June 2011 Archives

I did a quick/dirty installation of OpenBSD 4.9 (i386) to the Lenovo G555. I didn't install to the hard drive, but instead to an 8 GB USB flash drive, which provides far from optimal performance but allows me to test hardware compatibility.

That and the fact that OpenBSD can be installed in about five minutes makes it easy to test new releases.

I had been intrigued by a line in the release announcement that said:

Support for Mobility Radeon HD 4200 has been added to radeondrm(4).

Hey, that's my graphics chip in the Lenovo G555! Not that I was having any trouble with video in OpenBSD over the past few releases because I wasn't.

But any individual attention the the old HD 4200 is much appreciated.

I also wanted to test suspend/resume, a feature the OpenBSD developers have been improving from release to release.

My tests of OpenBSD 4.9 on both the Toshiba Satellite 1100-S101 (Intel 82830/830m) graphics) and IBM Thinkpad R32 (ATI graphics of some kind) had suspend/resume not working. But Linux suspend/resume doesn't work on these machines either.

But on the Lenovo G555 (AMD Athlon II 2.1 GHz CPU), Linux suspend/resume does work. I'm enjoying it now in Debian Squeeze (and after all the time I've spent having it not work on machines, enjoy is most definitely the right word).

So I loaded up OpenBSD 4.9, installing from a CD to the flash drive.

Once I booted into the new system, I turned on apmd as root:

# apmd

Then I issued the suspend command as my user:

$ zzz

Wouldn't you know it, the thing suspended.

I tapped the space bar and it resumed. Success.

Again, OpenBSD from this particular 8 GB USB flash drive on this particular laptop is not a pleasant experience (nor is running any other system booted from said flash drive, and I've done this with Debian and Ubuntu in the recent past).

And setting up a fully functional desktop system for my needs in OpenBSD isn't something I can do in an hour (or a day).

And ... (you knew there was another "and"), how can I give up on perfectly working Debian Squeeze no matter how much I want to run OpenBSD again?

I'll probably go there at some point, either with this laptop or another slab of hardware. I need to do a full GNOME desktop to see if the gam_server problem I'm having in Jggimi's live disc manifests itself in a "real" install (that problem being gam_server eating 80-99 percent of CPU on a continual basis).

But now I know that perfect graphics and working suspend/resume can be mine in OpenBSD. It borders on the profound (and no, I'm not exaggerating).

Mozilla has number-creep on the brain.

After dealing with Firefox 3.6.17's abrupt end of life in favor of 4.0, and then 4.0's deprecation in favor of 5.0 (and yes, I had to change repositories every time because Linux in general and Debian in particular doesn't force new software on users), now I learn that Thunderbird is jumping from 3.1 to 5.0.

I use the Debian Mozilla Team APT archive for all my Mozilla software needs (Firefox/Iceweasel and Thunderbird/Icedove). Now I'll be dipping into my sources.list files to up the Icedove repo from 3.1 to 5.0 once the Debian Mozilla Team offers the new version (they generally need a little time to package it up).

Thunderbird/Icedove remains my No. 1 e-mail client, and 3.1 brought features I wanted/needed that weren't in 3.0 (principally the Quick Filter functionality), so I had to make the move from the standard 3.0 in Debian Squeeze. Now I'll be on the 5.0 bandwagon.

Why am I tracking things so new on a base so stable (that being Debian Squeeze)? I believe that's the best way to do it: Run a stable base if your hardware is happy with it (and mine is ) and selectively roll in new versions of applications you need to be new from added repositories.

Fortunately or not, for me that includes the Linux kernel, since I need a post-2.6.32 kernel to solve sound issues on my Lenovo G555 laptop. Otherwise I'd ride 2.6.32 as long as I could.

Here are the apps I'm using newer versions of, including the original (in parentheses) and current-to-me (following, not in parentheses) versions:

• Linux kernel from Liquorix (2.6.32) 2.6.38.7 (I'm sticking with this one for awhile, even though 2.6.39 is out)

• Firefox/Iceweasel web browser from the Debian Mozilla Team (3.5.x) 5.0

• Google Chrome web browser from Google (Chromium 6.0.x) 12.0.x

• Thunderbird/Icedove mail client from the Debian Mozilla Team (3.0) 3.1

• LibreOffice from Debian Backports (OpenOffice 3.2.x) 3.3.2

That's it. I also have Dropbox, which I installed with a Linux Mint Debian Edition package.

Otherwise I'm running stock Debian Squeeze.

I'd like to try the newer gThumb, but I've got enough functionality in the Squeeze version right now. Still, I'll probably pluck the package from Sid when 2.13.2 rolls in.

ZaReason, the company that sells free-OS-running desktops, laptops and servers has a new netbook, the Teo Pro, that at $399 looks mighty good.

