8800GT TV Out solution

As you may already know, I recently bought a new PC. My new and cool PC is powered by Nvidia's Geforce 8800GT chipset based graphics card. When doing the market research about which graphics card to buy, I read comparisons, benchmarks and relied upon other people experience, thinking "Hey, I have a good experience with Nvidia, this looks like the best buy". Well, I learned that while all of this is important, it's even more important to look for known problems. Especially of other people using the same features you are planning to use.
A simple web search shows many people are having many problems connecting these cards to their TVs. The two problems I'm aware of

• DVI-to-HDMI would pass audio signal, even though there is no audio output. This would prevent the TV from playing sound coming over other cables (such as RCA).
• S-Video won't detect your TV, which forces you to use component or composite connections for CRTs. This sucks.

As for the later, I just found a solution. The easiest solution is to use Linux, and configure xorg.conf on your own (heard there's a wizard for configuring that, I should check that out). This would allow you to define your exact parameters for the connection: S-Video using PAL-B (Israel), and 1024x768 60Hz (written as 1024x768_60) should do the work.
If you insist using Windows, I can dispense some advice, assuming you use XP (don't have Vista, don't want to). The problem arises from the drivers (forceware) and Nvidia's control panel. It seems that the new control panel assumes you want your TV to be auto-detected, and if the auto-detection fails, you won't get any other option. Moreover, even if the control panel's wizard thinks it has detected your TV, but it actually fails, you are stuck. The most common phenomena is the TV screen flickers for a second or so with your desktop, and than returns to it's blue state. In the mean while, the driver is sure that everything works perfectly, so no further configuration is allowed. This is where most of the forum threads in the Internet says "replace the 8800GT with something else". And then I found this very helpful lead. Basically, it suggests to get back the classic Nvidia control panel, and try to configure stuff there, as the rest of his advice, it didn't help me. So i found this help site that explains how to do that. After changing the value of ContextUIPolicy to 3, I got the classic control panel. Now, I could use the wizard that says "My display is not in the list" and then turn on the "'Rigorous Display Detection". Finally, the TV connection wizard allowed me to use other settings then "Auto Detect", so I could choose "S-Video". After doing so, my TV came to life, and started to display my desktop as "Monitor 2". Some final tweaks were required, such as setting the region to "PAL/B" (otherwise, screen flicker would occur in Israel and western Europe) and setting the resolution, and that's it! Notice that the lead I found only managed to get B/W picture, while I managed to get full color.
Hope this guide saved someone from deep frustration, and happy TV watching everyone.

I've failed the Turing Test

Finally, I can see the end of my M.Sc in computer science. Almost all exams are over, and only the final project is to be completed. My projects advances kinda slow lately, but I hope things would get better. The topic of my project is spam (as in email-spam) and mechanisms to fight it. It is partially research, and partially technical. In order to test different web-mail vendors' spam filtering mechanisms, I've created free accounts at GMail, Yahoo! and Hotmail. Each of these vendors use the CAPTCHA anti-bot mechanism, so their services cannot be easily exploited. I've seen many of these CAPTCHA images before, and answered them correctly without a problem. And then came Yahoo!. According to them, I'm probably not human, as I failed the Turing Test three times, before I eventually managed to guess what's in the image, and pass the registration process. They had, by far, the worst looking CAPTCHA images.
I have a test of my own for computer software. My test tells if some software is usable or not. I call it "My Mom's Test". The test is quite simple: my mom is trying to use some software, with less than 1 minute training (zero is better). If she succeeds - the software is usable for everybody. If she doesn't - that's trash, and I remove it from her PC. Skype, for example, passed that test. The same test applies for web sites. Some sites are easy for her to use, while other sites frustrates her. For example, Google's search page is easy. Yahoo!'s registration process would fail "My Mom's Test" over and over again, as she would call me a hundred times before we can together "break" it's anti-bot and anti-human defense mechanisms. Therefor, it's unusable for my mom, and probably many other normal people. Too bad for them, as the profit from free services (such as web mail) comes from advertisement, which would bring more money if my mom would use it (she clicks ads).

The smell of something new

You know when buy something which comes wrapped, and you tear the wrappings off, there's a nice smell of something new. Well, I got to smell that two days ago. I bought a new PC. This is the spec:

• E8400 CPU (Dual Core, 3GHz)
• Asus motherboard, based on Intel's G35 chipset
• 2 Gigs RAM (800 MHz)
• Western Digital 320GB HD (SATA 2, 16MB cache)
  
• Nidia 8800GT based graphics card
• The rest of the stuff you put in a PC
• 20" LCD
  

Now I can play The Witcher, NWN2 and other games using their maximum settings. That's fun.
Also, installed 64bit Ubuntu 8.04. The new Ubuntu really rocks. Solid, fast and cool.
It took me some time to decide where to buy the new PC (or PC parts). After a long survey, I bought it on KSP. The decision was rather difficult for me, as I personally built my previous 2 PCs. Today I saw this post regarding Ivory (important: read the comments and follow-ups), and I'm more than glad I didn't go there. How rude can someone be, especially when your job is to give service. When my work included service-giving, it never crossed my mind to act like that.

