In which case it is not FRAND, because it discriminates against the individual and small startup.

It is not the term 'reasonable' which is being violated, but the term 'discriminatory'. Requiring per-item royalty payments not only discriminates against FLOSS but also against products which are sold (often at a high price) in source form for a one-off payment.

Also it would be nice if Google did not index the content of the ads. On numerous occasions I have found that the only occurence of my search is in an ad on the page.

Yes, but that is putting control in the hands of the owner of the device. Trusted computing and secure boot is doing the opposite and removing control from the owner of the device and placing it in the hnads of the supplier of one of the components.

Should they have to right to only allow 'official' games to played and prevent it running homegrown games or for third party independent developers to sells games to run on it?

Number portability should be for moving between providers while retaining the same number (to save having to give the new number to all contacts).

When I have moved a number to a new (PAYG) handset (keeping the same provider), the process required me to quote the IMEI of both handsets as well as answering security questions. For a contract phone (which one would assume is what a business owner would have), surely the only time the number should need moving a new handset is when the handset is changed as part of the contract - in which case it should not be possible to move the number simply by making a phone call.

They can bitch as much as they like, but what they must not do is count those overseas downloads as lost revenue and must not include them in their claims of the loss due to piracy.

I do not know about the USA, but in the UK although big business employs lots of people, the majority of people are employed by small/medium sized companies. While individually each of these does not contribute as much as the large corporations, taken together the small/medium companies contribute more to the economy than big business.

So what you do is calibrate the model on past data, then test it by generating some predictions and seeing if the predictions are accurate. Better still, use a number of different parameters all of which calibrate the model and then before actually using the model, wait and see which set of parameters generates the most accurate predictions. Only then use the model for actually predicting future events.

Maybe we need to wait for Hari Seldon to invent psychohistory (and for WikiPedia to morph into the Encyclopedia Galactica).