And I'm glad to see that not only will ZaReason ship you a computer running Ubuntu and derivatives Kubuntu, Edubuntu and Kubuntu (current release or LTS), you can also choose Linux Mint, Debian, Fedora or, for the purists, "no OS," and install your own.

Looking for a desktop system? Right now I like ZaReason's Media Box 4220, which -- befitting its media emphasis -- can be shipped with all of the OSes named above plus Mythbuntu so it can function as a media-PC right out of the box.

Man, that Google doesn't quit. First they tried to topple Facebook and Twitter with the privacy-challenged Google Buzz, then it tried to redefine Web communication with the resource-hogging, poorly understood Google Wave.

Those didn't exactly perform. But Google's no quitter.

Today they announce Google+ -- with a simple slogan: Real-life sharing, rethought for the web.

Sure it's another corporate monolith looking to grab your data, but unless Diaspora scales up a whole lot faster than it is doing right now, having a Google-ian counter to Facebook sounds like a pretty good thing to me. (Note: I do have a Diaspora account but am unsure exactly what to do with it ...)

Getting back to Google+, here's more from Google:

Sharing is a huge part of the web, a part that we think could be a lot simpler. That's why we've been working on adding a few new things to Google: to make connecting with people on the web more like connecting with them in the real world. We hope you like what we've cooked up so far. And stay tuned, because there's more to come.

Like Gmail back in the day, Google+ is at invite-only status right now, though Google is happy to take your e-mail address to keep you informed as to when you can get a Google+ account.

To find out way more, go to this cool slide-showy thing from Google.

Things I see right away:

The "Circles" concept, in which you share different things with different people (like in Diaspora but unlike Facebook, where everybody's a "Friend" whether you're sleeping with them, or you were on yearbook together in high school)

Hangouts, that let you gather people together for group chat. I really like this.

"Sparks," to find things you're interested in to "follow," Facebook style.

"Huddle" turns separate texting conversations into group chat. I'm not sure how this differs from Hangouts, other than that it happens in a smartphone interface.

This sounds to me like Google Wave refined -- an attempt to stitch together the various kinds of online communication we have in order to make the often disparate parts if not more than their sum, at least more cohesive and useful.

Right now, sans invitations especially, there's no "there" there. This could be the next Wave (in the bad, nobody can figure it out, and it eats CPU for breakfast way), or a new, compelling paradigm for social interaction over IP.

More on Google+

Inside Google+: how the search giant plans to go social by Steven Levy (A very long article by the Wired writer, well-worth reading)

I'll chalk this up to "nothing to see here," but this story is going around about the nD, a not-here-yet handheld gaming device that's supposed to rival the Nintendo DS in speed yet cost only $10.

The nD is said to be based on a 400 MHz ARM processor, include a 2.4-inch backlit screen with 320x240 resolution, 16 to 32 MB of RAM and 2 GB of internal storage. It is supposed to run a custom, embedded Linux OS.

The nD has its own website (which includes this news page) and I imagine a lot of people will be keeping an eye on this one.

As you can see above, they do have a video (or two).

I didn't have any trouble with the functionality of OpenOffice as it stands in Debian Squeeze.

But since the project is generally in turmoil, with the entire Linux world shifting its focus and support to the new Document Foundation's LibreOffice fork, that coupled with the arrival of LO in Debian Backports, I decided to make the switch on my own Debian Squeeze laptop.

Despite some warnings from readers of Life, the Universe and Debian, the whole process went very smoothly, and I now have LibreOffice running on my Squeeze machine.

More about my LibreOffice installation from Life, the Universe and Debian:

• Installing LibreOffice in Debian Squeeze from Backports without a hitch (With full text of the terminal session in which I did the installation)

• LibreOffice is now in Debian Squeeze Backports

I finally cleared out a drive for a test machine (my IBM Thinkpad R32), but this entry isn't about that. (It's running Debian Squeeze with LXDE today after a day with Debian Squeeze with GNOME. I might try Linux Mint LXDE or Fluxbox next ...)

What it is about is my main Debian Squeeze laptop, the Lenovo G555 (it's no Thinkpad).

I already have a newer kernel (from Liquorix), newer Iceweasel and Icedove, Google Chrome and Debian Multimedia.

Now I want/need/have to have gThumb 2.13 (they added sharpen!!!).

I could build it myself. But instead I'm going to use apt-pinning to pull the newer gThumb from either Debian Testing or Unstable.

The Debian Project is so non-commercial, they don't sell T-shirts, although some are available ... in Europe ... at conferences.

But true to the Debian way, the art needed to make your own shirt is open source and generally available.

So I decided to make my own Debian Squeeze Space Fun (the theme for this release) T-shirt. I could have gone with Cafe Press or Zazzle but decided to try SpreadShirt.

