My Extended Blog List (Q3 2009)

As part of the 24 Hours of PASS, Steve Jones recommended that bloggers actively maintain their blog roll (blogs they follow).  A little while ago, my blog roll was fairly short - seven blogs, to be exact.  In part, this was because the list of blogs I (currently) follow is incredibly long - at least I think so.  I was also only listing blogs that I was of the opinion were very on topic, and extremely helpful.  I do realize that point of view is valid, but I also understand why Steve recommended expanding such lists - or at the very least, I have my own interpretation of what Steve was trying to communicate.
Reasons To List a Blog Roll
Quid Pro Quo
(This is the "self serving" reason.)  If I tell you who I listen to, it's possible that those bloggers will hear about it.  If they know I blog and listen to them, they're somewhat likely to visit here, and there exists a possibility that they may start following me.  There's then a slight chance those bloggers may then decide to add this blog to their blog roll.  Sounds like Drake's equation... but maybe there is intelligent life elsewhere...
Establishing a Respectable Community
There's always search engines, aggregators, and even community sites like Steve runs.  If someone is looking for help or information, they'll probably be able to find it there.  But search engines don't rank on "trust".  (Maybe a little, indirectly.)  If you frequent this blog, and listen to the opinions and advice given here, then it follows that you have extended me a level of trust higher than a search engine.  If I then tell you that I listen to some other blogs, it should carry forward that those blogs also deserve a little more trust than a Google or Bing ranking.
Encouraging Commuity
At this point, I'm a "little guy" - at least I think I am.  I don't get much traffic - but that gives this point even more weight.  Anyone who ventures far enough into the outer reaches of the SQL/SSIS community to find me must not mind the neighborhood.  If that's the case, then they probably already know just fine how to get back downtown - but a little direction around the suburbs could be just what they're looking for.
My Reader
My list is (I think) rather long.  I believe that part of the reason is my reader and how I read.  I use Viigo on my BlackBerry, which gives me the freedom to not read feeds on my computer.  I don't know how unique I am for not having read basically any feeds using desktop software.  I do this for a particular reason: when I'm at my desktop (laptop, actually) I have the capability to get real work done.  When I'm away from my computer, and have a minute of time to kill, it's very easy to make that time at least partially useful by catching up on some news.  If you happen to have a BlackBerry, check out Viigo.  It needs (IMO) some improvements - especially on the desktop interface for managing feeds - but for usefulness on the phone itself, I haven't found any better.
The Blog Roll
I;ve updated the blogger blogroll, but there's no way I'm going to list all of these in that widget.  (That would just be too much work.)  Here they are in no particular order and uncategorized (maybe next time I'll do that):
CNET News
CBC Top Stories
Gizmodo
How To Change The World (Guy Kawasaki)
Lifehacker
TechReport - Articles
TechReport - News
SQL Server Central - Articles
Engaget
The Daily WTF
Paul Thurott's WinInfo
MS SQL Server Performance Team Blog
William McKnight
David Loshin
Presentation Zen
Database Journal
Ed Bott's Microsoft Report
Jamie Thomson (Old)
SQLTeam.com
Cracked
John Welch
MS PerformancePoint Team Blog
MS SQLCAT Team Blog
Techdirt
MS SQL Release Team Blog
MS Business Intelligence Team Blog
MSDN SQL Server Headlines
ZDNet Emerging Technology
SearchCIO
SearchDataManagement
Craig Schiff
SQLIS.com
Tom's Hardware - News
MS SSRS Team Blog
The Onion - War For The White House
Paul Thurott
WeGotServed.uk
Simon Sabin
MSDN Canada
MS C# Headlines
MS CodePlex Team Blog
Kevin White
Wade Dorell
Phil Brammer
Matthew Roche
Michael Entin
Applied Business Intelligence
Geoff Hiten
Paul Ibison
Donald Farmer
Mosha Pasumansky
Jimmy May
MS SSIS Team Blog
Alberto Ferrari
James Nix
Mick Horne
Douglas Laudenschlager
Krish Krishnan
Sid Atkinson Jr.
Cringely on Technology
Brent Ozar
TechReport - Blogs
Escape From Cubicle Nation
Sara Ford (CodePlex tagged items)
SSWUG Articles
MS Canadian Developer Connection
Jean-Rene
Rushabh Mehta
Rafael Salas
Andy Leonard
Jessica Moss
Duane Douglas
John Myers
Brian Knight
Dan English
Christian Cote
Joe Webb
SQL Server Integration Services - Community Buzz
SQL Server Integration Services - Recent Downloads and Resources from Microsoft
Devin Knight
Sarma Musty
Kirk Haselden
James Beresford
Microsoft Connect: Most Recent Feedback - SQL Server
BI Tinkering
SQL Lion
Benny Austin
BI Twibe Blog
SOA Blogs
Jitu Raheja
Karlo Bartels
Tomer Lev
Altius Tech
Jamie Thomson (New)
SearchWinDevelopment - News
SearchWinDevelopment - Architecture
Bootup Labs
Merv Adrian
Ronald Damhof
Steve Dine
Dan E. Linstedt
Mark Madsen
Susan Major
Ron Powell
Dan Power
Neil Raden
Colin White
Lyndsay Wise
Steve Jones
MS Windows Home Server Team Blog
Scott Stauffer

