I must've done this before, but I have no memory of ever taking a Windows XP Service Pack 2 installation and upgrading it to SP3.

But I did one tonight. It went pretty much without incident.

However, it took a long, long time. The installer/upgrader/whatever-you-call-it seemingly had a lot of prep, then a lot of backing up, then a lot of installing, then a lot of cleanup to do.

I think it took 3+ hours.

Yes, slower than the average Linux distro installation.

But the positives are that the box didn't blow up. After that long, long wait, it rebooted and is now running. Relief was sighed.

Yes it is.

The OMG!Ubuntu blog reports on the decision, however preliminary, at the Ubuntu Developer Summit in Dallas to remove the GIMP image editor from the 10.04 Lucid LTS release of the wildly popular Linux distribution.

Read the well-wrought entry linked above for the drawn-out reasoning behind moving the "professional"-quality Photoshop killer GIMP from the Ubuntu base (it'll be available in the Ubuntu Software Center, or your other favorite package-management tool).

Those assembled seem to think that GIMP is not used enough and is not consumery enough. And that the F-Spot photo manager can do basic photo editing and is much better for the average user.

Oh, do I have bones — plural — to pick over this one. I still haven't made my decision on whether I'm for Mono (using the Microsoft-compatible open-source tools) apps or against them (and F-Spot, along with Tomboy notes and, if you've added it, the Banshee music player seem in my mind anyway to be the highest-profile Mono apps in the GNOME world).

All I can say is that with the geek-political climate these days, more Mono rather than the same or less will just give more users a reason to jump off of GNOME (and Ubuntu) in order to keep one's collective hands, if not clean, than at least Microsoft-free.

Again, I haven't made a personal decision about Mono as yet, but I'm far from happy with F-Spot.

And yes, I've been using it somewhat regularly. For my purposes, I'm not crazy about having to import images into F-Spot. digiKam can deal with images in any directory structure, and I'd like my photo-organizing program to do the same. I understand that F-Spot is more iPhoto-like in this aspect. I still don't like it. It's OK for my personal images, but I can't keep my businessy images separate. Everything's in one big pile in F-Spot, except when you dig into the actual directory structure the app creates. Yep, just like iPhoto.

In F-Spot I can add a caption in the "comments" area. Unfortunately that data does not come up in any other applications I use to edit or view photos. I can't edit the IPTC data that 100 percent of professional photojournalists use (and those are the guys whose images I handle day in and out).

F-Spot will sharpen and adjust the color of images. It will crop them. But it won't resize them. Huge, huge deal-breaker for my "professional" use of this application. (And why would I use something for my "home" images that won't do the job with my real work if I don't have to?)

Truth be told, I don't require all that Photoshop offers. On the PC I use IrfanView. And basically my "quest" for a Linux/Unix image viewing/editing program runs along the lines of "give me something that does what IrfanView can do."

Even the GIMP (and Krita, too, O fans of KDE) can't deal with the IPTC data in JPEG images, which I absolutely need.

The digiKam image manager in KDE, through the great Kipi Plugins, CAN deal with this data, and pretty well, too (although the limit on the length of the IPTC credit line is a bit grating and seemingly unnecessary).

So I've been using digiKam for the past few weeks somewhat regularly. (Truth be told, I tend to work in IrfanView on my Windows box at the office about 80 percent of the time when editing photos; it's the environment I know, and that does what I want it to do.)

digiKam is a bit unwieldly. Like many KDE apps, there are menus for days, along with choices to match. It resizes. Good. It sharpens (although the results aren't as good, seemingly, as in every other app that sharpens images; there are, again, lots of choices, and I barely understand — and can't get a great result — from them. digiKam can crop, but you can't enter the exact dimensions of your crop in pixels and then drag the box around to make the perfect crop like I do in IrfanView. Not a deal-breaker, but not good either.

And did I say digiKam is unwieldy. Why are there separate "edit" modes for the metadata and the image data?

I've had little ol' gThumb on this Ubuntu machine for awhile. And hearing that the UDS suggested and then rejected it as a "replacement" for either GIMP and/or F-Spot prompted me to try it out. Sure I had opened a few images, but I hadn't yet done any heavy lifting with gThumb.

It was time.

