Ever heard the expression - if it aint broke, dont fix it? Terrorist may not be stupid, but until what they are doing stops working they will continue to do it.

I noticed that the MI 5 study is very specific, and not a general condemnation of profiling.  Perhaps you will want to reread it with that in mind?  Not all terrorists are arabs.  Not all arabs are terrorists.  But then you have a better chance of catching a terrorist (plotting against the US - the UK has its own problems) if you profile an arab male, than if you profile a scandanavian grandmother.

MI5 are great, so long as they don't step foot in the London Underground.

Don't portray the UK's services as being good just because they have a history of this stuff, the UK isn't the same as it was even twenty years ago, and it's history with the IRA is hardly a gleaming report. MI5 simply instigated mass-tortures of Catholic Irishmen the likes of which make Abu Ghraib and Guantanamo Bay look perfectly normal.

MI5 have gotten so disjointed in normal operations, you could call the John Charles De Menezes incident the MI5 equivilant of Katrina for the rescue organisations in New Orleans. The difference being, MI5 haven't shown that they've improved.

As a Brit, I tend to look to the FBI for the cutting edge stuff, ironically

Profiling is not about catching all, but in increasing the chances of catching some.  Profiling should never be used to the exclusion of all else.  If you do, then yes, Mrs. Gruendsnik from Oslo will catch you with your pants down.

The problem with having plenty of arrests is that you might not be picking up terrorists. Guantanamo Bay is a case in point. Thousands of arrests, thousands sent home without charge after years of miserable treatment. More time needs to be devoted to establishing guilt rather than a profiling match.

As Dr Guy said, there won't be that many sent home without charge. If there were, you'd get so many lawsuits about wrongful arrests. I was thinking more in general terms, though, such as people carrying guns in places where they're illegal (or in the UK, stop-checking teens for knives which I can vouch would return enourmous dividends) as much as terrorrists.

In terms of profiling for terrorism, I think it can work very well, but it is always important to remember the convictions of terrorists and it's not all this lust for virgins, it can be driven by hate, envy, anger, despair, or even by sub-normal intelligence (like the British example with the brain-damaged guy who was brainwashed in... Exeter I think). That would make profiling much more dificult, because it's such an extensive profile. But generally, if you look at religious-terrorists they are Aarab/East-African, male, and living comfortably in the West. It's weird, but I haven't heard of poverty-stricken terrorists before.

In this case, it can work very well - you know the kind of places they have in common (the mosque) and the kind of groups they move in. You know their predominant gender, and their predominant race. It's quite easy, from there, to focus your efforts on about 1.5% (900,000) of the population in the UK, if you only take religion/race/gender into account (note sure the figure in U.S.) and, although there will be times when they're exactly the opposite of the profile, those times will be few and far between.

Profiling becomes harder in places where terrorism is rife, like in Iraq, where they're beginning to use women, Islam is the norm, etc. and that's the time when profiling doesn't work. But it's all about the time and place - IRA are dead easy to profile, religious-terrorists are just as easy (if not easier) but Russian spies would be impossible to profile.