h0bbel.p0ggel.org

To follow up on the Microsoft Mythbusting Video Stu has dissected a Microsoft technical case study called How Microsoft Designs the Virtualization Host and Network Infrastructure.

Head on over to vinternals and read Microsoft Myths and Realities…. As Stu concludes, VMware doesn't need marketing when Microsoft provides us with this kind of real life examples on how their infrastructure works.

Thanks Microsoft, you just made my job much easier with regards to explaining to management why Hyper-V just doesn't cut it.

On the 3rd of April 2009, Microsoft released a video which was supposed to "bust" 10 "VMware Myths", but as far as I can see the entire video has backfired pretty badly.

I'll let some of the brighter minds highlight the video content and dissect the myths, but I have to say this: Microsoft, please, if you want to play this game do it right. Make your own cases, don't try to disprove other people by highlighting their errors, erroneously. That doesn't work, sorry.

Gabe posted "When myth busting goes wrong and becomes a myth to bust it self"

Stu posted "Microsoft Hyper-V 2.0 - NO Memory Overcommit!"

Be sure to read the comments on the original video as well. Can anyone spell busted? Or is that coming in R2?

Finally I want to highlight a post by Eric Sloof: "When I’m feeling blue" hits the sweet spot with regards to just how mature and stable Hyper-V is. I guess busted is a part of v1.0 after all.

Let me just start off by saying that I am in no way a javascript wizard. I'm even hesitant to call myself a developer of any kind as most of the "development" work I've done has been of the hack'n'slash type. For the most part I can get something started, or hack something existing to work the way I want it to.

This post isn't about my ineptitude as a developer though, but rather about the discoveries I made while investigating jQuery based plugins to do some photography presentations on the photo365.org site I recently hacked together.

All Lightbox inspired solutions were automatically disqualified, at least in the long run, as thats been overdone basically everywhere by now. I do use a Thickbox on the site right now though, but thats not the solution I really want to stick with in the end. Consider it a placeholder for now.

After looking for pre-made plugins that could help the site display the photos, I was left with a list of potential candidates:

• slideViewer
• Galleria
• Galleriffic
• uGallery

All of these, besides Galleriffic have one major flaw, at least how I see it. They all download the full size image you wanto to display and then resize them in browser. Not only is this slow, but it's a huge bandwidth eater. Trent Foley’s Galleriffic however, does this a bit differently. Like all the others, it's based on an unordered list that holds your images, but instead of just doing a simple html img off the fulle size image, it implements a way to show a specific thumbnail that links to the full image you want to display. You can even link to the original image for download purposes if you want.

In my particular case, where I on each of the participating photographers "profile page" will end up with showing a total of 365 images on a single page, the bandwidth waste is very obvious.

To break it down, we end up with some number similar to this (Based on 1 random image from the pool):

• Full Size: 54993 bytes
• Thumbnail Size: 6165 bytes

Now, if you add all of this up you will end up with a calculation that is something like this:

• Resize in browser total image size:
  365 images x 54993 bytes = 20072445 bytes = 19.14 Megabytes
• Separate thumb and full size display:
  (1 x 54993 bytes) + (364 images x 6165 bytes) = 2299053 bytes = 2.19 Megabytes

This equals an approximate difference of 17 Megabytes transferred on each page request that this particular profile page gets. I could of course split my profile pages into smaller chunks and help relieve some of the bandwidth usage per page load that way, but my I want to display all of the thumbnails all at once.

As I said, I'm no javascript wizard and I'm sure there are good reasons why most of these plugins behave the way they do. Grabbing the full image and then resize it "locally" seems to be the easy way to implement something like this and I guess easy is what we are all looking for. Doing this does make the html markup very easy, and if you don't have thumbnails readily available in your setup it's fine. In my case, I do have thumbnails and I want to utilize them properly.

Did I miss some great jQuery based photo showoff plugin that I should have looked at? I'm sure there are several that consider the needs of bandwidth concious users out there, and that I've just missed them. If that is the case, please let me know.

Now I'm off to try and actually get Galleriffic integrated into my photo365 project site.

I've been attending various Microsoft Technet events (link in norwegian) over the years, for the most part every time they find time to grace the little city of Bergen, Norway with their presence. For the most part I've been pretty happy with the events, they provide a way for me to get out of the office and spend a day listening to usage scenarios and sometimes they have even provided me with some valuable input.

Obviously, this is a Microsoft event designed to promote Microsoft in all it's glory, but I've been OK with that since they have been putting on these for free. I can bare the brainwashing sessions, they more often than not provide some value I can actually put to use later. Other sessions have had me actually correct the speakers, but that is an entirely different story.