I Am A Twit

At least now nobody can say I'm not averse to trying new things - even those things that sound absolutely rediculous on the surface.  I mean, I can understand how "cute" somebody was trying to be calling the thing "Twitter" (hee hee).  I even agree that if the service is called that, it makes sense to call the messages "tweets".  But it's a little lame to call those who tweet "twitter-ers".  Come on - that's just bad English. 

You're twits.

And now I'm one of you.

Let's hope this doesn't consume my time like a black hole, and instead provides some quick ways to find out what's new and interesting in what I'm interested in, and provides a meaningful way to interact with any of you who are interested in these things too.  If not, I'll let the two birds play in the bushes...

Using CNAMEs for Flexibility

I'm no "real" DBA - I'm an "accidental" one.  (And relatively bad at it.)  My servers probably have performance and disaster recovery holes large enough to drive tractor trailer rigs through.  It's a good thing I have some backstop from my hardware guys - who are just about as accidental as I am, but from the other direction.  Sooner rather than later, somebody's going to have to bite the bullet and get the gaps filled in to become the "real" DBA.  (I voted for them.)
Anyway, I wanted to pass along a neat trick that those gents in the server room came up with.  At least it sounds brilliant to me - and is working very well in practice.  In the process of planning to migrate some servers and applications around our server farm, they suggested we start to use CNAMEs to decouple our infrastructure a little bit.
What's a CNAME?
Truly, I don't understand the exact technical definition.  I would read this, but I have other non-network administration things to do, and I prefer to kill brain cells with C# and Hefeweizen.  (Separately.)  In layman's terms, it's an alias for a computer's name.
What Do You Need An Alias For?
Our main issue was how inflexible we were in being able to redistribute applications and services around our server farm.  From time to time, we needed to move a database, a web application, a networked service, or something else from one server to another.  We aren't completely virtualized - but even in the places we were that didn't help us out completely, because moving virtual servers from metal to metal just wasn't granular enough.  Doing piecemeal moves of apps and services wasn't extremely hard - but it can be complicated and involved to orchestrate taking something offline long enough to change configuration settings and the like to refer to the new environment.  And it's rarely easy to ensure that all the little "bits" out there get redirected to the correct server.
How Does a CNAME Help?
Again, the CNAME is an alias - and it's all in how you use it.  In our environment, since we're not a very large shop, we tended to name our servers from a "hardware" point of view - PRISQL for example for our "primary" SQL Server.  Again, since we're a smaller shop, we only used the default instance.  It's great to map server software to hardware that way - if you're focused on managing the hardware infrastructure.  But if you need to manage the software infrastructure that uses that - and need to move some of it off that server and onto SECSQL... you're looking at manual labour.
Using CNAMEs allowed us to create a set of application-specific names for services, and map those to hardware resources.  For example, we have an application well call "3PS" that uses a Citrix server for remote access, an App server for the business layer heavy lifting, and it (of course) has a database on SQL Server.  Lets say that those three parts of the application resided on servers named CITRIX04, APP03, and PRISQL.  We created three CNAMEs for this system: 3PS-CTX, 3PS-APP, and 3PS-SQL, mapping them to the respective "physical" servers.  Once this was done, if we wanted to move the database to SECSQL, we could completely and easily do so by changing the CNAME alias instead of reworking vendor-specific configuration.
I don't know if it will help for you, but it sure did (and continues to) for us.