Gthumb, little ol' gThumb (that's what I'll call it for the purposes of this entry), does almost everything I need:

-- Deals with images in their current directory structure
-- Resizes images to exact pixel dimensions
-- Crops images to exact pixel dimensions
-- Can edit/add IPTC caption info (to the main caption area only) with the "comments" feature
-- Allows for easy save-as of images

The only thing gThumb doesn't seem to do (and I could be missing it, though I don't think I am) is sharpen images. I can live without that, especially if gThumb can create and won't destroy existing IPTC data in JPEGs.

(Note: Besides Krita and GIMP, my previous favorite light image editor for Linux, MtPaint, is also an IPTC-data-destroyer and therefore can't be used for my "real" work.)

So thanks UDS people, for mentioning gThumb. And if you're asking my advice, and I know for damn sure that you're not, keep the GIMP or don't. I'll install it anyway.

But look deep into your geeky, geeky hearts and find it within them to replace F-Spot with gThumb. Or at very least make gThumb part of the Ubuntu base, make it the default image-organizing app, and let the rest of the free, open-source software-using world discover this most worthy of applications that for the most part can free me from the purgatory of Windows-based photo editing applications for good.

(And while I'm on the well-trod soapbox, let me mention that I wrote this entire entry using the newish Webkit-based Epiphany Web browser, another lovely bit of GNOME that I liked in its Gecko days but like even more now.)

(And sorry [really] about all those parentheses, within which I'm thinking all too often these days.)

I couldn't have said it better myself.

Writer of the post, Aaron Toponce, is so good, he's going in the blogroll. He also recently started a company called Root Certified that does "Managed Hosting Solutions as well as Linux and UNIX consulting." Worth checking out, I imagine.

Notice how I'm doing quick entries on WorksWithU just about every day? That's because WorksWithU is one of the very best sites out there on Ubuntu and the enterprise, with a lot applicable to the average user as well.

In its latest entry WorksWithU reports on Dell's plans for the next Ubuntu LTS, 10.04, due as its numerical designator hints in the fourth month of 2010 (aka April).

What interests me more than the entry itself are the comments that follow.

The commenters on the post are talking about the idea of a Ubuntu LTS "rolling release" in which the base of the system maintains some consistency over the life of the release but the applications on top of it are updated/backported as they develop over time.

This is more akin to the Windows/Mac (and RHEL) style of system in which a full OS upgrade isn't required to get a newer version of an application such as Firefox or OpenOffice.

For instance, to get Firefox 3.5 in Ubuntu without using a backport or PPA (the newish "personal package archive," which I'm eager to learn more about and start using on my own systems), you need to be running the current six-month release, Karmic (9.10).

Jaunty (9.04) released in April of this year, still runs FF 3.0.x, as does the current Ubuntu LTS (8.04), which is now a year and seven months into its three years of desktop support.

Whether automatically, or by user choice, major apps such as Firefox could be backported into the LTS, either by Canonical/Ubuntu, or through a special Dell repository, I think many users would be interested in running them in the LTS and not risk basic compatibility on their "established" systems.

I think the idea has more merit than one might think. Saying that Windows and Mac OS do this and therefore Ubuntu (or any Linux) should as well doesn't carry much water. But the fact that Red Hat decided to update major apps such as Firefox and OpenOffice, along with adding new hardware drivers to the kernel, between point releases of RHEL 5.x shows that desktop users want the combination of a stable base along with up-to-date versions of the major apps they rely on for desktop productivity.

In short, I definitely think the Ubuntu LTS should move in this direction. As it stands now, the Ubuntu LTS is recommended to the casual/non-technical user as the stable choice, but I don't see all that much attention paid to the LTS for desktop users once the Ubuntu community moves on to the next six-month release.

It would be great for Canonical and the Ubuntu community to devote more development and testing time to the LTS over its life, and by making it easier for users of the release to stay a bit more current in terms of apps (while keeping the same base) seems to me to be a great way to make the LTS much more attractive to the desktop user.

That way if things blow up, as they have for me anyway, between six-month releases, then the LTS would be a more viable alternative. In moving away from the 8.04 LTS, I didn't need new hardware compatibility or new bells/whistles. What I needed was Firefox 3.5 and OpenOffice 3.1. I could've and probably should've looked into backports or PPAs, but it just seems so much easier to upgrade the whole distribution.

