I will keep this short. A couple weeks ago I went to Star Wars in Concert. Amazing show. It was a very different way to experience the story. The music was flawless. Here are a couple videos I took.

Shame on you ATI for putting the Catalyst Control Center item in my right click menus without me agreeing to it, and not allowing me to remove it. This bugged me enough to find the registry key and remove it. What sucks is that every time I upgrade the drivers it will most likley come back. This blog post detailed the key to remove.

• Remove ATI Catalyst Control Center from the Desktop Right-Click Menu

Haven’t we all got to a point where software developers who make installer packages, and the business leaders that direct them, have some common sense. Doing stuff like this without the users knowledge and not giving them the choice is what got the world into this whole spyware problem. The worst part about this, is that putting options like this in the installer is a trivial configuration setting. But I suspect we will never get past software makers, trying to sneak other software onto user’s computers.  Apple is one of the worst. Nefariously forcing an install of the Mobile Me control panel, and sneaking an install of Safari in when running their software updater. This topic really should be a new subject in every computer ethics course across the world. It is the Tragedy of the Commons for consumer software industry.

So one thing that has been bugging me about Windows 7 is the jumpy navigation pane in the explorer. I am trying to figure out what Microsoft’s logic for this odd behavior is?

If I use the arrow to expand the folder, and there is only a few sub folders, few enough to already fit in view, it will not jump. If there are many sub folder, more than will fit in the current view, the root folder you are expanding jumps up to attempt to fit all the sub folders in view. If there are to many it jumps to the top.

Double clicking on the folder sometimes will operate in the same fashion. But, most of the time it will jump the root folder to the very bottom of the view but just high enough that some of the sub folders are in view. Sometimes if there is a couple folders, the folder will all be in view. This is the most annoying behavior for me. And there is no way, as far as I know, to fix it. There are a handful of threads on this issue and nobody has seem to figured out a fix.

• http://social.technet.microsoft.com/Forums/en/w7itproperf/thread/17a83954-8c3c-4d45-9dd6-03f6cdd1604c
• http://social.answers.microsoft.com/Forums/en-US/w7files/thread/fb07b088-bc03-42dc-a8d6-79b360154d7c
• http://social.technet.microsoft.com/Forums/en-US/w7itproui/thread/a934513a-e1a0-4457-b558-46c0de1a5362
• http://social.answers.microsoft.com/Forums/en-US/w7files/thread/29d69da6-0f15-427b-9b07-7360910fc871

Unfortunately, even with all threads going back to RTM release over 6 months ago, Microsoft hasn’t posted that they acknowledge the issue. Clearly there is something wrong here. There was a post in one of the threads talking about hopefully getting a fix in SP1. I am not holding my breath. I have to either live with this or find a explorer replacement, of which there are plenty. I just don’t know if I want to go down that road yet.

Many posts in the threads talk about the “Automatically expand to current folder” feature. But this really has nothing to do with navigating in the left navigation pane. It is related to navigating in the right pane and the left pane getting automatically updated. I can almost understand this behavior. However, the same odd jumping behavior seems to happen with sub folders. At least this one can be turned off if you don’t want it.

I also find it very annoying that I can’t reorganize or remove the Favorites and Libraries from the navigation pane easily. There are some registry hacks to do this, but seriously, just give me a check box option in a right click menu somewhere. I really don’t even care if it is hard to get to. I will find it eventually. To remove Home Group I simply had to stop the Home Group service. Again, not the best way to do this, but easier than applying registry hacks.

Also the auto hiding of the folder expansion arrows bugs the hell out of me. Just leave them there. I like knowing if there is sub folders under the folder in view without having to move my mouse over there.

I have been dealing with SQL Server for many years now. I have dealt with trace flags to alter behavior and fix issues for quite some time. Over the past couple months I have ran across a couple issues where I had to use trace flags to fix a problem. One was with our OCS database server, the other was with our Microsoft Project 2007 database server. Both Microsoft products oddly enough.

Our OCS database server would experience serious performance problems every couple months. After about a year of this happening, we finally figured it out that the TokenAndPermUserStore was growing too large causing a bottleneck. The whole issue is detailed on the SQL Server CSS site. We were already at SP3 for 2005, but we still needed to turn on some trace flags to alter some behavior to fix the problem. I set the startup parameter and interactively turned on the ones I wanted. This all went down a couple weeks ago so I have been monitoring the cache size since and observing the trace flags not having any effect. The are on, verified by running dbcc tracestatus (-1). We even failed the cluster over manually thinking that turning them on interactively doesn’t take effect. After a failover, the trace flag parameters appear in the startup messages of the SQL Server log, but they weren’t showing up when running dbcc tracestatus(-1). I turned them on interactively again hoping they would take effect, but no luck. I was baffled why the trace flags weren’t doing what I expected until today.