The completed shirt arrived today. It cost about $20 (higher than it would if I had printing on only one side). Here is what it looks like:

and the back:

In the interest of keeping this "project" open source, here's how I did it:

Note: I'm no expert on preparing images for printing on T-shirts, and one thing that made this particular project more complicated is that the color of the shirt is part of the design. That means the artwork wouldn't work on a white T-shirt.

I started with the art for Debian Squeeze, available on the Debian Wiki.

The T-shirt design actually shows a whole shirt, front and back, with designs wrapping around the body of the shirt and the arm. That looks pretty sweet. I'd buy it. If somebody was selling it, which so far doesn't appear to be the case.

Luckily the sources for these designs are freely available. Start here for everything, or here for the .svg and .png images for the T-shirt and other images, one of which I used for the back of the shirt.

Since SVGs (or scalable vector graphics) are not bitmapped images and can not only be reduced in size but increased in size with no loss in quality, they are perfect for printing large on a T-shirt.

My method was to use Inkscape to grab the portions of the image that I wanted (i.e. everything but the blue T-shirt itself, since I wanted to print on a T-shirt, not a picture of a T-shirt itself).

If I did this again, I would do some more research on the exact sizes of images allowed for the various instant-shirt-printing companies and make sure to product that size either out of Inkscape or in my eventual bitmapped image out of GIMP.

Getting back to the recipe, Once I selected all of the elements I wanted to keep from the Debian Space Fun T-shirt SVG file in Inkscape, I saved that .svg.

You can convert an .svg to a bitmapped image in Inkscape, but I prefer to do that in the GIMP image-editing application.

So I opened the .svg in the GIMP, then converted to a .png image, being sure to discard the background (so the T-shirt printer won't have a big white background behind my desired image, which happens to be mostly white anyway, and printing white on white wouldn't work at all).

Once I had my .png images ready, I "built" the T-shirt in SpreadShirt's Flash-driven application, choosing the style and color of shirt I wanted. I could've picked a T-shirt that was $5 cheaper and maybe should have, but the results are pretty nice.

Original Debian Squeeze Space Fun t-shirt SVG image (used for front of shirt):

http://svn.debianart.org/themes/spacefun/others/t-shirt.svg

Original Debian Squeeze Space Fun promo image SVG (used for back of shirt):

http://svn.debianart.org/themes/spacefun/others/banner-web.svg

Here are my images in SVG format (elements pulled from original SVGs with Inkscape):

Debian "Space Fun" swirl (front of shirt)
"Get" Debian Squeeze promo (back of shirt)

Here are the PNGs I made in GIMP from the SVGs:

Debian "Space Fun" swirl (front of shirt)
"Get" Debian Squeeze promo (back of shirt)

I'm no expert on these instant-printed T-shirts. Looking at the shirt now that I have it, I would have liked to use art that was more "solid" to get a clearer, whiter white onto the fabric, but overall I think it looks pretty good.

If anybody out there has tips on how to do this better, please let me know.

To improve this particular shirt (the one you see in the images at the top of this post), I would've liked to make the images bigger (and don't quite know how/why they came out relatively smaller). I'm not sure the type in the "banner" image on the back was the best for printing on fabric.

I was aiming for this shirt from debian.ch (and still wonder why I can't just buy it ...):

The free, open-source office suite OpenOffice has been dying by inches for quite some time. Now that Oracle has "given" the project -- sans any kind of monetary or developer support, it would seem -- to the Apache Foundation, I'd say it's time to close the lid on the box and drop it in the ground. It's not like Oracle didn't kill it first.

After all, Oracle succeeded in alienating OpenOffice's developers and users to such an extent that the project forked, with all the geeky air in the room going to the newly minted Document Foundation and LibreOffice, which took the free (and yes, open) OpenOffice code and began a project with no corporate overlord and a pledge to include rather than exclude, innovate rather than the other thing.

I was never sure just how strong Sun's commitment to OpenOffice was in the years and months before Oracle swallowed the SPARC-centric hardware maker, and I wondered why Sun needed an office suite -- other than to be a showcase for its Java programming language.

That Java made OpenOffice bloated and sluggish was (and is) another story. The barriers (real or imagined but probably more real than not) to code contributions outside Sun didn't help. Nor did the rise of Google Docs and other online solutions pushing OpenOffice over to eat their own share of Microsoft Office's lunch.

I guess Oracle thought the same thing. They ignored OpenOffice and its contributors after buying Sun. Sure they killed OpenSolaris first. It was only a matter of time before they ankled OpenOffice.

As soon as LibreOffice was born (or made, "Frankenstein"-style, with the guts of OpenOffice), all the major (and likely most of the minor) Linux distributions began subbing LibreOffice for OpenOffice and pledging to support the Document Foundation and its new endeavor.