But it's not easier. Stuff breaks. I think a "rolling" LTS, would not only help users but would also encourage OEMs to install and support Ubuntu on the desktop. Seems like the proverbial win/win/win to me.

After a bit of time at sea, in the wilderness, at the bottom of a well, what have you, Dell looks like it's finally headed in the right direction when it comes to Linux- (and specifically Ubuntu-) powered PCs. And yes, Dell Zino (where do they get those names?) is primarily a Windows product.

Johannes Eva offers a very extensive post-installation guide on what to do after you install/upgrade Ubuntu 9.10 on your computer.

While I didn't want to personally do most of what was in this guide, I did pick up a few things that I either have done or will do in the near future.

On the first day of Karmic, my true love gave to me ... LET'S STOP THIS RIGHT NOW ...

I'll just present the link:

Christopher Tozzi of WorksWithU describes his first 10 days running Ubuntu 9.10 "Karmic Koala," along with pros and cons.

Brief gratuitous quote I pulled from the entry:

"In my experience, Karmic is far from an unmitigated failure. But it's not perfect, either."

According to WorksWithU (where I'm checking daily for Ubuntu developments, by the way), a Ubuntu One Music Store is in the works for the 10.04 "Lucid Lynx" release. Though nothing is confirmed in the slightest, speculation is that Ubuntu parent Canonical is working with Amazon to meld its own Apple-fighting music service into Ubuntu proper in order to fill one of the more cavernous gaps in the Linux app stack.

WorksWithU's Joe Panettieri writes:

I am reaching out to Canonical for a briefing on the Ubuntu One Music Store. But in the meantime I can tell you this: When my kids run Ubuntu on a range of netbooks and desktops at home there's only one "consumer" application they miss: Apple's iTunes. If the Ubuntu One Music Store fills that void, Canonical could successfully push Ubuntu deeper into the consumer market.

Normally a change in the automounting of USB drives in Ubuntu wouldn't be a big deal.

But in my case I've been using shell scripts to back up my Ubuntu box to USB drives via rsync. And before Karmic, those USB drives automounted with the name "disk" and woujld be at /media/disk/ in the filesystem. That was perfect for my shell script to target for the backup.

Now for some reason those drives are automounting not with the name "disk" but with the unique UUID number for the given device. At first this was bad, but after I modified my scripts, I actually see some wisdom in what started out as just another Karmic fail.

Not that I can't code around this, because I can (and I could probably put any name I wish in /etc/fstab ...) but should I have to?

Update: I've modified my shell scripts to rsync to the USB drives by their UUID-generated names instead of the previously given name "disk." The scripts work, and everything is back to normal.

It's another in the now-half-dozen things that have broken in the transition from Jaunty to Karmic. All were fairly easy to fix, but my idea of a successful upgrade isn't having to devote time to restoring basic functionality after an upgrade.

So where is the "wisdom" in this Karmic fail?

Well, now that each drive is being automounted and automatically given a unique name that happens to be the same as its UUID, that means I can further automate the backup process without much additional coding.

Previously I could've modified /etc/fstab to give each of these drives a unique name when they were automounted. Truthfully I never thought about it until now.

But now that each drive is getting a unique yet predictable name, I can plug both into the computer at once and do my two separate backups (first is everything but Thunderbird mail, second is only Thunderbird mail, and yes, I have too much Thunderbird mail) with a single script if I wish.

So from the jaws of fail comes the thrill of geekish victory.

This is the kind of thing I'd normally just tweet about, but the problem with Twitter is that if it's not doing anything for you, i.e. bringing you (or me) traffic, you're just giving away your time.

Hence I pretty much have relegated Twitter to promoting my other writing. It's gotta pay the freight. After 600+ tweets, I'm stalled at about 130 followers. If I had 1,000 or more followers, I might feel differently. But I don't. And I don't.

That's a long preamble to this:

The Tech-no-media blog asks, "Is Ubuntu Broken?" and basically calls for Ubuntu to call the proverbial spade a spade, and own up to the fact that the six-month releases aren't exactly stable.

One thing they called for, and I agree, is that the Ubuntu LTS release could benefit every once in a while from some new drivers in the kernel in the same way that Red Hat adds hardware drivers to the RHEL kernel in point releases.

