Thursday, November 30, 2006
Chicago Bears
Man, where's the love? Why the constant "overrated" flag that many sports writers keep pinning on the NFC North leader? Before the season, I doubt any of them pick the Bears to be Superbowl contenders, so the fact that they're leading the NFC should really mean that they're beating expectations. Instead, it's a constant "they are who we thought they are." (To this day, I'm still trying to understand Dennis Green's quote. Is he saying that the he thought the Bears was a good but not perfect team who will beat his Cardinals? If that's the case, yes, the Bears showed flaws and did beat them even without an offensive touchdown...).
Friday, November 24, 2006
fedora core 6
This weekend, I decided to upgrade to Fedora Core 6 since M sent it to me on DVD. This would be the first time I've "upgraded" to a new version rather then start with a clean install and I was curious as to how clean it would be. As we all know, Windows can be a nightmare to upgrade and it's often easier to just do a clean install and re-install all the data and application.
My system was a Shuttle SN95GV3 with a Athlon 3500+ (64bit) CPU with Fedora Core 5. I haven't customized it very much and beyond my original spec that I detailed previously with the exception that I added another 1Gig of memory for a total of 2 Gigs. I've installed most software through yum and kept up with the latest packages.
Before I started, I ran yum update just to make sure I got the latest packages on FC5 in case the upgrade had some dependencies (don't think this actually mattered). I popped in the FC6 DVD and rebooted. I selected "upgrade" from the menu (as opposed to clean install) and from there it was pretty automatic. I had to verify a couple of settings in the menu (language, keyboard, etc.) by clicking "Next" and then it started the upgrade.
It took about 1.5 hours before everything was installed and ready for a reboot. The system came up fine until it tried to load up X which it failed because of nvidia video driver (I was using Nvidia's from livna rather then the stock nv driver). I didn't worry too much about this and just ran yum update since I figured it didn't have the latest version for the kernel that fc6 uses. Yum identified nearly 700megs of files that needed to be downloaded and updated so I said ok and let it run. I ran into one problem with the yum update because of a dependency problem with OpenOffice. This was resolved by me telling it to not update OpenOffice. Everything else got updated and following a reboot, I was back into X where I uninstalled the existing Open Office and then had it install the latest version.
My previous settings were all preserved and everything seemed to be working and then I noticed something... In the GUI boot up message, I noticed the following error message:
/lib/modules//kernel/arch/x86_64/kernel/cpufreq/acpi-cpufreq.ko
FATAL: Error inserting ...acpi_cpu_freq.ko .... No such device
Now, this is wierd... I never seen this in FC5 and FC6 still worked even though it had this error. After some research, I discovered that the error came from the cpuspeed daemon which I don't remember using before so I suspect it was activated through the FC6 upgrade. This daemon is what will switch the speed of the CPU depending on load to save energy and battery life. I decided to turn this service off even though it should be working since the Athlon has the Cool 'n' Quiet feature.
Two things bothered me, though, and I spent quite a bit of time researching even though they weren't critical to the workings of the server. First, the error message I saw only was in the GUI boot message display. Nothing was written to the log and somehow the boot.log file in FC isn't updated. I'm not sure why that is, but seems to happen in Fedora.
Second, why was cpufreq failing even if I didn't want to use it? In the logs that was written to, I noticed this:
powernow-k8: BIOS error - no PSB or ACPI _PSS objects
Research on the web seems to point that the kernel was having problems with Cool 'n' Quiet. Since this fails, it's no wonder cpuspeed failed later. I couldn't find any solution other then to try to upgrade the BIOS. This caused me two dilemmas. First, I never put in a bootable floppy drive in the Shuttle. Second, Shuttle makes it so hard to 1) find the bios version and 2) enable/disable Cool 'n' Quiet. For (2), people said Shuttle makes it always available if the OS supports it so there is no BIOS setting (stupid). For (1), I realize that Shuttle always show their XPC logo rather then the useful bootup messages (i.e. the BIOS version) but the way around it is to hit or unplug your keyboard so it'll stop the boot on error and display the boot messages (this work 'cause the default behaviour for the Shuttle is to stop on all errors).
I found out that my BIOS is a bit old but without a boot floppy drive I'm not sure how to upgrade the bios (might go buy a USB floppy). Whether this will fix the kernel error, I'm not certain but it's worth a shot someday.