Pumped for the PASS Summit 2009

I was just at VanPASS' chapter meeting for September, nicely convened at the Kingston Taphouse.  It gave me the opportunity to ask about the PASS Summit experiences of local PASS chapter members.
Being relatively new to this "community" thing - I'd never been to a PASS Summit before.  I "missed" last year's because I was snootily attending the Microsoft BI Conference the week prior (which was pretty good - but not great) - and I didn't know enough to twist the boss' arm to get me into PASS as well.  Since there's no MSBIC this year (next one's in 2010) I felt I should probably get to PASS.
But apparently I don't have enough influence to redirect our reduced IT budget from non-essentials like "offsite DR" and "UPS backups" in order to pay for conferences.  Isn't my continued education important?  We can throw the DR servers in my garage with a few car batteries, no?  I jest.  We're operating on a very tight budget - probably just like most of the rest of the world these days - and there's (unfortunately) no room for conference spending.
I thought this one over a lot, and eventually decided that the ten or so sessions I "had to see" warranted shelling out my own money to get to PASS.  I attempted to completely discount the fact that I'd get to meet up with some fantastic guys I'd bumped into here and there over the past year.  I had to, or the whole exercise of objectively weighing the pros and cons of going would get shot out the window.  You see, the conference is going to be a ton of fun - sharing information, experiences, and getting new tidbits of information.  And when something's fun, I immediately question how objective I can be about the business value of it.  Regardless of how beneficial all that information gathering and sharing among colleagues is in real objective terms - I'm highly suspect of putting any kind of large valuation on that.
So the practical and objective side of me says it's worth it to go in order to understand SQL logging and recovery, and definitely learn more about the next two or three steps in my BI journey: AS, RS, PPS, MDX, and on and on.  Enough said.  But taking it over the top is the spectacular opportunity to get some handshakes, beers (I hope), and conversation in with John Welch, Matt Masson, Joe Webb, JR, and whomever else that's going to be there that I've had the fine fortune to get to know.  I don't yet know where I'm staying - so if I haven't dropped you an email yet, send me one so we can not-so-haphazardly run into one another.  I'll see you in Seattle!  (I'll apologize in advance for blathering on about SSIS and bringing you up to date on the spectacular extensions to it that I'm involved with - the Kimball Method SCD component, the File Properties task, the Pause task, and the SSIS Extensibility Directory.)

I'm Looking Forward To: Windows 7 "XP Mode" Version 2 or 3

I'm not saying that "XP Mode" in Windows 7 isn't a good thing already - but we all know that Microsoft (or any company) needs some feedback and usage data in order to really polish off a product or feature.  What I'm certain that the current wave of interest in virtualization technology will do - eventually - is free us from the shackles of backwards compatibility.
The Backwards Compatibility Jenga Tower
How many layers of software are currently between your latest app and the hardware it runs on?  Let's count, making an assumption you're using Silverlight.  (If you're using Java, WinForms, or anything else, you may not start out in the same place, but your stack will look just as disgusting.)

• Silverlight
• Internet Explorer
• .Net Framework/COM
• Windows Kernel/Win32 API
• BIOS, Intel's segment/offset memory, and "AMD x64" emulation