The other issue happened about a year ago with Project Server. It was related to a SQL Server generating a bad execution plan for a query used to build it’s cubes causing tempdb to grow and causing the query to take hours vs minutes. The issue is discussed on this Project Server Blog. I set the trace flags in the startup parameters and turned them on interactively using so I didn’t have to restart to have them take effect. Problem solved. Then, about 4 months ago they started complaining about slow cube build times again. Another team member took the issue so I backed off and let him deal with it. A couple weeks went buy and I followed up and the issue was still open. Being too busy myself I didn’t have the time to dig into it. Fast forward a couple months and now my other team member is on vacation so a user contacts me directly. One, I was surprised that the issue wasn’t solved yet, and two I had no idea what had been done so far by my other team member to pick up where he left off. I decided to push some other things aside and try to figure out what was going on. I fired off some emails to people that were dealing with the issue to get some more background. Waiting for replies I decided to check over the server and see if I could spot anything. One thing stuck out. Running dbcc tracestatus(-1) returned nothing. Thats odd, the trace flags are in the startup parameters.

After having both these issues on my plate I realized I really needed to do some testing around enabling trace flags. Why weren’t the trace flags for the Project Server database server not on? I took a sandbox server and copied the startup parameter string exactly. Restarted and sure enough, the trace flags were not on. Here is the string after the standard parameters. Can you spot the problem?

"; -T1204; -T4101; -T4121"

To a rookie this would seem fine. I figured I would humor the Microsoft developers and try to insert a space between the “-T” and the trace flag number like below.

"; -T 1204; -T 4101; -T 4121"

No go. Humoring Microsoft even further, I tried removing all spaces like below.

";-T1204;-T4101;-T4121"

Bingo they worked. I then added the space between the “-T” and the trace flag number and that worked also.

In the official SQL Server documentation on trace flags, it doesn’t mention anything about whitespace and semi colon placement. It does however mention the difference between capital t and lower case t.

When specifying a trace flag with the -t option, use an uppercase “T” to pass the trace flag number. A lowercase “t” is accepted by SQL Server, but this sets other internal trace flags that are required only by SQL Server support engineers.

Until now I never knew there was a difference, I always just used a capital t out of habit. And in most cases I didn’t put a space after the semi colon when adding them in, except in these two cases for some odd reason. It is just natural for me to just add the space after you type a semi colon. You would think that parsing these parameters would handle the whitespace but what do you expect? The code to do the parsing was probably brought over from when Microsoft bought SQL Server back in the day and never touched because it always worked without the whitespace.

The other thing that sucks about this is that now I have about 450 SQL Server instances that have to be double checked. Writing a script to do it won’t take that long, but still. What a joke. I normally don’t rip on SQL Server, but this is one of those little common sense things.

Both the open issues are not fully solved yet, I want to do some more research into the Project Server issue before turning them on. Looking back I think the issue started up again when the server was rebooted for some OS patching, so I am pretty confident that just getting the trace flags on will solve the problem again. For OCS, I need to fail over the database server so that the trace flags are turned on at startup to see if the size of the cache stays limited as Microsoft says. We will see.

The moral of this story: when adding trace flags as startup parameters, you can not have a space between the semi colon and the “-T” parameter.

Here is my mini review of the two apps and which I like about each one. My use of both of these services is limited to the iPhone clients and their websites. The short, I like Gowalla over foursquare. Here is why.

I like the look and feel of Gowalla over foursquare. The icons in Gowalla are much more colorful and informing. The gray dull look of foursquare doesn’t do it for me. The categorization of the spots that is in Gowalla is much nicer also.

Friend management between the two were similar. Also, the badge/achievement system that each one does is comparable. Also, the whole mayor thing in foursquare would be comparable to the leader in Gowalla.

The list of spots in foursquare is pre-populated. In Gowalla, the list of spots is not pre-populated. All the spots are user created. A pre-populated list can be good because you don’t have to waste extra time creating a spot if you are the first to check in. However it can be confusing when you can’t find the spot you want to check in at. And you can’t manually add spots. At least I couldn’t figure it out. There was a lot of spots, like small parks and stuff that just wasn’t in the list. Plus, I like being the creator of some of the spots. There is a small bit of prestige along being the first.

