Kommandeering Developers Everywhere

Thursday, September 29, 2005

Theory vs practice

As the saying goes:

In theory, theory and practice and the same. In practice, they'e different.

Now I agree fully in theory with the stabilization and hopeful boredom of the desktop (due to completeness and stability). In practice, I'm not quite there yet.

My tech pride forces me to cover this, and not just send a footnote email to Aaron. When pride force you to post on a topic that will undoubtedly result in 102% of the readers to label me as incompetent (or cement their existing label of me being incompetent) and I do it anyways, it's ugly.

For my computer user profile, file me under "enthusiast" or "newer hardware" In other words, the linux kernel still kills me. I try to ride onto hardware as long as possible, but when I upgrade, I tend to buy fairly new components (but well below the insane price-threshold point of course, reserved for suckers and cash-fat businesses.)

That being said, I've never, ever, ever installed any Linux distro that has ever, ever identified and configured 100% of my equipment. Ever. Not with Red Hat in 1998 on my PII because of my SCSI controller and not since. (To be fair and for the record, in my personal experiences on this hardware sampling pool, Mandriva/Mandrake and Slax for live CDs do consistently recognize and configure hardware components as well or better than any other option)

Recent tribulation? Caved and built a new desktop. AMD64, socket 939, striped RAID Serial ATA, PCI-express video card. Not bleeding edge. At risk of being labeled a message board fanboy, I'll stop there - let's just say I shopped the linux hardware friendly sites and support lists to maxmimize my odds. Now, up through trying OpenSuSE 10.0, not a single distribution would even install through completion. Not Fedora/OpenSuSE/Ubuntu/Kubuntu (of course with same back end)/etc/etc. No live CDs will run. Not Knoppix/$other_version_here. 64-bit distros/32-bit distros, you name it(Note: I did not try the newest versions of Xandros or Linspire). So I install XP with no problems.

After installing VMWare, I tinker further. Even in VMWware tweaking the settings, no distro will install. Same goes for my other dekstops. other servers and other laptops. Conclusion: I've never had a Windows install finish *unsuccessfully* and I've never had a Linux install finish *successfully*. ie. my worst installation with Windows since 3.1 is better than my best Linux experience still to this day.

Of course, every individual experience is different, but I'm still waiting for the day I can buy standard brand components (Gigabyte, Nvidia, Asus, Logitech, etc) that have been out on the market for 6-9 months, are listed as being supported, and have a corresponding distro support them.

Conclusion: God I hope we get to the point where the KDE desktop is boring and stable, but I'm not even putting money on my desk yet to save for Linus' pina coladas for all his free time. In the interim, I'll still push heavily for vendor pre-installations.

Wednesday, September 28, 2005

Ding! Ding!

Ringside announcer 1: In the red corner, we have a braaaaaaand new Sun E6900 running Oracle 10g.

Aaaand in the blue corner, we have a single stored procedure written by the Master of Coding Disaster Waaaade Olson; pulling 66 fields from no less than 15 full-outer joined tables, each table having upwards of 4 million rows and joining on non-indexed columns, all fully sorted. *

Ringside announcer 2: This could be a fast fight....it's going to be a bloodbath!

Toady teammate:
Get a bodybag man!

Place your bets!

* Unbelievably dba-tuned and approved. Go figure. I'm washing my hands of this debacle. Based on client requirements, there is *no* other way to code this monstrous Kraken.

Monday, September 26, 2005

Two fun things

Now, on other days I might answer peanut butter and jelly. Or the Doublemint twins.

But today it's Klik and KimDaBa. Of couse, I haven't used either, but that's never stopped me from having an opinion before.

Klik-Among other great things, it's YAATT (yet another awesome testing tool). Virtualization, NX remote sessions, live CDs and now Klik; each having a different testing forte. It's boggling to think that the outgoing 4.0 code could eradicate the standard "buggy milestone release" mentality. That's it, I'm going to run val/cachegrind on a cppunit case of a qtunit case of code on a Klik application through NX of a live CD running in a XEN virtualized environment running on a second live CD. I'll tell you how it goes.

