Getting Compiz to work on Suse 10.3 with a ATI Radeon Mobility 9000 driver

Recently I had to try and configure compiz for a Suse set-up. Pretty much the same as on Ubuntu, but with some twists and more graphics. I present this post mostly so I'll remember how I did it later; eventually I'll add screenshots and y'know, real explanations to the thing.

First, open up YaST, go to the hardware tab and select Graphics Card and Monitor. In there, de-select dual head support (unless you need it), select 3D graphics support and click "options" next to your graphics card. In there, find the option that says XAANoOffscreenPixmaps, select it and hit ok.

Back up and then open /etc/X11/xorg.conf. Under the server layout section, add

Option "AIGLX" "true"

.

You should be good to go; however, you may have to get rid of libIndirectGL.so.1 by typing

mv libIndirectGL.so.1 libIndirectGL.so.1.old

as root. I did this, and my compiz works fine, but I may have been solving a problem I didn't actually have when I used it. Since I've finally got compiz working fine, I don't want to tinker. Also, it's possible to simply add

Option "XAANoOffscreenPixmaps" "true"

straight to the device section of /etc/X11/xorg.conf rather than selecting it graphically.

Now, just install compiz using the 1-click install, and you should be good to go.

Separate(ish) Bookmarks for Konqueror's Browser and File Manager mode

It's been ages since I've posted an honest-to-god HOWTO here. I suspect this is partly 'cause I haven't actually been tweaking my system all that much -- I've pretty much got my Feisty Kubuntu set up exactly how I like it, and the thought of messing about with it, just makes me feel tired. However, given that the 8.04 release of Kubuntu is going to come with the option of KDE 4.0, I may very well be tempted to reverse that stance, even though the KDE4 system is likely to be a bit buggy or featureless compared to the 3.5 series.

Anyways, onto the HOWTO. Once upon a time, I used Konqueror as both a browser and a file manager. It did both well, except I wasn't pleased with how it looked, and I wanted one set of bookmarks for the web, and one for my hard drive and not both at the same time. I also wanted one set of buttons for web browsing, and one for file browsing. How to resolve the dilemma?

Answer: metabar.

Ok, open up konqueror in file browsing mode: on kubuntu the easiest way to do this is to go to the systems menu applet (on a default kubuntu install, it's on the bottom panel next to the kmenu button) and open up, say, your home directory.

Now, under the view menu in konqueror, select "Show Navigation Panel". This can also be done by hitting the F9 button on your keyboard. A panel should appear on the left side of konqueror. From the choices on it, open the metabar.

The metabar basically provides you with information on each file you select: size, permissions, previews. It also comes with actions and links. Those last two are how you're going to customize konqueror.

Right-click anywhere in the metabar area and select "configure metabar". Under the "Actions" tab that opens, add any file-browsing specific actions you'd like (things like "open in terminal" or the "up" button.) Then, under the "Links" tab, and any file-browsing specific links you'd like, things like your home directory or network places. When you've got the configuration you'd like, go to the "Settings" menu and select "Save View Profile File Management" and save your configuration.

I personally find other options of the navigation panel to be useless and it's presence ugly, so I hide it by default. To do this, open a terminal and type:

kwrite ~/.kde/share/config/konqsidebartng.rc

and set the HideTabs value to equal true. Now when you open konqueror, only the metabar will be showing.

Finally, configure the main konqueror toolbar to include whatever buttons you like by right-clicking it and selecting "configure toolbar". I personally only keep web-specific buttons there, like back, forward, reload, home, etc. but customize it however you wish. Finally, consider adding some flair to your metabar by going to kde-look and searching for metabar themes. I personally like the Advance Blue theme from this set. To install them, simply download the file from the site, right-click the metabar, choose "configure metabar" and then under the "General" tab, press "install new theme" and select the file you downloaded. It will now be available as a theme option in the drop-down list next to the install button.

Two Things I Never Expected to Happen

Good grief, someone's read what I've written.

I started this web log mostly as a repository for the thoughts I have but didn't have anywhere to place. Sort of like a very public journal. I never really expected anyone to read what I write -- if this becomes a trend, I may have to actually start researching in more detail some of the positions I take on here. My last post is a good example: in the comments to my last post, you can see that Joe Shaw made the post I linked to as a joke. I like to think my point about the attitude in the community still stands, though I'm obviously mistaken about this specific instance. I apologize.

(as an aside: I'm hardly an expert on Open Source, either. My connection to KDE is the same one I have to Windows, in that I use it just about every day. I like to think that this affords a unique perspective, since most of the news I follow seems to come from developers. Still, I think it's important to see my opinions in that context.)

The other thing I never expected is the winning streak I've been on. The basketball team I coach, the Edge, is on a winning streak and that's a very strange place to be. This is my second season coaching youth girls basketball (third, if you count spring league last year) and the reversal of fortunes has been dramatic.