When trying to checkin, foursquare doesn’t sort the list by closest proximity to your location. Which inevitably leads to having to search for the spot taking extra time. Gowalla sorts by proximity to your location this and it makes it much easier to find the spot you want to check into or quicly realize its not created yet and create it.

In Gowalla you have to actually be near the spot to check in there. In foursquare you can check into any spot anywhere. I somewhat wish Gowalla did this because sometimes my GPS is so far off that it won’t let me check in, or the spot is so big and I am not close enough to the original check in point. In some of those cases then you have to create a secondary copy of the spot just to check in. But, the amount of spam check ins in foursquare just ridiculous (which is part of the reason why so many people think foursquare is winning), so if I had to choose I would prefer the Gowalla method.

After you check in on foursquare you can’t see a history of the people who have checked in at that spot. You can only see the people who have recently checked in (aka there). I Gowalla you simply see the check in history of the spot, which I like better.

I really don’t care about the whole gathering items game that Gowalla is adding. I somewhat play along because its there, but there is no desire to want to obtain more items. But its something over foursquare that adds a bit extra to the whole thing.

Foursquare does have all the commercial promotion stuff built in. Which from what I have seen it isn’t intrusive at all. So I could consider this a slight bonus for foursquare because you have the option to get a promotion. However I generally hate advertising embedded within apps. If either app had annoying banner ads I wouldn’t use them.

So there ya go, a quick list of some of the pros/cons and why I like Gowalla better. I haven’t tried out any of the other ones like BrightKite, Loopt, or MyTown. The gaming aspect of MyTown seems a little to over the top for me. Not sure I would like the competitive nature of that. I may try out Loopt, who knows.

It has been a few days now on Windows 7. The upgrade went fairly smooth. Only a minor issue with my hard drives. If you were following my Twitter feed, you would have only heard about all the bad stuff that happened. Really, things went pretty well. Most of my issues are related to the lack of real 64 bit support. Even the 64 bit version of iTunes isn’t really 64 bit. Only a couple drivers are, iTunes.exe isn’t.

I wonder how long it will take for true 64 bit applications to be the norm?  Unfortunately, the demand really isn’t there yet. And because 32 bit applications still work fine, it will be a slow conversion. I am assuming the average computer user doesn’t even really know what 64 bit means, let alone be able to distinguish between the two. Really, I could see another 3 years before it becomes mainstream.

So far I haven’t had any showstopping issues with x64. I just bugs me that I have to install “inferior” 32 bit applications in my crisp clean 64 bit environment. Here is some of the programs I have installed that are 32 bit.

• Adobe Reader
• ffdshow (there is a 64 bit version, but I didn’t use it)
• Haali Media Splitter
• iTunes (the iPod service is x64, the rest is 32 bit)
• KeyPass
• Medieval CUE Splitter
• Games for Windows LIVE Client
• Microsoft Office
• Windows Live Photo Gallery
• Sliverlight
• FireFox (there is 3rd party complied 64 bit binaries out there, but I heard that extensions and plugins don’t work)
• Picard
• QuickTime
• Steam
• uTorrent

This is a majority. And every game I have installed through Steam has been 32 bit. In fact, the only application that I use that is truly 64 bit is 7-Zip. At this point I am not too upset. Everything works at least.

To prepare for upgrading to Windows 7 I thought I would make a list of all the programs, not including any games, I use on my PC. This first group is the stuff I use pretty frequently.

• FireFox
• iTunes
• Steam
• uTorrent
• Office 2007
• Picasa
• KeePass
• Picard Tagger
• Ventrilo
• 7zip
• Bionic
• Video Codecs
  • ffdshow tryouts
  • Haali Media Splitter
  • Media Player Classic
  • Divx

The following things I use less frequently.

• Amazon MP3 Downloader
• CDisplay
• Curl RTE (for Puerto Rico Online)
• FileZilla
• Medievil CUE Splitter
• ImgBurn
• Audacity
• Paint.NET
• DAEMON Tools Lite
• CCCleaner
• Java
• HighPoint RAID Management Software

I thought I would post this quick for the community benefit. Here are some indexes that I create on all my SQL Server machines to speed up some default purging routines. The default indexes that come in msdb just don’t cut it.

