Welcome to the Q&A Forum

It can be good, to redirect users to simular content, but not how they have done it, but I think there are much better things to spend you time on.

What they have done is returned a Status code of 200, this will cause problems with soft 404s.

Also when you create a 404 page, you really should not have any external files such a images css or js files, this can cause a eternal loop if any of these files goes missing. a missing image will create a 404, that will call the 404 page that will cause a it to happen al over again. Crawlers will not like this. Bing once stated that they see your site of low quality of this is the case.

Is it interesting that people want to view it. If so people will link to it. another measure that search engines look at is clicks from search, if a user clicks on your searp, then retunes to the search and chooses another result, then the search engines assumes your result was not a good result for that search term.

technically, the content should crawl well, and have semantic meaning, you need to use schema.org to describe the content to the search engine to make sure that the SE understands what you are trying to say.

I would look at your links coming into those pages, and 301 the links only, I doubt all 100,000 pages have links.

A 301 only redirects link juice, so if no external links no problem, of cause you need to change your internal links. You should never 301 internal  as 301's leak link juice, you should fix your internal links.

I would look at it the other way around., am I concerned about what it fails me for.

yes I would go with responsive design, bootstrap is a good for layout,

yes I would try to get a good score on page speed,

Chris gives a good answer, but is it really a problem, bandwidth is very cheap these days, in fact here in Australia most accounts are unlimited,

I Host with Microsoft Azure and bandwidth is very cheap.

Sounds like nonsense to me, do they give any evidence to back this up?

I cant see how you can dilute a keyword, besides linking out with the keyword.

The reason people like to keep to one keyword per page, is that there is only one page title, one H1(should be only one), one url, not because more then one will dilute the other, and especially in your case, I cant see how the same keyword can dilute itself.
Don't keyword stuff of cause, but you are far from doing that.

1. unless you have a big thin content problem there is no gain

2. completely reversible, just remove and wait

3. you will have to decide, you seem like you are on the right track.

4. Question you should have asked, is there any downside to no-indexing these pages, Answer Yes there is, all links pointing to a no-indexed page will leak all their link juice, noindex is a last resort, I have never used.

if you must noindex a page, do it with a meta no-index,follow tag, note that was "follow", not "no-follow", then your link juice will flow into the page and back out again.

First do these pages have any external links?

if not then there is no gain from a 301, if some of them do, the redirect those pages

Good work Holmes

My guess is the large number of 404's was wasting your link juice(100%), also you will lose something from all the 301's (approx 15%),

as for the source code changing I would not worry about change itself. as long as the code has no problems. It a change in content that is the concern.

some other points,

2 weeks is not a long time for things to sort themselves out with this many pages.

You don't need to 301 redirect all the pages, just the ones that had external links.

Also just for interest, Bing treats a 302(temp) that has been in place for a long time as a 301, likewise a 301(permanent) that keeps changing is treated as a 302. Maybe google does the same, I don't know.

I would fix all the 404s and wait

I cant see any SEO benefit, if the images were on your own domain, then there might be.

this is my opinion, I believe there is a pagerank for images, at least that google keeps some sort of record of whos linking to what image. not only are these people not linking to your domain rather to the CDN domain, you are linking to the CDN domain also. I would rather host the images myself and not be linking out and have others linking to me.

I host at Microsoft Azure, if you host in the cloud there is no advantage to having a cdn as you already have good bandwidth.

A sub directory is never going to rank like the home page. you can look at linking to try and improve the rank http://thatsit.com.au/seo/tutorials/a-simple-explanation-of-pagerank I would try and push PR from all shop pages to the shop main page, not the root.

Myself I would create a sub domain shop.bouldersuk.com, you will stil be seen as the same site, but will be easier to have the main shop page seen as a home page

don't bother with w3.org for the reasons you point out.

I don't try to pass validation for the sake of passing, If the validation has logic behind it that concerns me then I take note.

I use Microsoft Visual Studio code analysis(fxcop) for server side code for performance and reliability , I use the JSHint and Web essentials (css)  for client side code, I use the Bing SEO API for SEO, all of these are built into Visual Studio, I also use the IIS Bing SEO Tools for a more detailed look at the SEO.

I just ran a site of mine though w3.org and they gave me 3 errors, all nonsense

for example
Line 5, Column 59: Bad value X-UA-Compatible for attribute http-equiv on element meta.

This is the correct tag to tell Internet explorer how to render the page, if you listen to the w3.org, then you page will not render correctly in IE.

no I don't think so, you would be cleaning up their index for them.

yes, but they should be noindex,follow so that the link juice flows back out of the pages, if no index,nofollow then the link juice will be lost

no not if your not indexing it, your doing what they want, cleaning up the index, but still letting the crawler find links to other pages

Just to add.

I believe the advise was, if you have a .au, .uk or in your case .be, they its not a problem. but if you have a .com then server location may be used. that's just from memory.

No, I have seen those videos and that was my belief up until a year or 2 ago, when I saw this contradicted, either in a video or a article, now with cloud hosting many people host outside there own geo area. I cant find the video/article but I was led to believe that the TLD is the important thing. That video is 2009.

Maybe, but I would put that one on your site also. The page will probably end up so deep in the site that the links would carry very little link juice, there will probably be so many links on the page linking back to their own site that very little would be shared with your links.

On the other hand, adding a page to your site will boost page rank just by existing  and with smart linking will increase your homepage rank for ever. See http://thatsit.com.au/seo/tutorials/a-simple-explanation-of-pagerank

I would recommend looking at ASP.Net webpages. Quite easy, good tutorials and videos, great free tools. and produces very clean code. http://www.asp.net/web-pages