Consistent growth – Not

Lavrusik argues that Twitter has always had steady growth, and will continue to do so. To illustrate this, he links to a Compete graph, which shows that Twitter grew about 20-fold in just one year. There’s almost NO negative growth, as well.

But here’s the problem. Twitter has really only hit the mainstream at the end of 2008, which is about when the Compete graph ends. Let’s look longer term, kindly provided by Alexa. Alexa shows that Twitter really starts taking off in 2009 – in 2008, growth was almost zero. A Google News timeline seems to illustrate this as well.

If you follow the Google News stories, you’ll see that most of the early stories from the mainstream are mostly articles about celebrities X Y and Z using Twitter. Additionally, Google Trends seems to indicate the first real spike followed the famed Ashton Kutcher-CNN battle. In other words, the data seems to indicate that users are added as celebrities begin using Twitter. That means there’s a limit: there’s only a certain amount of celebrity followers in the world. At a certain point, they start to overlap, as well. Additionally, most celebrities have their flunkies do all the updating, rather than post the blurb themselves. Finally, a lot of the celebrities followed are politicians. Without an election for another three years, most mainstreamers don’t really care about Fred Thompson or Dennis Kucinich. Therefore, to say that, simply because Twitter grew like crazy in eight months, Twitter will continue to grow forever.

Gaining approval from big companies – Really?

Lavrusik also argues that the big companies are using Twitter. That’s true – up to a point. Sure, tons of companies are joining Twitter. And some of them, like @zappos, do a really good job. But then there are companies like @amazon, who basically use Twitter as another outlet for their RSS feed. Just because companies “do” Twitter doesn’t mean that there are ANY advantages to visiting a Twitter feed rather than reading an RSS feed.

Twitter is the number one most social brand – Wrong

Lavrusik makes the argument that Twitter is the most popular social brand, based on this study. Wrong. Actually, Twitter is the most popular social brand mentioned most frequently on the Web during July 2009. Because of that, the study overlooks one very important alternative: Facebook. NO ONE links to Facebook. Why would you? Most of the stuff on Facebook is private anyway. On the other hand, everyone can link to Twitter because almost everything on Twitter is public. However, that in NO WAY means that Facebook isn’t more popular than Twitter. And despite this, Facebook is still #4 (beating OBAMA, for crying out loud!). The next generation is turning to Facebook, not Twitter, as the next social brand, as Matthew Robson argues.

Twitter receives millions in free media coverage – So?

Lavrusik points to an Ad Age study ($$$) saying that Twitter receives more free advertising than Bing will. So? This really goes back to my first point – people who joined because of celebrities can’t sustainably fuel a company for more than a few years.

Continued growth internationally – Unstoppable?

This is where Lavrusik and I do agree: for now, Twitter is incredibly handy for getting users in other countries to share news about oppression. But here’s the problem (or rather, the blessing in disguise for the rest of the world): these criseses don’t happen every day. You really can’t base your product’s growth on criseses, unless you’re the military. Twitter can’t tell business investors that the only reason they’re continuing to grow is because they’re used in crisis situations. It’s a great tool, but in its current form, it’s not sustainable.

Conclusion

Will Twitter continue to grow short-term? Duh. Will Twitter continue to grow long-term? Maybe, but if it does, I doubt it will be solely because of any of these reasons. Of course, the really important question: Will I continue to use Twitter? The answer: Duh, heck yeah. Whether in good times or bad, find me at @andrewmin.

The problem: current adoption methods

Right now, a popular Linux evangelism idea is to force Linux as a pre-install option onto retail computers and wait for the market share to magically shoot up. With the exception of the niche netbook market, this really hasn’t happened (and the netbook market share happened in-part because of customized netbook-specific Linux distributions). What’s going wrong?

One problem is that Microsoft subsidises Vista. They hand out cheap copies at bulk discounts to companies like Dell and HP. Then, you add on the marketing crapware like the 90-day McAfee trial or the infamous 999 minutes AOL trial, and the price goes down even more. That’s why you’ll often see Vista computers at a cheaper price than Linux computers, especially at large retailers like Dell.

A second (and much harder to combat) problem is the lack of knowledge about Linux. No one actually realizes that Linux has apps that will handle most file formats around. In fact, most people still think that file formats are tied to operating systems. Heck, most people still think that Windows = IBM. No wonder everyone’s scared of Linux.

The solution to the problem: gamers

Gamers are adventurous folks. That right there is a positive sign. Linux adopters often need to be adventurous in order to even install a new operating system. But even better, gamers often build their own computers, either from scratch, a barebones kit, or a stripped down retail box. And as I pointed out above, what is often in the top 3 most expensive items on many gaming computers is the Windows Vista retail CD, ranging from the $214 Home Premium to the $249 Ultimate Edition. Gamers, therefore, are a ripe target for the open source community.