I said this was a Jenga tower, which isn't quite accurate, but I was going for the "rickety stuff built on top of other rickety stuff" feeling.  (Even though those technologies are pretty darn rock-solid.)  Even if all those layers work well - there's still a bunch of layers in there that effectively duplicate functionality.  Each layer is intended to "abstract" complex details below it, and make using the system "easier".  The problem is that each layer is only doing part of that job, and the lower layers are doing it pretty badly - as they were designed very long ago.  Despite their architects' best intentions, those lower layers are getting in the way, but they can't be removed or significantly altered because that would pull the rug out of the upper layers.  (There's the Jenga metaphor again.)
The Attempted Solutions
Solving this problem has been tried before, of course, at many different levels.  The .Net Framework is intended to replace COM... yet COM lives on.  RISC processors from Motorola et al, and Intel's Itanium "real" 64-bitness tried to change the foundation, but those are virtually dead.  We've seen the claims that taking a different architecture from the bottom would cut a lot of fat out of the middle and allow that tower to be streamlined and double, triple, or quadruple effective FLOPS.
Why didn't those attempts work?  Some, like .Net, won't truly "work" because they can't automatically upgrade older apps that use COM - COM is still required, which means all it's lower layers are required.  .Net may streamline that "layer" while providing more flexibility and usability, but it cut out the requirement to still have COM in Windows.  Rearchitecting the lower layers (chipset) failed miserably because it shook the very foundation of the tower, causing ripples of instability in "ported" OSes and applications, and lack of availability of drivers and compatible apps.  Enterprises don't deal well with instability, or systems with one-dimensional support.
So We've Stuck With This Sh*t?
I'm hoping not.  Virtualization technology could be the way out.  Up to now, it's been used to consolidate servers, and run "old" OSes that may not accomodate newer hardware advancements.
What I'm looking forward to is what I see as the inevitable conclusion to this trend that "XP Mode" seems to be the first step of.  Wouldn't it be great if Microsoft could cut the Win32 API and COM entirely out of Windows?  At the same time, they'd have to rework the .Net Framework to plumb deeper/closer to the hardware.  But wait, doesn't that put us in the same place as prior failed attempts, screaming coming from enterprises that run ten year old software?  Not with seamless virtualization, it doesn't.  If Microsoft can get that "seamless" part right - and that why I'm sure it will take a couple versions - then an "old" application requiring COM running on Windows 8+ that doesn't have it will simply start up a virtual host with a Windows OS that does.  Sure, we've just added (at least) one more layer that old application has to fight through to perform - but we're seeing results today where the virtualization layer isn't that onerous a burden, and there's always the continuous progress on the hardware side that should more than make up for the added fat.  (And who needs great performance from old apps?  Rewrite them if you do! :P)
I See a Win/Win/Win Situation
If Microsoft can pull that off, they suddenly open the door for much faster current generation applications that use the best framework/interface/whathaveyou that Microsoft's put in Windows for us to use.  And they've guaranteed backwards compatibility for old apps until virtually forever, and given themselves an upgrade path for radical reworking of those layers in the future.
If Microsoft can demonstrate this new "backwards compatibility" architecture as reliable and relatively efficient, they'll get plenty of buy in and interest from other "layers" in the computing supply chain.  They'll be able to work with the motherboard/chip manufacturers to rework those layers to benefit the "new", while only having to "maintain" backwards compatibility for one application - virtualization.
What do you think?

One Year Milestone

I've seen other bloggers do this - and reflection's not a bad thing.  It's a little late, so I won't take long here, but I thought I'd pass on some interesting tidbits.  (At least they're interesting to me.)  Unfortunately, I only have Google Analytics data from January on, so that will have to do...
Most Hit Posts
(Page views per day that the post was available - otherwise this would be severely skewed to older posts!)

1. Converting Strings to Dates in the Derived Column Transform
2. Inserting Records AND Getting the Identity in SSIS
3. SSAS Quick Reference: Attribute Key Cannot Be Found

Most Reviewed Posts
(Highest "average view time"... although this seems to be unreliable.)

1. Converting Strings to Dates in the Derived Column Transform
2. Upgrading Custom SSIS Components from SQL 2005 to SQL 2008
3. General Quick Reference: Batch File Tips

Most Commented Posts

1. Converting Strings to Dates in the Derived Column Transform
2. SSAS Quick Reference: Attribute Key Cannot Be Found
3. (tie) Converting Numbers to Dates in the Derived Column Transform
   File Properties Task for SSIS

We'll see what happens this year...