KimDaBa-Not in love with the name, not sure how it'll differentiate itself with Tenor or other general meta searches in the future. But from what I've seen it's got a simple intuitive interface thus far - nicely done. More importantly I love the initial flash demos I've just seen. A great first swing! Simple, straightforward, just a bit of humor, and broken into logical features. Already, it seems to me that it wouldn't be imposing at all for brand new KimDaBa users to watch those videos.

Imagine in the not to distant future some KDE some application certification process. You know that for documentation, application-based help, web-based help and multimedia help are going to be expected in various forms. I'd love to sit down users in front of those videos and find out in their own words how it helps. Good work! (and by extension, good work to those like Physos for pushing forward with such technologies).

i'm now off to find some peanut butter and jelly.

The sound of inevitability

I wish I could write about my rock star weekend, but after a mentally draining work schedule, a weekend of rain and a day-long headache, I can safely file this weekend under the "Fast and Forgettable" category.

So pulling from the Rantomatic9000 device I have (you'll have to imagine the whirring computer beeps and clicks yourself), let's talk about the downfall of the roman empire (aka MS). I know, I know...covered before...but it's Monday morning, you're not going to get 'The Importance of Being Earnest' every post, are you?

If you're not in the mood for a rant, I'd bail here.

The premise is simple: Over time, regardless of the ecosystem, paradigm shifts lead to balance and stability, which leads to another paradigm shift. Whether you're talking about animal populations, governments or businesses; you can always make a case why each circumstance is unique, and they never are. Now, the more competitive and mature the area of interest, the faster the frequency (and I do mean 'frequency' in the wave-length sense - we're talking harmonics people!).

Example: Whereas as a nation may have held power over hundreds if not thousands of years, that time has been whittled down to where the U.S. will have historically been considered unchecked for less than 75 years.

In the world of business, look at the U.S. in the last 150 years. Consider alone such disruptive events such as the automobile, electricity, telecom and maybe even raw materials (industrialization and commoditization of coal/iron/etc).

In each case, the tipping point and eventual ubiquity of a product is done through a single entity (or very small core group of entities). Universality is effectively accomplished through monopoly. A hundred years ago it was Ford, today it's Microsoft. Different sector, same path.

Once a product/service becomes habit (such as the phone or electricity in your wall socket), new groups struggle to climb over the barrier to entry. Is the task of rezoning land, getting permits and stringing thousands of lines of wire different from today's software patents and millions of lines of code? Are the special interest groups and soft money any different? The hurdles thrown out by the leader in an attempt to maintain their place are only stalling tactics.

Eventually, things settle down. The threshold of the barrier to entry is crossed and cottage competitors spring up. Further down the road, in complex endeavours such as writing operating systems or or creating massive distribution networks for automobile assembly, some form of oligopoly emerges. Not ideal, but better. Whether we're talking global economies such as North America/EU/Asia, or businesses such as the Big Three auto makers in the U.S., Coke/Pepsi/RC Cola, McDonalds/Burger King/Wendys, you'll find a majority leader who makes the rules, a strong competitor, and a tertiary entiry. Will Apple fall naturally into the Pepsi/Burger King secondary role? Don't be so quick. They sell hardware and multimedia. The desktop is the end to that means.

So, Napoleon and Alexander the Great couldn't keep power. Ford couldn't keep power. Micrsoft won't keep power. If MS and Intel want to be known as the powers that put a computer on every desk, they'll also have to accept the future that comes with it, and start making room.

Moral of the story? Am I making sweeping generalizations here? Of course, it's a freaking blog. But not as many as you think. I love the sense of urgency in fighting software patents, IP and developing code maturity. Don't stop! But to take a step back and look macroscopically, the railroads tracks are laid and the course of action is set. KDE has to be identified as a participant in the dismantling of yet another monopoly and will gain explosive growth.

We drive different cars on the same roads, talk over different phones on the same network, plug different items into the same wall socket, and purchase different materials of the same grade to manufacture our widgets. Our OS and application situation will be no different.