(Read the Document Foundation's June 1, 2011 statement on Oracle's donation of OpenOffice to the Apache Foundation. Meanwhile, OpenOffice.org hasn't issued a news item since April 12, 2011)

Not that I have any statistics to back this up -- because I most certainly do not -- but I have a pretty good feeling that the Linux/Unix users of OpenOffice/LibreOffice are more about mindshare than market share. It's on the Windows desktops at corporations and nonprofits, in educational institutions and businesses of all sizes where the majority of OpenOffice users can be found.

LibreOffice is already doing what it should be doing to gain the advantage -- welcoming new contributors, pledging to be way more open than OpenOffice ever was, adding features and cleaning up the code.

Can OpenOffice, without any Oracle help, get back on its feet? It doesn't look good.

And now that LibreOffice is a going concern, it doesn't really matter.

Not being beholden to Oracle is certainly a plus for the Linux distributions. The biggest thing you can tell potential converts from Windows (or even Mac OS) is that there's a free, full-featured office suite waiting for you on the other side. Ubuntu, Debian, OpenSuse -- they've all lined up behind LibreOffice. That didn't take long.

But Linux and BSD users are mindshare not market share. We're not even a blip compared to the number of users of OpenOffice and perhaps now LibreOffice on Windows machines around the world in businesses, educational institutions, nonprofits and other unclassified-here entities large, medium and small.

Not everybody wants to steal Microsoft Office (or has the dubious skills to do so). Not everybody wants to run Microsoft Office.

And while I have no statistics to back me up, I know that adoption of OpenOffice in Windows is huge.

So can OpenOffice continue to covet this market, or will LibreOffice become the new go-to free office suite (if Google Docs isn't already holding that title)?

Without Oracle's help, I doubt OpenOffice as a project can go forward. Unless I'm missing something, development will halt as the remaining members of the OpenOffice team -- whoever they may be -- try to regroup (or "group," as it were).

And LibreOffice is going (full steam ahead / like gangbusters / like no tomorrow / insert your favorite cliche).

LibreOffice has flung open the doors to new contributors in a way OpenOffice never did.

Development is proceeding at a rapid pace. New features are arriving. Dead hunks of the code (they're everywhere I hear) are being lopped off.

I've already installed LibreOffice on a few Windows machines. My frozen-in-amber Debian Squeeze laptop still runs OpenOffice -- LibreOffice has been packaged for Debian Wheezy and Sid but not for Debian Backports.

And at this point OpenOffice and LibreOffice are close enough that for my casual-use case I'm fine remaining with the OO my Debian system offers. For now. Sooner rather than later LibreOffice will come to Backports, or I'll pin it from Wheezy.

I do a lot of work in LibreOffice in Windows. And while it's seamless to go between LibreOffice and OpenOffice at this point in both projects' development, that won't be the case forever.

Ubuntu, OpenSuse and Fedora already ship LibreOffice. I bet Debian Wheezy, when it gets closer to becoming the project's Stable (capital "S") distribution, will do the same.

And mindshare is a very funny thing. In my role as office geek, I've done a half-dozen LibreOffice installs. Why would I roll OpenOffice onto a box when OO appears to be dead?

While many have a soft spot (or, shall we say, less hard spot) for Sun Microsystems -- I still have my Sparcstation 20, but you're welcome to it if you want to come in here and haul it away -- very few geeks on the ground have anything resembling warmth regarding Oracle. You've got your indifference. Your outright hatred. But warmth? No.

Sure I use Google Docs, but now with Dropbox syncing my files between my Windows and Linux machines, I'm using "local" applications more and more. That includes OpenOffice and LibreOffice.

LibreOffice has the momentum. I'm sure the project's leaders would like to get the OpenOffice name back and continue on with the brand recognition that the formerly Sun-driven office suite has enjoyed for years.

But as the months pass, name recognition will mean less and less. More users and more admins will learn about LibreOffice. And LO will only get better with all the developer it's bringing to the table.

Sure I've seen my share of OpenOffice 1.1 installations still going not-so-strong on Windows PCs. (Hint: An in-place upgrade mechanism for Windows would go a long, long way.) I can't see the admins of today -- or tomorrow -- sticking with OpenOffice when LibreOffice is going so well.

But as much as I'm reading the last rites for OpenOffice, it's not dead yet. There's still an OpenOffice 3.4 beta out there. But the OpenOffice mailing lists don't look terribly active. Not a good sign.

(In case you're wondering, the LibreOffice mailing lists are very active.)

OpenOffice could subsist by incorporating the open-source improvements being made in LibreOffice, just as LibreOffice started with the free, open-source OpenOffice codebase.

I'd say there's a better chance of The Document Foundation, one way or another, getting the OpenOffice name. Whether it'll be of any use depends on how long that takes. Once LibreOffice gets its legs -- and I'd say it already has -- the OpenOffice brand could be no more popular than MySpace in the Facebook era.