Overall, I feel the install went extremely well and the upgrade was for the most part very smooth. It made me aware of a few problems that wasn't related to the upgrade so the bulk of my time wasn't trying to upgrade but to figure out some long standing problems.
Update: Instead of buying a USB floppy drive, I made a bootable CDROM with the latest BIOS update from Shuttle, but even after flashing the BIOS with the most recent version, I still see the same error message so it didn't seem to help.
My system was a Shuttle SN95GV3 with a Athlon 3500+ (64bit) CPU with Fedora Core 5. I haven't customized it very much and beyond my original spec that I detailed previously with the exception that I added another 1Gig of memory for a total of 2 Gigs. I've installed most software through yum and kept up with the latest packages.
Before I started, I ran yum update just to make sure I got the latest packages on FC5 in case the upgrade had some dependencies (don't think this actually mattered). I popped in the FC6 DVD and rebooted. I selected "upgrade" from the menu (as opposed to clean install) and from there it was pretty automatic. I had to verify a couple of settings in the menu (language, keyboard, etc.) by clicking "Next" and then it started the upgrade.
It took about 1.5 hours before everything was installed and ready for a reboot. The system came up fine until it tried to load up X which it failed because of nvidia video driver (I was using Nvidia's from livna rather then the stock nv driver). I didn't worry too much about this and just ran yum update since I figured it didn't have the latest version for the kernel that fc6 uses. Yum identified nearly 700megs of files that needed to be downloaded and updated so I said ok and let it run. I ran into one problem with the yum update because of a dependency problem with OpenOffice. This was resolved by me telling it to not update OpenOffice. Everything else got updated and following a reboot, I was back into X where I uninstalled the existing Open Office and then had it install the latest version.
My previous settings were all preserved and everything seemed to be working and then I noticed something... In the GUI boot up message, I noticed the following error message:
/lib/modules//kernel/arch/x86_64/kernel/cpufreq/acpi-cpufreq.ko
FATAL: Error inserting ...acpi_cpu_freq.ko .... No such device
Now, this is wierd... I never seen this in FC5 and FC6 still worked even though it had this error. After some research, I discovered that the error came from the cpuspeed daemon which I don't remember using before so I suspect it was activated through the FC6 upgrade. This daemon is what will switch the speed of the CPU depending on load to save energy and battery life. I decided to turn this service off even though it should be working since the Athlon has the Cool 'n' Quiet feature.
Two things bothered me, though, and I spent quite a bit of time researching even though they weren't critical to the workings of the server. First, the error message I saw only was in the GUI boot message display. Nothing was written to the log and somehow the boot.log file in FC isn't updated. I'm not sure why that is, but seems to happen in Fedora.
Second, why was cpufreq failing even if I didn't want to use it? In the logs that was written to, I noticed this:
powernow-k8: BIOS error - no PSB or ACPI _PSS objects
Research on the web seems to point that the kernel was having problems with Cool 'n' Quiet. Since this fails, it's no wonder cpuspeed failed later. I couldn't find any solution other then to try to upgrade the BIOS. This caused me two dilemmas. First, I never put in a bootable floppy drive in the Shuttle. Second, Shuttle makes it so hard to 1) find the bios version and 2) enable/disable Cool 'n' Quiet. For (2), people said Shuttle makes it always available if the OS supports it so there is no BIOS setting (stupid). For (1), I realize that Shuttle always show their XPC logo rather then the useful bootup messages (i.e. the BIOS version) but the way around it is to hit or unplug your keyboard so it'll stop the boot on error and display the boot messages (this work 'cause the default behaviour for the Shuttle is to stop on all errors).
I found out that my BIOS is a bit old but without a boot floppy drive I'm not sure how to upgrade the bios (might go buy a USB floppy). Whether this will fix the kernel error, I'm not certain but it's worth a shot someday.
Overall, I feel the install went extremely well and the upgrade was for the most part very smooth. It made me aware of a few problems that wasn't related to the upgrade so the bulk of my time wasn't trying to upgrade but to figure out some long standing problems.
Update: Instead of buying a USB floppy drive, I made a bootable CDROM with the latest BIOS update from Shuttle, but even after flashing the BIOS with the most recent version, I still see the same error message so it didn't seem to help.