Sadly, for me this seems to be something of the past. Starting from the next event here in Bergen, Microsoft will require an early bird payment of 1000.- NOK + VAT, roughly $140 USD, to participate. After the early bird period passed, the price is raised to 1500.- NOK + VAT.

I can understand that the events have a cost, and I can see how Microsoft haven't been able to profit directly from them in the past, but does Microsoft really think that people will pay to attend them? In my case I would have to be away from the office an entire day, and then try to get my employer to pay the fee as well. The total cost of that is way over the actual value the event itself represents.

Even if Microsoft tries to "award" attendees with some kind of scoring system that could potentially win you a goodie bag of related Microsoft stash, or a "Change the world or go home" hoodie (?!), I suspect that many others will think the same way as I do. Normally these events have been fully booked, and you might not be able to actually get a pass unless you sign up early enough, my feeling is that they will see significantly less attendees than they have before.

Perhaps that's even their intention, as they also limit the number of student passes (half admission price) to 20 in each city. Even if I'm no economic guru, I do understand the current economic climate and why Microsoft would want to try to get some of their investment back. I just don't think everyone else feels it would be ok to both spend an entire day away from the office and pay to get brainwashed at the same time. Somehow I think that Microsoft would be better off spending a bit more on getting people interested and hyped up about their solutions, instead of charging them for it.

After all, the current climate might just be what Linux and other FLOSS solutions needs to really gain momentum in Norway?

It would be in Microsofts best interest to get their vision out there, and convince people. They won't be able to use these events to get any convicts now, as those who are looking elsewhere already certainly won't pay to be convinced otherwise.

Bad move Microsoft, bad move.

Does any other vendors like Microsoft charge for small events like this? I'm not talking about Lotusphere or VMworld, but rather small local "evangelism" meetups?

I've been using Thwirl for a long time as my Twitter client, and I was very happy with it. Then, all of a sudden, it seemed like the rest of the world had gone mad and started talking about TweetDeck.

I guess everyone knows about TweetDeck by now, but in short it's an enhanced Twitter client that allows split your data stream into smaller pieces. Groups, searches and other goodies are a part of the main interface, making it easy to administer larger amounts of twitter data. While this is the strength of TweetDeck, it's in some ways also it's weakness. Compared to other clients it takes up a lot more screen estate, but the thing that bugged me most was that currently there is no real easy way to keep your groups and other settings synchronized between different installs on different computers. Maintaining your user groups etc. on a computer by computer basis just isn't fun.

Thanks to @gabvirtualworld and @TweetDeck I found out that the configuration and settings TweetDeck uses is located in c:\users\[username]\AppData\Roaming\TweetDeckFast.[random string]\Local Store\ (In Vista/Server 2008. XP and others hav a slightly different path). In that directory there are two files that are of interest:

• preferences_[twitter_username].xml
• td_26_[twitter_username].db

If you set up TweetDeck on one computer, then copy those files with you to a new computer your get to transfer your current settings. You still end up with two separate installations to maintain though, but you do get to transfer your setup from one to another.

I wasn't happy with that either, hard to please I know, so I set out to find a way to have those files located at one place and have several TweetDeck installs point to the same ones.

Enter DropBox. DropBox lets you have a central storage for files and have that area syncronized to your local computer(s). I run several applications from my DropBox space already, like Putty and KeePass, so I figured there had to be a way that I could utilize DropBox for my TweetDeck configuration synchronization needs. Long story short; Yes. You can use DropBox for this, and still keep TweetDeck happy.

This is where the magic happens:

Sorry, but there is no magic, but it does work. The steps required are the following:

1. 1. Stop TweetDeck
3. 2. Move your preferences_[twitter_username].xml and td_26_[twitter_username].db from your initial "master" TweetDeck install, to a new folder on DropBox. I chose apps/TweetDeck
4. 3. Start a command prompt, and navigate to the c:\users\\AppData\Roaming\TweetDeckFast.[random string]\Local Store folder and run the following commands:

   mklink td_26_[twitter_username].db "[your path]\My Dropbox\Apps\TweetDeck\td_26_[twitter_username.db"
   mklink preferences_[twitter_username].xml "[your path]\My Dropbox\Apps\TweetDeck\preferences_[twitter_username].xml"
   

5. Repeat step 3 on any other machines you want to keep syncronized

The trick here is the mklink command. This creates a NTFS based symlink on your filesystem, that pretty much works the same way as a Linux symlink. In essence what you end up doing here is making symbolic links for the files that TweetDesk stores it's configuration in, thus making the different installs use the same set of files. Stroring it on your DropBox install makes it available for all the computers.

Of course, there is a downside to doing this and that is if you keep TweetDeck running on several computers at the same time. If you do that, you will end up with save conflicts in your DropBox, but you should avoid running Twitter clients on several computers at the same time anyway to avoid going over the rate limits in the Twitter API.