The problem with the solution to the problem: no games

Unfortunately, therein lies the problem: gamers need games. Sure, Linux has some decent games like Sauerbraten, Enemy Territory, Tremulous, Wesnoth, and Unreal Tournament. But what else? Civilization: Call to Power (a 1999 clone of Civilization). An old, unsupported version of America’s Army. X-Plane, a flight simulator. The original Medal of Honor. Just about the only recent brand-name games released are Savage 2 and EVE Online. Meanwhile, most of the free games on the market, built by volunteer developers in their spare time, have graphics that remind some of the original NES or are playing catchup to games created five years ago.

The solution to the problem with the solution to the problem: Linux companies

Here’s what Linux distributions need to do to even have a shot at taking a few percentage points of the market share. They need to start sponsoring companies like the old Loki Software and Ryan Gordon’s subsequent efforts. And they have a chance now. Steam, the popular game store, is possibly thinking about porting games to Linux. Companies like Canonical, Red Hat, and Novell would do well to sponsor some of that work.

In addition, Linux companies can start doing their own in-house porting. Canonical has a huge community of willing developers. Certainly, not all of them would feel comfortable porting proprietary code to the Linux platform. But some would. And this is very feasible: many Linux companies have lots of experience working with commercial developers. Canonical even has its own commercial repository for its “partners”.

Finally, Linux companies also need to start paying attention to the open source gaming community. Why? It’s lacking. However, gamers can get excited about free games. They just have to be up to par with commercial games. The problem is, commercial companies pay hundreds of employees to build a game for several years, while many competing gaming projects only last several years before the developer moves on. It’s time for open source developers to start getting paid for their jobs. Who better to pay them than the companies that benefit most?

Desktop Look

The first thing to do is to get Ubuntu into the Christmas spirit with some great themes and styles. Gnome users have it easy: all you have to do is download and install the Christmas Theme, courtesy of TheeMahn. It is unfortunately a little hard on the eyes (especially the green version), so I wouldn’t do this on a productivity machine. You can also get the Xmas Theme by Keith Hedger. It’s a lot easier on the eyes, but certainly not the most elegant theme in the world. I much prefer Fred McKinney’s Yulebuntu, a GTK+ style and color theme. The author also includes recommendations for a matching Metacity and icon theme. Emerald users may prefer using Amethyst Plum’s Christmas Emerald theme (Compiz and Emerald, of course, are required). It has a great holiday feel, and it actually looks nice as well. Speaking of Compiz, you can also get a nice Christmas 3D Desktop Cube (created by Rodolfo Ramone) for your Compiz or Beryl cube. It comes with wallpaper and two awesome-looking skydomes. It was originally meant for Beryl, but you should still be able to tweak it to work with Compiz Fusion.

KDE users have it much easier. They have a beautiful little theme called ChristmasTime (created by Eazy Eazysson), though it’s unfortunately only available for KDE 3.x. It comes with all the standard window decorator, icon, KDM, Kicker, KBFX, and mouse cursor skins, plus a Rainlendar skin and several SuperKaramba skins. The author is hinting that a KDE 4 version might be in the works soon. That could be enough to make me switch back to KDE.

Finally, the wallpaper. There are tons of Christmas wallpapers around, but there aren’t many Linux-specific wallpapers that look great. Good distro-agnostic wallpapers include acryline’s Christmas Tux and VaeloStudio’s Christmas Tale. Fedora users will enjoy mola’s Fedora Christmas, while Ubuntu users will most likely want to download kane77’s Ubuntu Christmas wallpaper, Patrick’s Ubuntu Christmas Remake (no relation to kane’s), or acryline’s Christmas Tux Ubuntu (a remake of the original Christmas Tux with Ubuntu and Debian-specific wallpaper).

Widgets

Now that your desktop looks Christmas-y, it’s time to gear it up with some holiday desktop toys. xsnow is probably the oldest example of this. Created way back in 1984 by Rick Jansen, this is a very simple X Windows app that lets snow fall down on your windows. It’s available in most distributions’ repositories, and is even bundled with a few. Just install it, make sure the root window is shown (see the README for more), and run it. You’ll even get the bonus of seeing Santa go by with his reindeer.

Unfortunately, xsnow was made back in 1984, and many components, especially the graphics, haven’t been updated recently. There’s more elegant solution available: the Snow plugin for Compiz Fusion. It’s currently unsupported, so you’ll most likely have to install an extra package or two to get it (Ubuntu has it in compiz-fusion-plugins-unsupported). Enable it in CompizConfig (look in the Extras section), and then launch it with the hotkey (default is Super-F3). Elegant snow will begin to fall. You can also change the snowflakes to any other image you want (for example, hot air balloons) by playing around with the configuration options.