Saturday, October 28, 2006
Search Bombs
A typical morning often finds me getting online and checking the news and reading some blogs before I head to work. This past week, I read an blog written by Jeremy Zawodny titled "You're welcome to google on Yahoo." The entry was in response to Google's attempt to get people to not overly generalize the word "google" which got a lot of attention this week because Google blogged about it on their official site (even though the news itself is actually pretty old).
As I was listening to the radio on the way to work, the reporter started to talk about "google bombs". My first thought was that people might be trying to hack Google, but then I realize that they were talking about search engines being tricked by political campaigns to return negative attacks on opposition candidates. The tactic is that if you are X running against Y then you put up a lot of pages that speaks negatively about Y so that when users search for "Y" they would find bad things written by X. This really didn't have anything to do with Google and has a distinctly negative connotation.
It is this type of usage that I believe Google is trying to prevent against. Imagine what other companies might do if their names were being used completely out-of-context in a negative way... Although I often find that people give Google a little too much credit, in this case I actually admire that they attempted to explain their action to the public in a way to help people understand. Too often, the languages use by company when they try to justify their actions are either too full of legal speak or they just don't say anything to the public at all.
As I was listening to the radio on the way to work, the reporter started to talk about "google bombs". My first thought was that people might be trying to hack Google, but then I realize that they were talking about search engines being tricked by political campaigns to return negative attacks on opposition candidates. The tactic is that if you are X running against Y then you put up a lot of pages that speaks negatively about Y so that when users search for "Y" they would find bad things written by X. This really didn't have anything to do with Google and has a distinctly negative connotation.
It is this type of usage that I believe Google is trying to prevent against. Imagine what other companies might do if their names were being used completely out-of-context in a negative way... Although I often find that people give Google a little too much credit, in this case I actually admire that they attempted to explain their action to the public in a way to help people understand. Too often, the languages use by company when they try to justify their actions are either too full of legal speak or they just don't say anything to the public at all.
Tuesday, October 10, 2006
Lost Season 3
I just watched the season 3 premier of Lost and while it was a pretty good episode, I can't help but keep thinking... "Why is Jack being such a cry baby...?"
Sunday, October 8, 2006
Volkswagon commercials
Those new Volkswagon commercial really annoy me. It seems to be a scare tactic because the commercial doesn't really give information on what might make their cars safer then others. In fact, can't I interpret their commercial to mean that if I own an Volkswagon then I'm more likely to get into a car accident?
Saturday, October 7, 2006
Updated Homepage
After nearly 5 years, I finally made a major update to the homepage. Some might be wondering what the heck I'm smoking since the page looks nearly identical to the previous version and that was the intention. The changes were mostly under the covers as I dragged myself into the modern web layout world by going with a CSS table-less layout. Besides learning more about CSS layout, there are other benefits.
It was more difficult to code, however, but that could be that the original was such as simple table design.
It's good to give myself a refresher on CSS and layout. I've been looking at more advanced JavaScript such as OOP Javascript and DOM manipulation, so hopefully I'll be able to find some time to experiment some on the site soon.
- The code layout is better as the content and presentation is better seperated and there are no more inline CSS styles.
- The size of the download has been reduced by nearly 50% although I did include some YUI CSS libraries to help clean up the fonts. This brought the initial size to about the original but once the page is loaded the first time those libraries won't be downloaded again.
- The page should look consistent between IE and Mozilla. By applying a different stylesheet, the page should also render in other devices.
It was more difficult to code, however, but that could be that the original was such as simple table design.
It's good to give myself a refresher on CSS and layout. I've been looking at more advanced JavaScript such as OOP Javascript and DOM manipulation, so hopefully I'll be able to find some time to experiment some on the site soon.
Saturday, August 19, 2006
Open source on Windows
I've been looking around at various open source IM clients such as GAIM and Miranda since I want to learn how IM clients are implemented. I downloaded the source for both and tried to compile both. GAIM is supported on multiple platform while Miranda is an Windows-only IM client. Compiling GAIM on Linux was basically running "make" and everything compiles. Compiling it on Windows was more complicated but GAIM has updated their site to include a script that will download all the various components that are needed in addition to the GAIM source. This made it a lot simpler to get a working compile, but even without the script the site had decent instructions on what was needed. For most UNIX/Linux open source projects except the most advanced and complicated, compiling the source code is pretty straight forward and it allows newbies to dive into the code.