I'm not sure whether or not it was mentioned in the post referenced above, but I've been reading about a change in policy for the next Ubuntu LTS, 10.04 (if you haven't cracked the Ubuntu release-number code, 10.04 stands for the fourth month of 2010, i.e. April, when the release is due).

Usually Ubuntu is built with packages pulled from the Debian Unstable branch, but for the 10.04 LTS, the Ubuntu developers are supposedly pulling from Debian Testing, which is more stable (and slightly older) than Unstable (or Volatile, for that matter) but less stable (and potentially a great deal newer) than Debian Stable (currently Lenny).

Whew!

All that means is that the packages in Ubuntu 10.04 LTS should be a bit more mature out of the box when the release is finalized in April 2010, and we should have a better-working LTS out of the box.

Or it could mean nothing. I'm no expert in how long it takes a package to move from Unstable to Testing, and I suspect that pulling from Testing = pulling from Unstable + 15-30 days, meaning that the potential is there for the Ubuntu LTS release to be more "stable" for the first month of release but just about the same after that, since patches from upstream tend to make their way from Debian to Ubuntu anyway.

I'll be upgrading at least one machine from the current LTS (8.04) to the new one when the time comes, and I always plan to bring the rest of my Ubuntu machines up to that LTS some time in the first six months after its release with a plan to use it for about a year before re-evaluating whether or not to stick with the LTS, move up to the next six-month release (as I've done now with 9.04 and 9.10), or go in a different direction entirely (generally to Debian Stable or Testing).

Here's my problem. I need a calendar app that rudely beeps to tell me when to go to meetings and such.

In Ubuntu, that means the Evolution mail client, which has an extensive calendar function, or so I'm told.

But I don't run Evolution. I use Thunderbird to manage my mail, and Thunderbird doesn't have a calendar function ... or does it?

Allow me to digress briefly: I first tried the Orage calendar app from Xfce, which I already have on this Ubuntu box because I have Xfce (but not the full Xubuntu) on it. But Orage, while working generally well for what I need it to do, for some reason is incapable of playing sounds to alert me to ... my alerts.

I did a bunch of Googling, checked bug reports. Nothing about Orage and a lack of sound in Ubuntu.

So I moved on.

I learned about Mozilla's Sunbird project, which is a full-fledged calendering app, and I also learned that there is a Thunderbird add-on called Lightning (Thunderbird and Lightning ... get it?) that brings Sunbird's calendar features to the Mozilla mail client.

Well, I downloaded the add-on, added it to Thunderbird ... and I was unable to create an event. Full stop.

So I backtracked. I removed the add-on and did what I should have done in the first place: I went through the Synaptic Package Manager and added the lightning-extension package, which brings along with it the calendar-timezones and calender-google-provider packages. (Presumably this means Google's calendar can somehow feed off of this ... I'll explore that later.)

I'll repeat for the West Coast audience: If you're running Thunderbird in Ubuntu, downloading and installing the Lightning calendar add-on from Mozilla won't work. Instead use the version in Ubuntu's repository.

Since I generally run Thunderbird all the time for my mail, having my calendar/alerts in there is the perfect solution.

Once I installed the three packages, I started Thunderbird. Right away the app asked whether or not I wanted to import my calendar settings from Evolution. Since I have nothing there, I declined.

Once in Thunderbird, I had Lightning. It works. I did a test event, sound worked, and I'm ready to start creating recurring events and alerting myself to their imminence (and/or eminence).

All this makes me think about the huge value we as users get from Mozilla. I'm waiting for music-manager/iTunes-killer Songbird to get better, and I'm already benefiting from Sunbird in the form of Lightning. ... and that's all on top of Thunderbird and Firefox. Very nice, indeed.

The quick version: To add calendar functionality to Thunderbird in Ubuntu, don't add Mozilla's Lightning add-on directly. Instead, add it through Ubuntu's own repositories, in my case using Synaptic to add the lightning-extension package and its dependencies. Then you'll be calendar-ready in Thunderbird.

The take-away: Don't want to use Evolution (I prefer a true cross-platform application for e-mail, and Thunderbird fits that bill very well) but want calendar functionality in your mail client? Thunderbird and Lightning seem to play well together.

Follow along: Developers of Sunbird and Lightning update things at the Calendar Weblog.