Sometimes, it’s better to have the snow falling during the screensaver. That’s where fuzzyflakes (created, apparently, by Barry Dmytro) comes in. It’s an xscreensaver that comes with the xscreensaver-data package, so it can work with both the Gnome and KDE screensavers as long as they have the xscreenserver backend installed. Enable it, and fuzzy snowflakes come down on a pink background. Perhaps not the most Christmas-y screensaver around, but the closest you can find.

Finally, a little something for the KDE inside of me. I’ve always been a fan of SuperKaramba widgets. If you are too, you’ll love the Christmas Tree SuperKaramba theme by chichovoto. It simply displays a big Christmas tree on your desktop. When you double-click on it, it will play Jingle Bells in xmms or Amarok. It’s probably the least useful app… but it’s fun to show off.

Games

One of the things that happens around Christmas time is that you get lots of fun gifts. These include games. So, why not install a bunch of Christmas games for your Linux system?

If you were hoping for the Wii and failed to receive it, you can attempt to console yourself with SuperTuxKart. Sure, the graphics might not be up to par with the Wii (or the GameCube, DS, Nintendo 64, or possibly even the NES), but it’s basically the same gameplay as the old Mario Kart games. Based on Steve and Oliver Baker’s original and now defunct TuxKart, SuperTuxKart is a 3D cart racing game starring (you guessed it) Tux the Penguin. There’s even a snow track, hence the snowy Christmas theme lives on.

If SuperTuxKart didn’t cut the Nintendo cake for you, SuperTux will. Believe it or not, it really has no relation to the above SuperTuxKart, except for an overlap in characters (both games feature Tux and his girlfriend Penny). It’s very similar to the classic jump-’n'-bump Super Mario games, but rather than a boring old plumber, SuperTux features, well, the Linux mascot. It’s still a very much work-in-progress, but it’s a fun timekiller.

If you love arcade games but are getting sick of the whole copy-Nintendo theme, Extreme Tux Racer is a great game. It has a long history: it’s the fork of PlanetPenguin Racer, which in turn was a fork of OpenRacer, which in turn was a fork of the commercial (and abandoned) Tux Racer. Despite these complications, Extreme Tux Racer is quite a simple game. You just slide down a hill, eating as many herrings as possible.

If you’re just plain sick of the entire arcade game idea, you might be more interested in a real time strategy game. Joey Marshall’s SnowballZ is the perfect Christmas-y answer. The concept is great: just have your penguins collect lots of fish to feed your army without getting killed by enemy snowball-throwing penguins. Sounds simple? Ha! Think again. And once you discover the LAN, you won’t be able to stop.

Note: the author is not responsible for lost jobs, angry spouses, or any other damage incurred by the above game.

If you like using your brain, it might be time for some puzzles. Look no farther than Guillaume Cottenceau’s Frozen Bubble, one of the most ultimate puzzle games ever. Simply shoot colored balls (using, of course, penguins as your operator) at similarily colored balls. Putting like colors with like colors will cause all contigious colors to drop. Unfortunately, it’s not as easy as it sounds. And yes, this is even more addicting than SnowballZ.

Of course, if you’re really feeling brainy, there’s nothing better than science to satisfy you. That’s where Snowflake comes in. Created by Don Johnson (from security pioneer Raph Levien’s original idea), this program, originally a cryptography exercise, generates unique snowflakes using bit patterns from ASCII strings. It’s very nifty, even if you’re not a security freak. Please note that Don’s page has since had technical issues, and you’re better off trying to look for Snowflake in your distro’s repository.

Now that your brain is fried with penguins, bubbles, and snowflakes, it’s time for a little relaxation. My favorite method of relaxation is playing a good old first-person shooter, and I can’t think of any better Linux-based shooter than Wolfenstein: Enemy Terroritory. And for Christmas, the perfect map is North Pole, created by Rikard Lindgren and “Loffy”. The Allies’ mission is to defend the “Christmas Star”, a symbol of hope and joy, from the Axis invaders. You’ll see Santa, reindeer, and, of course, the North Pole. Awesomeness.

Note: I am not a Google employee. I have hosted several of my apps on Google Code, however.

The Log-in Requirement

The author makes the legitimate point that a Google login is required. But my response to that is simply, “who cares?”. Yes, all your search queries and so on are tied to your email address. But that’s how many projects work. Bugzilla, for example, ties everything to your email address. If, on the other hand, the complaint is that you are automatically logged into other services, what’s the big deal? If you don’t like them, you don’t have to use them. You can also use a tool like CustomizeGoogle to annonymize yourself if you’re really that worried.

