The comprehensive changes sound generally good, abstract though they are. Since we're dealing with abstractions, I'll keep it abstract:
If English schooling is anything like American (apologies, I haven't got much of a clue--I'm coming at this as I hear it as a "yank"), it will take time to implement these changes, and by the same token, it's comparatively easy to write a wishlist of changes that "sound good." The pessimist within murmurs that this seldom translates to real systemic improvement, which demands close attention to detail, a willingness to piss off certain obstructionists to realize a certain goal, thoroughly invested critical thinkers at the helm, and long-term planning. When I hear Brits describe their government, those are seldom the things suggested (seems to be a pretty universal phenomenon), but we'll see. This could turn out well for them. I'm not going to start on American schooling.
To the points about all the new comp/info sci, and the programming in particular, there are several problems with this that I notice:
First of all, teaching elementary or secondary school-aged kids multiple programming languages is a terrifying proposition. Firstly because it won't really work for most (if any) of them, for the same reason the "New Math" programs in the US failed (most children learn concretely most of the time; those few who can reason abstractly won't see the value in it until they have a grounding in the intuitive, concrete principles, which New Math glosses over) and for other reasons as well, like the fact that much of any abstraction will utterly fail to impress the other 80% of the kids in that age range. Secondly because programming languages are idiosyncratic and all demand practice, for which there's no substitute. It's far better to practice and gain experience with a single language than it is to swim in several (for the experience of saying "Oh! The language wants me to do it this way." rather than the expectation of growing up to become an [insert language] guru). Are we to assume English schools are going to create "immersive" programming environments in which children learn to bootstrap every facet of applications they themselves build to do the rest of their schoolwork? Thirdly because programming languages change so quickly that all but the general ideas and practice of building logical systems will be obsolete by the time they're adults. Fourthly because I seriously doubt the UK really has the resources in terms of trained and practiced teacher-programmers to do this right. All of these taken together mean that if this proposal is realized, it will waste a lot of time, money, and resource.
My personal advice to the British government and education system (because they have no doubt been sitting on the edges of their seats in anticipation of my pronouncement) is to bootstrap computing education with little exercises accomplishing little but hopefully meaningful tasks like turning a lightbulb on or off using a very simple programming environment and a very straightforward language. Make no mistake: this is a feet wettening exercise and little more. Anyone who shows extra interest can be given supplementary material. Then you slowly show how programming is the same thing as algebra, according to the same principles--slowly enough that just about everyone gets it. You ramp up the involvement of math in solving problems in science and computing. Use tons of visual aids: graphs of functions and networks and et cetra. Before they're through middle school, they must take a basic logic/critical thinking/statistics/epistemology course to establish correlation doesn't equal causation and how argument works. By the time they're in high school, the school should be organizing internships in computing-related fields and sending especially bright students to take classes at colleges and universities (full disclosure: this is what I did for a lot of my high school and it's one of the best things that happened to me). At every stage, for all the reasons mentioned, and the fact that you'll be having 40-60-year-old teachers trying to provide some of this instruction, you must stress keeping things simple and actually getting the important concepts.
This proposal isn't perfect, but it's something to work with. Importantly, it's much more realistic than the "teach kids several programming languages so they can reason abstractly and build robots before they hit puberty" notion.