As long as KDE continues to grow and evolve, we have effective leadership and put in the hours, adoption will explode over time. Whether due to economics in emerging markets, or political factors such as avoidance of U.S. reliance. The U.S. loves to stubbornly make bad decisions and stick with them (see: weight and distance measurements), but when one of our own states bail on proprietary formats, the wheels are definitely in motion.

So sit back, work hard and pat yourself on the back for participating in Nature's Grand Balancing Act.

PS-I promise my next post will be lighthearted. Like my dog biting me or spilling on my keyboard.

Thursday, September 22, 2005

The importance of boot times and USB

So, $DEITY_OF_CHOICE decided that my area of the U.S. hadn't been slapped around by the weather yet, and was probably overdue. Of course, the only information that had been on local news sites and news broadcasts was bad weather miles and miles and miles away, so I was blindsided by this all.

Cue the thunderstorms and tornadoes. So power goes out very early on into this debacle, and I have no clue where I'd have a battery powered radio.

But I do have a UPS that reminds me that it has engaged with its shrill intermittent beeping. Only the essentials are plugged in: Cable modem, firewall and mail/web server (Suse 9.2). The Suse box is of no good to me, because it won't recognize any USB keyboard/mouse from Logitech plugged in after BIOS bootup. Awesome.

So, spurred into action in the dark, I boot up my laptop (Open Suse 10.some_way_too_early_release) with its own UPS (aka full battery).

And wait. And wait. And the blinking of the splash screen icons seem to synchronize with the beeping of the UPS just to rub it in. And wait. And finally when the UPS is about to give out, I get in and head to a weather site...and no net connection. Now in the past, no power did not necessarily mean no cable net access, but that was the case this time.

Moral of the story? The lack of cable access was the villain, but not by much. To everyone working on speeding up OpenSuse boot times and working on USB PnP and drivers, I salute you.

Monday, September 19, 2005

Now how did I get here?

Major tactical error.

I'm supposed to get paid to draw on whiteboards and talk about ambiguous things like "strategy" and "vision".

Now I've got two weeks to write a dumptruck full of code (restricted from saying more).

Haven't been in CodeMode in a while: Cancel all meetings until 9/30, make sure I have 24-access, functional VPN, Mountain Dew and Pepto Bismol.

I need an intern.

Friday, September 16, 2005

Standing on the shoulders of dragons

Listening to Aaron on the TLLTS show today, two things struck me:

1) Roughly 62% of information wants to be free. For two Appeal projects, Plasma and Oxygen, reasons were cited why work-in-progress was not the most visible and public.

Call it building anticipation, call it keeping others from plagarism (reference title here). I wondered if some listeners were ruffled by the insinuation that Open Source doesn't imply Open Plan and Open Progress. I'm all for a little privacy. Finding certain IP practices irresponsible won't be consolation when they're used against you.

Microsoft tries to code agile with more visibility and KDE starts holding its cards a little closer. Good times.

2) Questions about Gnome distro growth and desktop percentages. Just as Linux is reported to greedily take from AIX and Solaris and such, people like to discuss Gnome and KDE succeeding at the expense of the other. If you think the ~90% share that Microsoft is big, look at the numbers of people that will start using computers in the next decade. In emerging economies, many will look at their paycheck, look at the price of MS software, look at the paycheck again and shrug. The battle isn't against Gnome or Enlightenment, it's against a pirated copy of Windows. As my broker always tell me every month he loses me money, "It's not the percentage of the pie you get, it's the amount of pie that you get." KDE kicks ass and it's poised to be on several million new desktops over the next several years.

Rule of competition: The market leader never acknowledges the lower competitors (and Microsoft *always* fails this rule). The second place always differentiates from the leader. And the new entrants (OSS desktops) pitch the new paradigm and proposition. Debate is counterproductive, you're selling the validity of the market shift. I just hope my blog isn't outsourced :).

Thursday, September 15, 2005

Recursive meritocracy?

The merit karma of the participants battle it out to determine the formula to gauge the merit karma of the bug filers.

Like sports fans helplessly watching events unfold with their favorite team on the television, so do casual bug reports tremble with anticipation at their assigned ranking.