I've been contemplating an installation of Mac's OS X operating system on my old Power Macintosh G4/466 (466 MHz PowerPC CPU, currently 384 MB of RAM but a maximum of 1.5 GB on the motherboard).

That circa 2000/01 machine has been very happily running the PowerPC build of Debian Etch Linux for quite awhile now. But to truly be a work machine for what I do, I need to have Flash capability, and that's something that just isn't easy to do (and do well) on a PowerPC system not running OS X.

It's the tyranny of Flash as the predominant video format over the Internet. Flash is a proprietary system that is wholly controlled by Adobe, and both the apps that make Flash as well as those that display it are tightly controlled by this single company.

And while Adobe appears happy to code Flash players for Intel-based Linux, it is not so happy to do the same for other architectures in Linux (including PowerPC) as well as for other Unix-like operating systems such as FreeBSD (which uses the Linux version to some degree of success, as does OpenBSD, but in both cases on i386 only and not on PowerPC, which is what I'm aiming for).

I'm already running OS X on our iBook G4 laptop, and I figured that a "backup" OS X machine wouldn't be a bad idea all the way around.

I left the Debian Etch drives in the G4 (there is space for three drives on the bottom of the box, and I have one Debian "root" drive and another devoted to backups) but unplugged them and added a 40 GB IDE hard drive I pulled from a dying Compaq desktop a while back.

Once I figured out with the help of my Mac guru (and fellow LADN online worker) Tom Gapen that the OS X installer wouldn't even recognize my new/old hard drive until I used Apple's Disk Utility to put an OS X-recognized volume on it, I was able to continue beyond the first few screens of the install process.

Even then, I had to create the HFS+ volume in the Disk Utility and reboot before the installer would allow me to actually begin the installation.

Since this Mac is so comparatively old, it doesn't have an internal DVD drive. Being CD only, I used my OS X 10.3 discs to install the system. I figure I'd try later to hook up a Firewire DVD drive and upgrade to 10.4.

Once I had the disk-volume issue out of the way, I just let the installer run. The first thing it does is painstakingly check the first of two CDs for errors. Probably not a bad thing, but time-consuming.

I just let the installer run as I did other things, changed discs when needed, and then entered the barest of personal information when the system asked for it to make the first user account.

So I now have a relatively old Macintosh G4 with a single 466 MHz processor (we still have a few dual-500 MHz G4s in service at the Daily News; they run well with OS X 10.4) probably not quite enough memory at 384 MB, but a spare 256 MB module that I'll stuff in there as soon as I can (and the hope that I can scare up one or more 512 MB PC100 or PC133 modules to build it out).

I don't have the box connected to the Internet yet, and I'll have to load some software — especially the Firefox Web browser — in addition to first patching the OS X 10.3 installation and then upgrading to 10.4 and patching that ...

And if I can get networking into the box (I'm thinking either powerline networking or stringing some CAT5e from the home network to the box, which is far away from said home network), I will hopefully have a somewhat serviceable OS X machine on which to do work at home.

And if it doesn't work out, I can just unplug the OS X drive, replug in the Debian drives, update the Etch installation to Lenny and return to all the goodness that Linux has brought to this box in the recent past.

I have a certain, not-at-all-new entry in this Movable Type blog that has been attracting its share of spam lately. Frankly I'm not sure how it's making it to Published status. I could "turn up" the spam filter, but that tends to make every comment go into the spam pile, no matter whether it's a "real" comment or not.

One thing I could've don in this particular case, since this is a sign-in only blog in regard to comments, was to ban the specific account that is making the comment.

I decided to take a different tack. Since the entry was old. I went into it in the Movable Type dashboard and, under Feedback, unchecked the "Accept Comments" box.

No "real" readers are going to want to comment on it anyway, and since the spammers are focusing on this one entry for some reason, this should stop their vector into this particular blog.

Later: For good measure, I banned the commenter as well.

Canonical has been touting its Ubuntu One cloud-storage solution, which allows you to mirror up to 2 GB of files for free and up to 50 GB for $10/month.

The service also allows you to sync Tomboy Notes and Evolution contacts across multiple Ubuntu installations.

I gave Ubuntu One a try on my recently upgraded Ubuntu 9.10 system, and it appears that Ubuntu One just doesn't do very much that I need.

And whether it's the service's simplicity or lack of decent documentation, it took me awhile to figure out just how you get stuff synced with Ubuntu One.

