Yes, I read it. I also recall you previously believing that there was an implication of auto-spanning. When MS comes out and explicitly states that they are modifying the .NET framework to support auto-spanning - then I'll believe it.

Why was in-memory processing a norm in FP, but strongly recommended against in NET? The stated reason in the 2005 MS white paper referred to resource utilization, which is why SP was presented as the better option in NET.

Now Anders is talking about in-memory processing. How will he deliver that? And why?

My main interest is persistent, indexed, in-memory resultsets. My focus on memory-disk spanning since 2005 was because that's how FP delivers it while NET does not. If Anders has a way to deliver what I want without memory-disk spanning, cool!

It is still early in the game for Linq to SQL - but in looking at some of the SQL code generated by LINQ to SQL, I still believe that stored procedures are a better starting point.

Starting point? As opposed to the ultimate? ;-) OK, I'll go with that. Lets review as and when things mature.

You claim you just want to "share technology ideas"...I've been trying to find out if you're using the CTP or beta for anything, and you claim the question is demeaning.

After about 50 messages specifically expressing opinions about the CTP, a participant asks me whether I've looked at it myself. The question carries a message.

Keep in mind that you're the one who previously tried to make this about me. (or do I need to dig up prior modified posts?)

I'll post it myself. If anybody is interested (which I doubt), here is what KG is referring to- it relates to early in the thread when another poster had stated that the typed DataSet (as advocated by KG) was rendered obsolete by the Entity. Rod Paddock then posted that datasets can carry change tracking, and I replied:

That's why it's not correct to say LinQ for SQL entities/objects can replace the typed dataset in April 2007 unless you're doing 2-tier stuff. Pass the entity between layers and no change tracking details accompany it unless you concoct your own which is hopeless. IMHO. So KG gets to sit there with a big grin and I don't get to tell him he's using obsolete technology, which I'm sure you'll agree is a real shame. ;-)

It was addressed to Rod Paddock who has written books with KG but does *not* use typed datasets. If anything, it's a compliment- KG's views are borne out by research. KG interpreted it as some sort of attack on him and asked me to remove it, so I did. So there it is, of great interest to everybody I'm sure but posted now so that nobody can post oblique refererences and imply fault.