"Wade Olson: Your rank shall be -1. All of your reports will be routed to /dev/null. Go about your business."

I love these 'mailing list' discussions on the blog. There's an article percolating in the back of my head about the growing pains of KDE. With the argument of "we work on what we like, if you want something, here's a manual" slowly fading away, KDE (due to its own success) is faced with typical business/non-profit resource allocation issues. Must do vs. Can do Vs. Should do Vs. Want to do. Cage match!

Bug tracking/documentation/translation doesn't turn a lot of people's cranks, but such attention to detail can alter the trajectory of KDE's path through the stratosphere. And the analogy is about the atmospheric layer between the troposphere and the mesosphere, not the ultralame Las Vegas hotel.

So keep the ideas and debates coming. If there wasn't any contention, KDE wouldn't be growing and wouldn't be feeling any external pressure. Just put some tape over the "Hobby-to-Job" meter next to your monitor.

My idea to lob into the fracass: Add a reporter_type table that hold values such as "Packager", "Distro", "Individual", "Contributor", "Standard", etc. Give every reporter in bugzilla a reporter_type of "Standard" and modify acounts of specific people as you see fit. Then change some pages like http://bugs.kde.org/query.cgi to allow querying based on reporter_type. If you want to use the system as is, that's cool. If you want to search for more "likely to be quality" bugs such as packagers, go ahead.

Either way, a massive ticket closing session will have to be done closer to the next release as I'm curious how many bugs will no longer be valid after a year of coding and a Qt4.0 port.

Tuesday, September 13, 2005

Hip to be Square

Microsoft has unleashed a new site to discuss their upcoming Vista Gadgets.

Rich media gadgets and blogging. Why don't we ever think of these things?

To combat this "cool" arms race, the Plasma team will be wearing sunglasses at all times. A team tattoo appears to be in the works. MC P0z3r is working on the rap anthem "Supah Karambah".

Nonetheless, I'd encourage everyone to spend some time reading about MS's announcements at their Professional Developers Conference with an open (hard) and critical (easy) mindset. Press releases abound at this time of year, and it's a good time for everyone to do a mental competitive analysis for their field of interest on what's getting better and worse.

I'd love to hear objective feedback from the Appeal and Usability teams on Office 12 and Vista screenshots.

The software pricetags suddenly seem sensible when you realize they're including a 21" flat panel to handle all that wasted real estate.

Sunday, September 11, 2005

Seven horsemen appear on the horizon

And is anyone terrified? Terrified as a potential captive user: possibly. Terrified as a competitor: possibly not.

To be more clear, Paul Thurrott is speculating and postulating about forthcoming Windows Vista versions, and hopefully you have your slide-ruler and aspirin ready.

If true, 15-year old electronic store employees could have less than 18 months to master Vista version nuances in their quest to push new PCs and extremely valuable insurance service plans. I weep for the future.

Admiral Akbar is telling me it's a trap, and I'm not convinced either way. With closed source comes closed tactics. What in the Wide World of Sports is Microsoft thinking with each new Vista development? Dropping features and adding chaos. Will Zack Rusin be putting DRM into Xegl to restrict true HD content just to compete?

As with the console wars, we have three parties with fairly major desktop/OS releases planned in roughly the same time frame. Can KDE really be in the passing lane?

Where in the world is Carmen Sandiego?

At 12:41 AM Minneapolis time (the Greenwich of the Midwest as they call it), I send a message to clee.

He responds within 7 minutes.


It's too late (read as: I'm too lazy) to see if he's on this map, but unless it's Toyko or Fairbanks Alaska, the timezones just don't make sense for a response that fast.

What can we deduce? His dedication is second to none. Or he has a robot answering his mail.

Either way, if you don't hear back from him in 10 minutes, he and his robot are most likely ignoring you.

Great job Chris!

Saturday, September 10, 2005

Back in the saddle

Ok, I've deleted page after page of Hello Kitty fan club entries, and we're back at square one.

This blog is being refurbished and repurposed for KDE-specific topics.

int iSuccess = total_readers - readers_that_dont_care - readers_that_disagree;

I'm guessing a negative result coming my way.