The Lack of Freedom

The argument here is that Google Code is closed-source, which is bad, and automatically locks you in. Yes, Google Code is closed-source. I have news, folks. So is Sourceforge (it runs SourceForge Enterprise Edition). So (as far as I can tell) is Launchpad. And as for the lock-in, that applies to any host, open source or not, unless of course you host it yourself. And even if you host it, you might be in for a headache.

The License Favoritism

This is really where I have the least amount of objections. The AGPL isn’t allowed on Google Code, and it’s odd that Google hasn’t explained why. Still, I’ve never personally licensed anything under the AGPL (the GPL is good enough for me), so it really doesn’t affect me. As for the argument Google is just protecting itself, this makes absolutely no sense. A piece of software is more free when you can do whatever you want with it.

The Lifetime Limit

There’s a simple piece of reasoning behind this: Google can’t host every single hello world application ever created. It doesn’t have the space for it. And they are reasonable about getting more hosting.

Use Something Else

Great in theory, horrible in practice. Google Code is the cleanest and easiest to use project host I’ve ever seen. OK, maybe Launchpad is cleaner, but not by that much. Sourceforge has tons of features, but it’s ultra cluttered. Since GForge is based on Sourceforge’s old code, Roman’s suggestions of using Alioth and GForge still has the Sourceforge clutter problem. Heck, I spent ages figuring out that you needed to SFTP to your Sourceforge account in order to create a webpage. And I’m a techie guy.

Run Your Own

Great if you already have hosting. But I’m not going to pay to host an open source application when I can just use a free host. Doesn’t that kinda defeat the purpose of being “free”?

Conclusion

I do want to note, however, that this may not apply to everyone. Some may favor their freedom over everything else. Me, I’m a pragmatist, and most of the idealistic points don’t really make sense in practice, to me at least. As for me and my apps, we shall use Google Code.

LacKing in appliKations

Even while I was a KDE user, I was frustrated by the apps that Kubuntu came with. Amarok was awesome, to be sure. K3B was pretty darn awesome. And KDE’s plasmoids are a ton better than Gnome’s Screenlets. But that’s about the limit. Pidgin’s interface was so much more clean than Kopete’s. Strigi made 10GB indexes of my 80GB hard drive. Konqueror was lousy at loading webpages. KDE email notifiers drove me crazy (I resorted to CheckGmail). And even the highly-touted KOffice paled in comparison to OpenOffice.org’s friendly (albeit slightly foreign) interface.

Of course, most would ask why I didn’t just use Gnome apps in KDE. It might just be me, but it seemed to that Gnome apps are quite unstable in KDE. Pidgin would routinely crash (though that could have been the fact that I usually run a lot of plugins with it) and CheckGmail would periodically fail. Additionally, Gnome apps look horrible in KDE. Pidgin always had incorrect tray sizes, so it always had an enormous tray icon compared to normal KDE apps. And apps like HandBrake look repulsive in the K Desktop Environment.

KDE StaKility

KDE 3.5 was rock solid. Like, rock solid. But KDE 4.0 was really buggy. Of course, KDE 4.0 was really just another release candidate, so I was willing to give it some grace. And KDE 4.1 under Hardy was kind of buggy as well, but I was using the bleeding edge libraries, so I expected that. But when Intrepid came out with KDE 4, I was really upset. Canonical has a long and rich tradition of having the software they ship just work, and Kubuntu had normally done just that. However, this version was horrible. Maybe it was the desktop effects, but KDE 4.1 on Intrepid was almost as unstable as KDE 4.0 was on Hardy. I lived with it for about a month, then gave up.

KanoniKal support (or lacK thereof)

Canonical is one of my favorite companies around. However, there’s one thing I really don’t like about them: they focus much more on working with Gnome than with KDE (or, for that matter, Xfce). Maybe that’s just because Gnome is more popular with Canonical developers. I’m certainly not complaining. But I am upset. Jockey-KDE lagged far behind Jockey-Gnome. GDebiKDE had lots of memory usage bugs, but GDebi’s just fine.

At this point, most KDE diehards are scratching their heads and trying to figure out why I just didn’t switch to another distribution that supports KDE. I tried. But SimplyMEPIS just didn’t work. openSuSE’s package manager was annoying. And Debian is… meh. I know I’m going to get a lot of comments about why I should switch to Xandros, Red Hat, and all those other great distros. Let me say this now: I like the .deb, I like the Ubuntu Forums, and I like Canonical. I’m not leaving the *buntus.

Conclusion

One day, I snapped. I opened a terminal, uninstalled all traces of KDE, aptituded ubuntu-desktop, and I had a clean Ubuntu system all ready to roll. I loved it. Evolution rocked the socks off of me. Gedit was comparable to Kate, if not superior. Firefox was its normal self. I did have to install a few KDE apps to make me happy (Amarok ftw!), but at the core, I have a Gnome desktop. And you know what? I’m happy.