Testing the Zune Content Ingestion Pipeline

I'll admit I haven't watched many videos on C9 since way back in the scoble days, but I wasn't really impressed with this one.

First, the whole thing seemed like a commercial for the DB version of Visual Studio, which would be fine if you showed me how it helped.  All I saw though was what looked like normal T-SQL, and I don't see what benefits there could be over using something like SQL Server Management Studio.  It didn't sound like she's firing up Visual Studio every night to run all these tests, but rather that they were automated.  I guess I can take Meher's word that it saves her time (and might save me time), but if I was pricing out new dev software I'd need a little more convincing.

Also, I felt the interview was too rehearsed.  Meher was standing there ready and waiting, which really contrasted with the "impromptu" feeling of walking down the hall with Mairead.  And having the stuff already drawn on the whiteboard seemed a little too well planned.  IDK, I suppose that you are actually saving me time, but the "feel" just seemed a bit off.

Finally, there was just a lack of substance.  Sure, I know that the data moves through a few databases on the way to the final database, but there was nothing really cool.  Things like statistics, problems that have been encountered, oddities, etc.  Knowing there are 130 tables was kind of interesting, but that was just in a transitional database anyways.  How about how many TB data are in the main DB?  # of GBs that get processed per night?  Who the content providers are?  Do you have to completely reprocess the data if there were errors?  How do you get the data again to reprocess it (physical media or what)?  Do errors ever slip through into the production catalogue?

Maybe my interests are just way different than what this interview was designed to be about, but it just seemed disappointing.