Unfortunately I haven't had much time to play with the early Beta of SQL 2005.  However, from what I have seen and heard, the locking will remain largely the same...

Thanks for the clarification. I had not thought of that interpretation of what I said...

It is generally very bad practice to wait for user input inside a transaction.  But the scenario you lay out would be a worst-case.  Not only would you have to...

I'm not a programmer and don't pretend to be one so I won't be the one to effectively address your COM+ transaction questions.  I could address the issue at a...

That is exactly what I meant, but I did not mean to imply that the cost was entirely paid in performance or that it couldn't be mitigated by other factors. ...

Thank you.  The next article (scheduled for publication on Nov 24th will deal pretty extensively with error handling.  The third (submitted but no publish date yet) deals with Locks and Isolation...

You misunderstand, I am emphatically stating that XML is NOT the answer to anything.  At its best XML is a bloated and inefficient method of data exchange.  I am aware that...

Incremental and Differential backups are the same thing, or the terms are used interchangebly.  However, differential backups DO NOT truncate the transaction log, only a transaction log backup does that.  So...

Don't register the nodes.  Use the virtual server name or IP.  The node you registered will cease working if the other node becomes the active one.  If you use the virtual...

Paste the query in Query Analyzer and then hit Ctl+L, this shows the estimated execution plan for the query.  You can then look at each part of the query and...

Sounds like a bad idea to me.  The ObjectID, as I understand it, is not guaranteed to be unique over time.  In other words, a given ObjectID will be unique...

David,

Let me guess...The vendor in question probably doesn't use any sort of DRI either.  That way nobody can make heads or tails of their database design (assuming there's any sense...

Whether the key is "natural" or a "surrogate" like IDENTITY makes no difference to this particular subject.  Yes it is always possible for someone who has direct INSERT/UPDATE access to the...

You are correct in saying that the JOIN to a MUCK table works, but what you are overlooking is that the result of the JOIN (or any other query) is only as...

Very well said!

I might add that the practical result of following the scientifically sound approach (the relational model) is that most of the complexity should be handled at the database level...