Windows open-source projects, on the other hand, are different. The last few projects I've tried to compile were filled with problems. One problem is that there is no standard development environment for Windows. UNIX has the GCC suite which is open source and easily obtainable (if not already on the system) while Windows' primary development environment is the commercial Visual Studio tools. The developers themselves seem to not have standardize on a single platform. Miranda, for example, is developed on MingW by some and Visual C++ by others. Microsoft also doesn't make it any easier despite releasing the free Visual Studio express editions (which they admits isn't meant for windows development anyway).
Thus, my attempt to compile Miranda wasn't too successful (I don't normally do much development on Windows so I'm not as familar with it to begin with.). I installed Visual C++ Express only to find out that I also needed the Platform SDK. Once that installed, I still needed other things, but it wasn't clear to me where I get them. In the end, I wasn't able to compile my own working executable. I'm sure I'll be able to with a little more research, but it definitely takes a lot more work then on Linux.
While Microsoft talks about the importance of the developers, they really mean commercial developers. While they are doing many things to try to fend off open source competitors, they should really go back and do something OS vendors used to do (and UNIX vendors still do): include development tools as part of the OS.
Windows open-source projects, on the other hand, are different. The last few projects I've tried to compile were filled with problems. One problem is that there is no standard development environment for Windows. UNIX has the GCC suite which is open source and easily obtainable (if not already on the system) while Windows' primary development environment is the commercial Visual Studio tools. The developers themselves seem to not have standardize on a single platform. Miranda, for example, is developed on MingW by some and Visual C++ by others. Microsoft also doesn't make it any easier despite releasing the free Visual Studio express editions (which they admits isn't meant for windows development anyway).
Thus, my attempt to compile Miranda wasn't too successful (I don't normally do much development on Windows so I'm not as familar with it to begin with.). I installed Visual C++ Express only to find out that I also needed the Platform SDK. Once that installed, I still needed other things, but it wasn't clear to me where I get them. In the end, I wasn't able to compile my own working executable. I'm sure I'll be able to with a little more research, but it definitely takes a lot more work then on Linux.
While Microsoft talks about the importance of the developers, they really mean commercial developers. While they are doing many things to try to fend off open source competitors, they should really go back and do something OS vendors used to do (and UNIX vendors still do): include development tools as part of the OS.
Monday, August 14, 2006
Getting Linux to work in Portrait mode...
This weekend I swapped out my 17" LCD (Dell 1705FP) with a 19" (Dell 1905FP) version that can be used in portrait mode. There is not particular reason why I used the portrait mode other then desktop space. I first tested it in Windows and it was just a matter of plugging it in and then using the Nvidia control panel to set the rotation to portrait.
I then switched over to Linux (Fedora Core 5) and ran into a few snags. There was a dependency conflict between the latest nvidia driver and the kernel I had installed and so the system was starting up with just the native nv drivers. Once that was resolved, I realized that the Linux version of the control panel didn't have the options that the Windows version had including the ability to rotate. It took awhile, but with the help of a friend and Linux guru, i was able to get Linux to go into portrait mode.
First, to get the Rotation option to appear in the Nvidia control panel, I had to add to the Device section:
Option "RandRRotation" "on"
This gave me the option to change the location once I log in, but the problem is that until I do everything appeared in landscape mode so I would have to tilt my head 90 degrees in order to log in and get to the contral panel.
Digging around the Nvidia drivers appendix, I found the option to have X rotate automatically on startup:
Option "Rotate" "left"
This did the trick and I now have less strain on my neck, but then I noticed that things were a bit off on the screen. I realized that Linux was running the driver at 75hz refresh rate which is supported by the monitor but LCDs run optimally at 60hz. I had selected the right monitor, but checking the /etc/X11/xorg.conf file, the monitor section have a range for the VertRefresh rate up to 76hz. This is ok, but Linux only used 75hz and didn't let me select anything different. Even when I modified the VertRefresh to just 60, it still used 75.
Again, my friend came to the rescue and suggested I added the following to the Device section:
Option "UseEdidFreq" "false"
Restarting X and all was well. ^^b
I'm still getting used to having my monitor taller then it is wide, but the only troublesome thing is finding wallpapers for it!