-- To speed up execution of sp_delete_backuphistory
IF NOT EXISTS(SELECT * FROM msdb.dbo.sysindexes WHERE id=object_id('msdb.dbo.backupset') AND name = 'ixFinDate')
CREATE nonclustered INDEX [ixFinDate] ON msdb.dbo.backupset (backup_finish_date ASC)
go
 
IF NOT EXISTS(SELECT * FROM msdb.dbo.sysindexes WHERE id=object_id('msdb.dbo.backupset') AND name = 'ixMedia')
CREATE nonclustered INDEX [ixMedia] ON msdb.dbo.backupset (media_set_id ASC)
go
 
IF NOT EXISTS(SELECT * FROM msdb.dbo.sysindexes WHERE id=object_id('msdb.dbo.restorefile') AND name = 'ixHistory')
CREATE nonclustered INDEX [ixHistory] ON msdb.dbo.restorefile (restore_history_id ASC)
go
 
IF NOT EXISTS(SELECT * FROM msdb.dbo.sysindexes WHERE id=object_id('msdb.dbo.restorefilegroup') AND name = 'ixHistory')
CREATE nonclustered INDEX [ixHistory] ON msdb.dbo.restorefilegroup (restore_history_id ASC)
go
 
-- To speed up deletes from sysjobhistory
IF NOT EXISTS(SELECT * FROM msdb.dbo.sysindexes WHERE id=object_id('msdb.dbo.sysjobhistory') AND name = 'ixRun')
CREATE nonclustered INDEX [ixRun] ON msdb.dbo.sysjobhistory (run_date ASC, run_time ASC)
go

Every once and a while I will replace a track in my music library with a better quality one. This little VB script that I wrote allows me to manually set the playcount of a track so that the new one has the same playcount as the old one. Enjoy.

Dim iTunesApp, currTrack, newPlayCount
Dim prompt, title, defaultValue
 
Set iTunesApp = WScript.CreateObject("iTunes.Application")
Set currTrack = iTunesApp.CurrentTrack
 
prompt = "New playcount:"
title = currTrack.Artist & " - " & currTrack.Name
defaultValue = currTrack.PlayedCount
 
newPlayCount = InputBox (prompt, title, defaultValue)
 
If IsNumeric(newPlayCount) Then
  If newPlayCount >= 0 Then
    If Len(newPlayCount) > 0 Then
      currTrack.PlayedCount = newPlayCount
    End If
  End If
End If

A few months ago when I really started to get into Twitter, I had the brilliant idea of setting up a locally running script that would automatically tweet the current track I am listening to in iTunes. I know there is built in integration with some of the music services out there, but I don’t use any of those. I figured someone out there would have done this already, but after a few minutes of searching I didn’t find anything so I whipped something up. Here is the whole script. Explanation is following.

dim sUrl, sUsername, sPassword
 
sUrl = "http://twitter.com/statuses/update.xml"
sUsername = wscript.arguments(0)
sPassword = wscript.arguments(1)
 
 
dim iTunesApp, currTrack
dim CurrentTrackName, CurrentTrackArtist, CurrentTrackAlbum
dim PrevTrackName, PrevTrackArtist, PrevTrackAlbum
 
Set iTunesApp = WScript.CreateObject("iTunes.Application")
 
while 1=1
 
  Set currTrack = iTunesApp.CurrentTrack
 
  CurrentTrackName = currTrack.Name
  CurrentTrackArtist = currTrack.Artist
  CurrentTrackAlbum = currTrack.Album
 
  if PrevTrackName <> CurrentTrackName or PrevTrackArtist <> CurrentTrackArtist or PrevTrackAlbum <> CurrentTrackAlbum then
    wscript.echo "track change detected, tweeting update:"
 
    sMessage = "Currently listening to """ & currTrack.Name & """ by """ & currTrack.Artist & """ from the album """ & currTrack.Album & """"
 
    wscript.echo sMessage
    wscript.echo
 
    HTTPPost sUrl, sMessage, sUsername, sPassword
 
    PrevTrackName = CurrentTrackName
    PrevTrackArtist = CurrentTrackArtist
    PrevTrackAlbum = CurrentTrackAlbum
 
  end if
 
  wscript.Sleep 3000
 
wend
 
 
Function HTTPPost(sUrl, sRequest, sUsername, sPassword)
  set oHTTP = CreateObject("Microsoft.XMLHTTP")
  oHTTP.open "POST", sUrl,false,sUsername,sPassword
  oHTTP.send "status=" & sMessage
  HTTPPost = oHTTP.responseText
End Function

There isn’t any real standard for naming objects and variables or formatting. The script simply runs an infinite loop with a 3 second delay each iteration. It takes your Twitter username and password as parameters. It uses the iTunes.Application COM object to obtain the current track information. It saves the values to see if the song is changed from the previous iteration. If changed, it simply sends POST request with the value of the tweet. Simple.

A more advanced implementation would use sc.exe to create a Windows service that runs this script in the background. But beware, every time you change song, it tweets is. Your followers will get very annoyed very quickly.