This discussion is becoming very baffling (for me at least - no big surprise).

1.  It seems to me that our team is happy to have ARM out of the instant litigation.  This is (IMO) precisely because if we were to settle with ARM, that may prevent us from pursuing their many, many customers.  That said, why would our team be interested in settling with ARM (unless it was for some unGodly amount of money)?  Lots of potential low-hanging fruit, so why cut down the tree?

Then you say: "The problem lies in the MMP Portfolio in that it is NOT individuated, it covers ALL patents or NONE."

2.  Begging your pardon, but, via the current litigation, the MMP has been "individuated".  There are seven patents in the MMP, and it appears that in the current litigation, we're down to two.  To resolve the current dispute, we need only settle on those two.  This leaves the door open for additional settlements with the Js regarding the remaining patents (including, ultimately, the '584).

3.  As for the '584, do you really think that the intent of our team in fighting this through the USPTO is necessarily so we can later go after ARM?  I seriously doubt it.  I think it is so we can, when/if we prevail, go after all those ARM customers. 

4.  As for your last question, I strongly suspect that IF we negotiated a settlement with the Js for the entire MMP, we could still go after their technology supplier (ARM) for their sales to all their OTHER customers.

5.  Would it be smart to settle with the Js on the entire MMP at this point in time?  It may be good from a "marketing" standpoint, but I see little other benefit.

6.  IF our team is confident in its effort with the USPTO, and prevails, the '584 will be "worth" a great deal more in terms of later settlements with the Js.

7.  Actually, if I were in control of our team, I'd take the '584 off the table in negotiations with all potential licensees - again, if there is confidence that we will prevail at the USPTO.  The "value" of the '584 at this point in time is questionable, at best, in any license with anyone.  License the rest of the MMP (MMP-1), and keep the '584 in your hip pocket.  That patent alone is so broad in application that another full round of licensing could occur.  And with a win on the '148/'336, and resultant dramatically increased licensing fees, the $ precedent will be set for that '584-only second round of licensing.

8.  I note that you mention indemnification (i.e., such a clause in ARM's contracts with the Js).  Though I was a strong believer in this probability in the past, there has been no clear indication that this was ever the case.  Possible, but not absolute by any stretch.  While indemnification would have necessarily brought ARM into the instant litigation, it may be that they joined in simply to argue claim 29 of the '584 (the one they thought they could effectively argue) for the benefit of their customers.  The reason I now very much doubt that there was ever a patent indemnification clause in place is purely because I'm convinced that ARM designs infringe on the '148 and '336 (especially since ARM references those patents in many of their own patent applications).  As you KNOW, patent indemnification clauses are all-encompassing (i.e., they don't selectively identify patents where indemnification is granted, as this would effectively be an admission of infringement by the granting party).  So if there was such a clause in ARM contracts, I would think they would have remained in the fight to battle contentions regarding these patents as well.

JMHO, and yet another novel.  Going back to number paras for ease of reference...

SGE