I then switched over to Linux (Fedora Core 5) and ran into a few snags. There was a dependency conflict between the latest nvidia driver and the kernel I had installed and so the system was starting up with just the native nv drivers. Once that was resolved, I realized that the Linux version of the control panel didn't have the options that the Windows version had including the ability to rotate. It took awhile, but with the help of a friend and Linux guru, i was able to get Linux to go into portrait mode.
First, to get the Rotation option to appear in the Nvidia control panel, I had to add to the Device section:
Option "RandRRotation" "on"
This gave me the option to change the location once I log in, but the problem is that until I do everything appeared in landscape mode so I would have to tilt my head 90 degrees in order to log in and get to the contral panel.
Digging around the Nvidia drivers appendix, I found the option to have X rotate automatically on startup:
Option "Rotate" "left"
This did the trick and I now have less strain on my neck, but then I noticed that things were a bit off on the screen. I realized that Linux was running the driver at 75hz refresh rate which is supported by the monitor but LCDs run optimally at 60hz. I had selected the right monitor, but checking the /etc/X11/xorg.conf file, the monitor section have a range for the VertRefresh rate up to 76hz. This is ok, but Linux only used 75hz and didn't let me select anything different. Even when I modified the VertRefresh to just 60, it still used 75.
Again, my friend came to the rescue and suggested I added the following to the Device section:
Option "UseEdidFreq" "false"
Restarting X and all was well. ^^b
I'm still getting used to having my monitor taller then it is wide, but the only troublesome thing is finding wallpapers for it!
Tuesday, August 8, 2006
MS doesn't want you to code for Windows?
I like the idea that Microsoft decided to release the express editions of their development tools for free to encourage people to program on Windows (although this doesn't mean I like Windows). I thought it was a smart move for them to lower the barrier to build for the platform since the other alternatives are gcc and mingw which are both very nice but not the easiest to use for newbies. Then I discovered that if you install the express edition, you can't actually write any native windows application.... Huh???
Instead, one has to download the Microsoft Windows Platform SDK seperately before you can use VC++ Express Edition to write Windows program.
Instead, one has to download the Microsoft Windows Platform SDK seperately before you can use VC++ Express Edition to write Windows program.
Tuesday, August 1, 2006
Web 2.0?!?
The Web has gone through many cycles but there seems to have always been two main groups that influenced it's direction: engineers and publishers. The web started out as static content but it was originally a realm of the engineers who learned HTML and coded everything by hand. To publish something, you need to understand computers and possibly had to set up your own web server.
As it grew it fell into the publishers' hands as they started investing into online business. A lot of early commercial sites were simply online publications where the layout reflected that of magazines and newspapers. I wonder how much real growth in the industry happened during this time other then that the web go prettier?
The next cycle shifted back towards engineers as they transformed web sites into web applications. This is when the web really became "useful" and began to be part of our daily lives. Now the web isn't just for looking at information, it allowed us to find information. We didn't just read financial news on the web, but now we can trade stocks on the web.
I think now we might at another transition point. The web as an "application" is more the norm then a new idea. It's a proven commodity and that means the publishers are now wanting to get back on board and take over. The result is that they've coined the term "Web 2.0". Web 2.0 seems to me that it's just publishers saying that they want to be the ones putting stuff on the Web. I'm not necessarily opposed to it, but Web 2.0 has nothing to do with technology.
As it grew it fell into the publishers' hands as they started investing into online business. A lot of early commercial sites were simply online publications where the layout reflected that of magazines and newspapers. I wonder how much real growth in the industry happened during this time other then that the web go prettier?
The next cycle shifted back towards engineers as they transformed web sites into web applications. This is when the web really became "useful" and began to be part of our daily lives. Now the web isn't just for looking at information, it allowed us to find information. We didn't just read financial news on the web, but now we can trade stocks on the web.
I think now we might at another transition point. The web as an "application" is more the norm then a new idea. It's a proven commodity and that means the publishers are now wanting to get back on board and take over. The result is that they've coined the term "Web 2.0". Web 2.0 seems to me that it's just publishers saying that they want to be the ones putting stuff on the Web. I'm not necessarily opposed to it, but Web 2.0 has nothing to do with technology.
Subscribe to:
Posts (Atom)