out to be a mistake, as the DNS-323 proved to be unable to recover from a simple RAID1 rebuild/simulated hard drive crash test. It didn't lose data, but I couldn't get the device to rebuild the RAID1 array and keep it from going into degraded mode until I reformatted. Not only that, after the simulated failure, it became very unstable, timing out and giving random read only and permission denied errors as I attempted to test working with the device in degraded mode.

I really liked the form factor of the DNS-323 though. That is basically what I am looking for - not much bigger than two hard drives put together, temperature sensor, quiet fan that adjusts its speed based on how much cooling is needed.

I guess it is time to start looking at the Mini PCs. It's hard to believe that what I would like to have is that "far out" and that nobody has attempted to develop something along these lines:
- dual 3.5" drive support
- gigabit ethernet
- small, really small form factor
- x86 based, prefer freedos to windows vista
- no monitor, don't care about graphics performance
- ~50W power consumption
Hardware folks, please design something like that and I shall buy.... at least one.