Microsoft's BI Certifications

Duane Douglas posted a thread for discussion on the Integration Services forums a while back that I'd been thinking of myself.  The basic thrust of the conversation has been whether the current structure of the Microsoft Certifications relating to Business Intelligence make sense.  Indirectly, it also questions the value of such certifications in general - which seems to be a perennial question.
In my opinion (shared by several others in the thread), the current certification makeup doesn't make tremendous sense - although I can understand why it would be the way it is.  Currently, there are two certifications - Microsoft SQL Server 2008, Business Intelligence Development and Maintenance, and Designing a Business Intelligence Infrastructure Using Microsoft SQL Server 2008.  In general, the "TS" (Technology Specialist) certifications are geared toward lower-level (as in nuts and bolts) as well as lesser-experienced candidates.  The "ITP" (Information Technology Professional) certifications tend to target higher-level (as in design and architecture) as well as more-experienced candidates.
What's Wrong With That?
In reviewing the TS and ITP certs for BI last year when I decided to take one, I found not much difference between the two (on paper).  Not having taken the ITP exam yet - I can't comment directly on the relative difficulty of the two exams.  (And by the time I take it, if I do, I'll surely have forgotten all about the TS exam.)
The difficulty that Duane, I, and others have with the current structure and philosophy of SQL BI certifications is that it doesn't truly seem to reflect the path of progression of a DBA/developer/information worker/whatever you might call yourself as it relates to the SQL Server BI stack.  From my own experience, it seems as though one would have to concentrate in one "product" (such as Integration Services, Analysis Services, or Reporting Services), learn it, then move on to the next.  Now, I can totally see that not everyone would progress from IS to AS to RS.  In fact, I can see individuals hired to work on one, and only one, of those products in a larger shop.  However, that also seems to lay more of an argument against having just one "TS" exam.  If I was hiring for an AS position, there's no doubt that someone with the current "TS" certification would get good consideration - but would I be unnecessarily excluding someone who had barely worked with IS or RS, which I don't care about?
Taking another tack on the same issue - what's the utility of the "ITP" certification when compared to the "TS" certification?  It seems like a question of perspective at the moment - implementation (TS) versus design (ITP).  But what's the progression path to the ITP certification?  How does one get there in practical employment?  By designing, planning, and working with all three products and making sure they interoperate smoothly.  That's different than the requirements of the TS level... but not that different.
So, Mr. Smarty Pants - What Would You Do?
It certainly seems to me to make an awful lot of sense to split the TS level certification into three.  The BI certification space in Microsoft would then look (and function) as follows.  There would be one exam per BI stack "product": IS, AS, and RS.  Completing one of those exams would earn a corresponding TS certification.  Completing all three - possibly in conjunction with completing a "design-level" exam - would earn the ITP certification.
What Are You Waiting For, Then?
Now, this obviously hasn't been done for some very real reasons.  The first and very obvious one that I can think of is cost vs. benefit.  If you take a look at the current numbers of Microsoft Certified people, it's quite clear that the BI certifications aren't all that "popular."  The space just hasn't matured all that much for Microsoft (...yet... hopefully).  As of writing this, there are 4,191 BI-related TSs, and 1,626 BI-related ITPs (05 and 08).  Compare that to 55,461 TSs and 12,501 ITPs for SQL Server non-BI roles.  All of those other SQL certifications are earned through single exams.  If the cost/benefit equation doesn't justify expanding (at least) the TS exam offering for SQL Server administration to different flavours - Installation/Configuration, Backup/Recovery, Database Design, Security, etc... - then how is adding exams and certifications to the BI stack justified?
Unfortunately I think we're kind of stuck with the current situation unless something changes on the numbers.  Either the numbers of people involved on using the BI stack, or Microsoft's costs of developing and administering the exam and certification process.  Given those constraints, it may be more worth our while to suggest amendments to the exam creation process, rather than the exam variety itself.  In doing a little research on this topic, I stumbled across an older blog post by a then-employee of Microsoft Learning, talking about "crowd-sourcing" certification exam questions.  I didn't read through everything there or follow every link, but I think the concept certainly bears investigation. But seeing as that post and discussion are years old... either that discussion's been had and died, or nobody was really listening during the "boom" years.  Personally, it would certainly make sense to me to have "the crowd" submit questions for "expert" review/rework - with that review and rework done behind closed doors so as not to leak exam contents.
You're Just A Ray Of Sunshine Today, Aren't You?
Better to be practical than overly optimistic, I think.  So don't hold your breath about having the Microsoft SQL BI stack certifications change anytime soon.

Filtering RSS Feeds

In reviewing the Wiki for my Kimball SCD component on CodePlex, I found that there had been some comments placed there that I'd never known existed!  When I'd first set up the project, I sent the whole list of RSS feeds available for the project to Viigo on my BlackBerry.  I soon grew very tired of watching my own activity pile up and obscure any interaction I may have been getting from other users - so I unsubscribed from everything except the Discussions and Workitems sections.  Unfortunately, that caused me to miss two comments placed on my Wiki pages.
Well, I had to solve that problem!  I've added a request to CodePlex (always better to fix the root of the issue, I say) - but who knows if or when they could solve it.  So a quick Bing of the 'net (still sounds weird, but I'm giving it a shot) found me FeedRinse.  It's a free service that basically allows you to place "allow" and "block" rules on any RSS feed, generating a new RSS feed from the results of the filter.  The interface is very easy, and it certainly seemed to do the trick, allowing me to filter the project Wiki feed to only pass through articles with "Comment" in the title.  Excellent!
Maybe it's time to filter some noise out of my other channels... mwuahahaha...