Back to the fun part. Now I have 2 PCs, a laptop and some other WiFi enabled devices, so it was time to setup up a WiFi router. Bought an Edimax router. The setup was very easy, including WPA2 encryption (actually the encryption is AES in WPA2). The firewall and NAT were a bit tricky, but 10 more minutes and all was up and running. Now I have gazillion devices that all require electric power in the same room, so it's time to buy a little power-plant.

Correction

Today I had an exam in communication security (SSL, IP-Sec, Crypto Hash, etc.). The exams are the main reason I hadn't blogged in a while. But excuses aside.
One of the questions was about how a CA signs some certificate (attribute certificate), and includes the validation key of Alice (so Bob can be sure that Alice really signed the data in the certificate). Anyhow, the lecturer, who wrote the exam, had a mistake. He wrote A.s which means Alice's signature key (private), while he meant to write A.v which means Alice's validation key (public). So, when our lecturer found out about the mistake, he wrote to the board "Q2 s/A.s/A.v/". I don't know how many students appreciated the little joke during the exam (did I mention it was difficult?), but I'm sure it made me smile.

Proud

Computer-people becomes proud very easily. Every time a big chunk of code works, we brag about it. If we make a piece of software work a little faster, we can't stop showing off. Not to mention how happy we are when we do something big such as a major upgrade.
For me seeing a review written in my native language, about an open-source project I've participated in, does the trick. Actually, I left that project as I currently haven't got the time to volunteer, but I'm still very much connected to it.

Web 2.5

Few hours ago I had a conversation with a Sharepoint administrator I know about the new Google Sites service. He told me that their announcement was the talk-of-the-day, and couldn't go unnoticed. We talked about what threat could the new service be for MS Sharepoint or Office at all.
For those of you who are familiar with organization/enterprise oriented portals, the idea of Google Sites is not new: an individual, a team or an entire organization, can build up a site around an idea (a project, a shared point of interest, etc.) simply, quickly and without having to write a single line of HTML. Afterwards, they can add to this sites, lists of objects, such as documents, pictures, spreadsheets, calendars, task-lists, etc. This is done very easily, and with full integration to the user's desktop applications (MS office, OpenOffice, Web-based office, etc.).
Now for the exciting part. Google's product might not be better than the competition (yet) in the organization level. BUT, they are the first to bring such a polished, full-featured product, to the Internet. Now, everyone can create sites around ideas, projects, shared interests, etc. Using the same tools they are familiar with: Google Docs, Picasa, Youtube, RSS feeds and such.
I like to call this new level of services: Web 2.5. Why? Because in Web 2.0, the control over the content of the web, was passed to the individual. Now, the control over the content is passed to group of individuals – collaborating. Imagine how powerful it might be for a team of students, working on the same project, to collaborate using such tool, and share their results with the client (lecturer or a real client, doesn't matter). This is the evolution of Wikis.
Of course there are limitations: 10MB per file, integration with GoogleTalk yet to be made, workflows around documents aren't possible, and a lot more. But hey, we can call it Web 2.49 until then.

SysAdmins who write code

At work, under my team's responsibility, there's a dozen or more systems, which have few dozens development environments. This adds up to a large number of servers (with many virtual servers) which runs variety of enterprise software. Managing all this software, requires each sysadmin to be expert in our favorite language: Perl. Also, since all the servers are Windows, some cmd scripting knowledge is required. Rarely, for our internal development, other programming languages are used, such as C# or PL/SQL. Today i read, on the only Microsoft blog I read, that in Windows 2008, sysadmins who write code would have full power. This means that PowerShell will allow us to do what we already do very well, and enable us to do that for MS products (never wrote a Perl script to administer IIS).
According to my experience, this PowerShell technology exists for a year-and-a-half, and yet I haven't written a single PowerShell line of code. As for the MS products we administer (Windows, IIS, etc.), we just don't write complex (more than cmd) scripts for them. I only hear the Exchange guy crying about MS removing functionality from the Exchange administration GUI, and that now in order to do simple stuff, he has to use PowerShell. Brutal marketing that is.
Now don't get me wrong, I'm all in favor of finally being able to control the rest of our systems using scripts. It's just weird that the way to do this was to remove functionality from the GUI and to wake up that late (who knows when we will upgrade to Win2k8 with the new IIS, etc.). I'm all for .NET technologies (I use mono at home), but I don't see how I'm going to replace "du -h" (we use MKS/cygwin) with "get-childitem | measure-object -property length -sum" (taken from wikipedia), especially if one day I'll replace a Windows server with a Linux server, and all my scripts would mean nothing.