Does Neo-Darwinian evolution require change ?

You have yet to show this. Also, as a population dwindles over a few generations this can reduce the deleterious mutation load far easier than in a growing or stable population. You can have a cycle of booms and busts that negate genetic meltdown.

You can also reduce a mammalian population from millions to a few thousand and re-establish the species.

Just in case the horse isn't dead . . .You have made the mistake of conflating functional with transcribed. Even taking the kindest view of the ENCODE results, the best one can say is that nearly all of the genome is transcribed at some point. However, just because an mRNA is produced does not indicate that it has function. Even more, the number of transcripts per cell is so low for most of these mRNA's that function is doubtful. To use an analogy, you are looking for a signal in the static that doesn't exist. Sometimes you just have to admit that it just is random noise.

In the short term, yes. The American Bison went from 10's of millions to just a few thousand and they are now making a comeback. The same for the walrus. Other species, such as the cheetah, show definite signs of a sever bottleneck indicative of intense selection.

And for each round of genetic recombination in the gametes there is a loss of multiple mutations. Also, individuals with a higher load of deleterious mutations will be selected against allowing for selective sweeps. The bust cycle resets the mutational load, so yes it has changed. You are only measuring the load since the last bust cycle instead of the origin of all life.

You are forgetting about genetic recombination and selective sweeps. If selection is looked at through the lens of reproductive success then you up the chances of hitting a beneficial or compensatory mutation that can then spread through the population separately from the background of slightly deleterious mutations found in the individual in which the beneficial mutation occurred. IOW, you up the chance for beneficial mutations. Are you taking the size of the E. coli genome into account? The human genome is 1,000 times larger than the E. coli genome. So is the deleterious nature of these mutations. If their affect is so slight as to be a non-factor then it really doesn't matter.

To use the analogy I used before, this is like saying that since you can hear static on your radio that this opens up the possibility that there is a radio station on every part of the dial. At the same time, there are sections of the genome that have accumulated mutations at a rate consistent with neutral drift. These include processed pseudogenes which were the first to be called "junk DNA". If they do serve a function it is probably independent of the actual DNA sequence. The more we learn of genetics the easier it is to find junk DNA. Even more, scientists have removed over 2.3 million base pairs from the mouse genome with no visible change or loss of function. This is one half the size of the E. coli genome. Obviously, there is much more vestigial DNA in the eukaryotic genome than in the prokaryotic genome."We deleted two large non-coding intervals, 1,511 kilobases and 845 kilobases in length, from the mouse genome. Viable mice homozygous for the deletions were generated and were indistinguishable from wild-type littermates with regard to morphology, reproductive fitness, growth, longevity and a variety of parameters assaying general homeostasis."
Megabase deletions of gene deserts result in viable mice - PubMed

Mattick leaps to the conclusion, as seen in this sentence:"Moreover, it is now evident that these non-coding sequences are transcribed in a dynamic manner, to produce tens, if not hundreds of thousands of noncoding RNAs, and that most complex genetic phenomena are RNA-directed, which suggests that there exists a vast hidden layer of regulatory RNAs that control human development and brain function."Mattick leaps from "transcribed" to "has function" without ever showing that there is function. Also, if Mattick is relying on the ENCODE data sets I think most would agree that it is way to early to make any judgements. Some have even challenged the ENCODE data in that some of their data may be related to shorter products from known genes. Also, the number of transcripts from "junk DNA" is very low compared to RNA from known genes.I would say that Mattick's conclusion is not supported by the evidence in hand.Edited by Taq, : No reason given.