24 Hours of PASS

The short: A fantastic FREE program put on by PASS!
I didn't attend every session, obviously - I'm no Jack Bauer (or Thomas LaRock) - but I did happen to catch all the sessions I wanted and each one was fantastic.  Regardless of how much you think you know about a given topic, there always seems to be at least one thing you pick up that improves your practical knowledge.
If you didn't happen to catch them live - yes, the speakers were live - the recordings are scheduled to be made available for a limited time on the main PASS website in mid-November.  Kudos to all involved, the production value was high - everything went smoothly on the technical end with slides, video, and audio smartly delivered.
Recap
Here's a (very) brief review of the sessions I attended...
Delivering Good Performance Consistently with SSIS - John Welch
(This was hour 4 - about the time when the first layer of the "conspiracy" would have been revealed...)
An exceptionally good (especially given the time constraints) coverage of "what can be slow" in SSIS.  From parallelism in the Control Flow to diagnosing memory (buffer) issues in the Data Flow - a soup to nuts review of what you need to watch in SSIS performance-wise, what "easy" arrangements within a package typically lead to poor performance, and simple techniques to improve efficiency.  John must have also seen my name somewhere in the list of the hundreds of participants, because he gave my Kimball Method SCD component a shout out right at the end.  Thanks, John!
Reporting from Analysis Services Cubes Using Excel 2007 - Peter Myers
(This is hour 14 - not being Jack Bauer with a portable caffeine drip, I would now be completely confused about who the newly identified criminal syndicate is, and what nefarious plot they were 15 minutes away from executing..)
A very good look at how to get at Analysis Services Cubes using Excel 2007 - and even a look at what publishing those spreadsheets into Excel Services on Sharepoint can deliver.  I definitely picked up some tips on little things you can do contextually within a PivotTable to hide members of hierarchies, filter, and group data.
Data Warehousing: Laying the Foundation to Success - Erik Veerman
(This is hour 15 - rogue states using nuclear weapons have to be part of the plot now - ah, yes a DW topic...)

Great coverage of the technical side of starting to understand Data Warehousing.  IMO, that's the "easy half" - there's the political side to manage as well, but I'm not sure you can put that process on a slide deck...  The newer part for me was the discussion on server infrastructure.  Not something I've had to dive into yet, but sooner or later I'm going to have to - if just to have a meaningful conversation with my hardware/infrastructure guys.  Read up on Microsoft's Fast Track reference architectures for yourself.

Building a Better Blog - Steve Jones
(This is hour 17 - just about the time when Jack gets blindsided by a "personal" angle in the plot...)
I was really looking forward to this one.  Personally, I think my blog needs some work - or should, as everything has room for improvement.  I also like to hear non-technical talks too - we shouldn't always get all wrapped up in APIs, performance monitoring, and coding patterns.  Unfortunately, the links on PASS's site for this session didn't work!  Live Meeting claimed the meeting info wasn't recognized, or I wasn't authorized. :(  It took me 15 minutes to hunt down an alternative link (bad searching protocol by me?)
Back to the content!  Excellent, again!  Steve gave plenty of great guidelines and examples of what he considers good ways to improve your blog.  Steve's point of view was primarily that you should use your blog not to educate others, but simply chronicle your career, giving potential (future) employers more "meat" to review.  The many reasons I'm blogging do overlap (educate, chronicle, participate), so I plan to make some changes based on his thoughts:

• Expand my blog roll (damn, that'll be long)
• See if I have Pingbacks enabled
• Try to write less per post :)

Loading a Data Warehouse in SSIS - Brian Knight
(This is hour 19 - now and only now is the final depth and horror of the full conspiracy revealed after having chased and dismantled three or four paper tigers...)
Not to diminish Brian's presentation - it was just fine - but I've seen it before :)  I knew that going in, but it's a good idea to review content once you think you understand it - you'll pick up more.  In my case - using HASHBYTES... I'd seen that before, but I need to look into it again.  For me - that probably means looking at incorporating it into my Kimball SCD component.
What a Great Deal
Wow - that was a great set of presentations, and I know the ones I didn't see were just as good.  The moderators said that "select presentations" will be available on PASS by September 9th, and all of the content will be available "for a limited time" in November/December.  Review the session list and keep a watch on the PASS site for your opportunity to catch sessions that you should have to pay for!
(Now, if this had been an actual season of 24, I would have seen every episode.)