It turns out you can't pick and choose directories on your Ubuntu box to backup/mirror across your various Ubuntu desktops (and right now I have only one system — this one — capable of using Ubuntu One).

Instead, to have files saved/mirrored, you need to drop them in the Ubuntu One folder/directory that is created in your Home directory. You drop stuff into that folder and the app mirrors it, where you can see the results at https://one.ubuntu.com

What I'd like to use Ubuntu One (or any other cloud-storage system) for is to back up the directories I select. I don't want to have to dump everything into a single folder. That would seriously mess with my file-management mojo.

And I don't use Tomboy or Evolution, so syncing those notes and contacts holds little appeal for me ...

And ... (you knew there was another and) I'd like to access these documents on other desktops that aren't running Ubuntu, or even running Linux.

The good news (and the biggest reason why I probably WILL use Ubuntu One at least a little bit) is that any client computer with a Web browser can access the shared Ubuntu One files at https://one.ubuntu.com ... this thing could be useful after all. I just uploaded a file into Ubuntu One from my Windows machine and did see that file in the Web interface on my Ubuntu computer. Once I remembered to "turn on" Ubuntu One by clicking on the little cloud icon in the upper GNOME panel, I received a notification that files were syncing. And sure enough, a few seconds later I saw the file in question in my Ubuntu One directory.

So for my purposes anyway, Ubuntu One is a somewhat useful FTP/sandbox area in which to share files across the many computers I use in a given week.

I suppose something like JungleDisk at this point is much better suited to the specific task of backing up to the cloud, but a free 2 GB is something that can come in handy.

Given the need to place items in the Ubuntu One directory in order to get them into the cloud, I really can't use the service as a true backup system. And if I exceeded 2 GB and had to pay the $10/month but didn't use the full 50 GB, I'd be paying more for Ubuntu One than I would for Amazon S3 storage in JungleDisk (although that doesn't factor in download/upload charges, just the monthly per-GB storage costs of about 15 cents/GB plus JungleDisk's $2/month fee).

I'd rather have a Ubuntu One that could pick/choose among my system's many directories what gets backed up/mirrored, just like I do with the rsync scripts I use to back up my system to USB drives.

I have a pretty good feeling that Ubuntu One is under heavy development (the Web site labels it as "beta"), and perhaps a more sophisticated way of both choosing what will be mirrored and even perhaps making the service function more as a backup application (i.e. allowing a user to archive files that aren't necessarily on the hard drive at all and keeping them in case of hardware/software disaster on the local machine) ... and syncing Firefox bookmarks, Thunderbird contacts and more just might be in the offing.

Again, there are a few services that compete quite well with Ubuntu One — including the Linux-compatible, very-much-cross-platform single-folder-syncs-all (and identically priced up to 50 GB) Dropbox, which looks a lot more built out than Ubuntu One at this point.

Still, a free 2 GB is a nice enticement for any given cloud-backup service. And watching what Ubuntu One eventually becomes looks to be a popular sport among FOSS proponents. I'll be one of them.

But right here, right now, the cross-platform nature of Dropbox makes it the better service for anybody who wants easier access to files from non-Ubuntu (and non-Linux) machines.

Later: I just took a look at the Dropbox blog, and that very blog subtly tells Ubuntu users that Dropbox is ready for them and has its own Karmic-specific repository that they can add to their /etc/apt/sources.list for automatic installation and updates of the Dropbox client software.

This still isn't quite what I want. Complicating matters is the fact that a) I don't exactly know what I want, and b) I don't know what's possible on the currently offered cloud-backup services.

I do know that for the most part I want a certain subset of my overall files synced across all of my desktops, but I also want another subset of files archived and accessible from all desktops yet not synced to them. And then there's the problem of stored e-mail. I have about 2 GB of mail in Thunderbird right now, all POPped down to my main PC. I'm not anxious to go back to IMAP, but the overly large files that Thunderbird's mbox format creates are laborious to back up and potentially hard to sync (OK ... I really don't know how easy or hard it would be).

At one level, I think all of this would be much easier if I finally pulled the trigger and piped my mail through Gmail, using POP through Thunderbird as a secondary way of dealing with that mail. That's another overly boring geeky tale for another overly boring geeky blog post.