I began coaching last year because the C level team didn't have anyone to coach it. Parents usually don't like taking the role. They often don't know much about basketball or coaching and if they do, they often don't fancy the time commitment. Typically however, if someone steps up and offers to be head coach, a parent will volunteer to be assistant coach. And so it happened; we actually had three coaches that year. And we lost every game but one.

Part of the problem was inexperience. Most of our team hadn't played basketball before. Another problem was sheer desire. Teams that win are often teams that want to win: the teams that run hard, that box out, that give their all out on the court. We had our most success when girls gave second or third efforts, but it was hard to get that kind of thing going team-wide. Of course, my inexperience probably contributed to the problem, in that I didn't know how to fire the team up.

This year though, things have been different. Whereas last year I focused more on individual skills that the girls lacked due to inexperience (shooting, dribbling and passing), this year I'm coaching the B-level squad. The main difference is that I've been able to focus more on the team than the players: running things like an offensive framework, a couple of defenses and getting the two lines to gel together. And it's been great to see, as a coach. Not just the winning, which is a nice change of pace but ultimately secondary to seeing the team improve. And not just improving, but improving in the exact areas I've been coaching them in and seeing that translate to success in games. That's the kind of thing that gets you excited as a coach, and is why I continue to do it.

We're looking to move up from B3 to B2 after Christmas, and I think we can be successful in that caliber of ball. Good grief! Who knew I'd be saying that, back when the season started in September?

Credit Where It's Due

I try to keep up with various open source news, since the software tends to interest me -- a quick look around the blog would tell you as much. So naturally, even though I am a KDE user, I follow Planet GNOME. And as I read it today, I had to laugh as I came across this post. Essentially, the poster is mad that a co-founder of GNOME, who is now working for Novell, has been looking over Excel, and said it was a nice piece of software. The poster now wants to know when Novell is going to stop sucking up to Microsoft.

Unfortunately, as it turns out, Excel is a nice piece of software. I mean, credit where credit is due: Excel hasn't got a lock on the industry by being a poor product. The program runs fast and has a better interface than OSS equivalents (I especially like how Excel will remind me about formula syntax when I type them in). Gnumeric doesn't have all the features Excel does. Open Office comes closer, but although it nominally has some of the same features as Excel, some of them (optimization and conditional formating spring to mind) are not nearly as advanced.

Posts like the one I linked too are worrying, if for no other reason than they show a polarization that doesn't make sense in software. It becomes a war, an 'us-versus-them' mentality where anything produced by the enemy is automatically considered inferior. And by trying to learn from what they did well, you're 'sucking up' rather than improving.

I realise that this is in part due to Novell's current pariah status in the community, but at least part of that status is this idea in the community that anything coming out of Redmond is automatically inferior and damaging. This isn't a healthy attitude. The Linux community needs to start looking at proprietary software and asking why those programs are succeeding, and what we can learn from them. Fact of the matter is, Microsoft's put a lot more effort into finding out what's important for users rather than developers. It's entirely possible that all that testing has resulted in unique insights -- and good software.

Surviving Group Projects

With few exceptions, at some point in your university life you will have to form a group with others to complete some massive project. If you're in more of the academic subjects -- english and history come to mind -- you may be able to avoid this fate, but for the rest of us, having other random people you don't know from a hole in the ground determining at least in part your success is an unfortunate reality of university. And, it turns out, the working world.

Group projects are easy when you have a solid core of people who all contribute to the project, or when the project is small enough that one or two decent people can do it all even if most of the group is useless. But what do you do when you've got a huge project and some clear underperfomers to contend with?

Read my blog, that's what. I'm here to help. Here's 13 guidelines for surviving a group project when it's all on your head to do a good job.

1. Exchange contact information. This includes specifying a preferred method of contact (likely email) and an expectation of how soon responses will be given. Make sure there is a penalty (most group projects have a peer evaluation component) for late responses to queries. A good rule of thumb is to forfeit 2% the grade if a group member does not respond within two business days.



2. Divide up the project by tasks, not pages. Make sure that "writing the report" is its own task, unless the report is so long that each section should be written individually. If you divide up a report by pages, it doesn't flow well and sounds like it was made by many different people writing in many different styles.



3. Hand in hand with point 2, don't worry about being exactly fair in the division of labour. Maybe someone has to do a bit more, but you'll have a better project if people (or maybe a smaller team of 2 or 3 people) handle a specific area. One area might be more work than the other, but it's better to have a little unfairness and a solid product than a poorer project that had the labour divided exactly equal among group members.



4. Set up expectations for deadlines, as well as the marks lost if those are missed.



5. Once you've identified underperfomers, focus them like a laser beam on whatever you hate to do. I hate market research, so I tried to focus my Marketing group on getting that done for me. I still had to do some (after all, they are underperformers) but it took the pressure off where I wanted it off the most.



6. Set up collaboration tools. Especially Word's "track changes" feature. If multiple people are editing a report (and they should, even if one person wrote it), then you can easily have four or five version of the thing floating around. Compiling that can be a nightmare. So at the minimum, track changes. At best, slap the report (or lab, or database, or whatever) on a server like Google docs and allow for multi-person edits to the same document.



7. Don't get mad; shame people. I find it more effective to go to someone who hasn't been pulling their weight with some of their work already done and a request that they finish it.



8. Unless someone is a total slacker, it's best to not fight very much about lost marks due to work quality. If it's measurable (such as not responding to e-mails or missing deadlines), then by all means dock them marks, but even then it's better not to get into a war over it. All you'll do is gain an enemy, and you won't have what you really want: sleep, and that person's part of the project done.



9. Don't be afraid to take a leadership position, especially if you're doing all the work anyways. Some people aren't really underperformers, they just need someone to tell them exactly what to do. If they are legitimately a bunch of slackers, then you're probably going to be up until 2 am finishing the stupid project anyways, and that makes your word law.



10. If you do take a leadership position, always explain why you think your ideas are best. Accept criticism. Just because someone is an underperformer doesn't mean they don't have good ideas. If any arise, don't be too prideful to use them.



11. For the love of God, don't do everything yourself. Just because you're surrounded by slackers doesn't mean you can't use their stuff. Keep it as a base, and shore it up with the grammer/research/wording/effort it requires to shine.



12. Have face-to-face meetings as often as you need. If it has to be at 6 pm on a Saturday, do it. You often can't replicate that input over email.



13. Make sure meetings are on task. Establish a rotating chair for the meeting, and make it that persons job to set the agenda and keep things on task. Everyone has better things to do than meetings, so keep them short, sweet and to the point.

Finally, I should stress that you should still be a team player throughout all of this. Never criticize people, just their actions (i.e. do not say "you're an idiot". Instead say "I don't think this section has been properly cited/researched/whatever", and then say why you think that.) Remember, the goal here is not to do everything yourself, but to get as much out of your underperforming group so you don't have to kill yourself doing everything to make the deadline. And finally, remember that not all groups are populated by slackers, and that people do have other things going on in their lives than the project -- so take that into consideration before you give into the urge to blow up at them.

Here on the other side of November...

Wow, it's been awhile since my last post. November is project month at university, and since I'm taking business and that means almost every class has a group project. Most of them, naturally enough, are due sometime in November. However, the biggest of them, my Marketing project, has been mostly retired and this has given me something akin to breathing room.

Which is why I'm going to take the opportunity to blog about technology.

Has anyone here seen IBM's Lotus Symphony? My opinion is best summed up by: high hopes. Based on Open Office, the premier office suite available for the Linux desktop, Symphony gives the office suite on Linux more developers, and God willing, a better interface and functionality.

Open Office is a fine product, especially given it's price. But it's not nearly as good as MS Office, especially the 2007 version. This is more of a problem than with other apps, since the fact is people were not typing out documents on Wordpad before switching to Linux. For example, most people have not bought Adobe Illustrator, and therefore an equivalent if less feature-full program being available for free on Linux (Inkscape) is a selling point. But most people have been using some version of MS Office before switching, and therefore the higher cost of that program is considered sunk. So having a less feature-full program available for free becomes a liability, not an asset.

Just looking at the screenshots for Symphony though, it's looking like the interface is improving, and more importantly, moving away from being a simple MS Office clone. Spreadsheets needs some attention: optimization, anyone? Or how's about reminding me about formula syntax as I write them? Also, presentations needs some serious love. The key here it a) make more templates and b) hire decent graphics designers so they're slick. Oh, and would it kill anyone to make a "handouts" layout that includes lines for notes?

More than Symphony, though, I'm looking forward to KDE 4 making an appearance on the Linux desktop. Eventually.

I tried out what they're calling "Release Candidate 1", though the system is clearly not ready for production use. Having said that, I see enormous potential for KDE 4. The widgets, the smooth animations that are scattered throughout the system, the icon theme -- all look very slick and very professional. Assuming some non-KDE technologies can be easily integrated into the system, or ported over for KDE equivalents, I think that the Linux desktop will begin to match or surpass the offerings from private companies, though we'll still lack in 3rd party apps.

The non-KDE technologies I'm thinking of here are PulseAudio, Avant Window Navigator and Compiz Fusion: the most exciting Gnome-based technologies out there. I know Kwin now does compositing, but I'd like the zoom, desktop cube and tab windows features from Compiz (some of the visual flash would be nice too, but not necessary) as I find those rather handy for windows management. As far as AWN goes, it's the best damn dock going. Period. With some KDE-specific plugins, I think it could feel right at home. Finally, I don't know much about PulseAudio, but being able to control sound on an app-by-app basis sounds very, very cool.

NaNoWriMo!

It's that time of year again!

Yes, it's officially National Novel Writing Month, so naturally that means I'm already 2 days behind in my writing. Still, this year is my year: 50 000 words by November 30th or bust baby!

(And in true Geek fashion, I'm gonna craft the thing entirely in FOSS -- most likely KWord, since I'm a bit of a KDE fanboy. For awhile I was contemplating writing the thing in Vim using Dvorak, but then I realized even I have my limits.)