I think it's fine to have these options as extensions or even inside about:config, where the user would never disable by accident.
If they don't read, then they wouldn't be using Firefox, since all web pages (barring a few exceptions) would be gibberish to them :)
> And showing message box every damn time one site has Javascript/SSL
That's not what I'm advocating in that particular recommendation. I'm more advocating for some sort of heuristic analysis when Javascript is disabled. There are lots of sites that do silly things like rely entirely on Javascript for rendering text (for example); those should be easy-to-detect as scenarios where a warning would appear.
Also, most similar warnings presented by Firefox already (usually about outdated plugins and such) have a way to permanently dismiss, or to remember a setting for a particular website, or some other way to mitigate the understandable annoyance of always throwing warnings. A "don't ask me again" would immediately resolve the problem you identified.
> I think it's fine to have these options as extensions or even inside about:config, where the user would never disable by accident.
I think that's fine, too. My comment was more about identifying the correct cause of various effects - i.e. that the harmfulness of the checkboxes being criticized is due to their non-obviousness rather than their existence.
Not can't read; don't read. See http://blog.codinghorror.com/teaching-users-to-read/, http://www.joelonsoftware.com/uibook/chapters/fog0000000062....
Users don't read material that's put in front of them. Modal dialogs get dismissed without reading, non-modal dialogs (Firefox's doorhangers, Chrome/IE's notification bars) get ignored completely or dismissed.
Basically, a heuristic browser-chrome view in place of what used to be a site-author's <noscript> view.
You can't get them to read prompts at all, let alone text on a page.
And from those discussions, and from my observations of those users, 99% of the problems discussed would be resolved if it was clear what options actually did. Users don't know or care what "Javascript" or "TLS" are, but you can bet your ass that if the relevant checkboxes had at least a basic explanation of why they should be checked (i.e. "Don't uncheck this box unless you know what you are doing; doing so will cause a lot of websites to break"), the vast majority of end-users will happily leave that box unchecked until they ask someone more knowledgable about it.
But in this case, we're dealing with users who somehow managed to disable JS, but are still surprised by the effects and don't read prompts.
I'm pretty sure such users exist, however instead of directly basing your UI descisions on this scenario, why not trying to investigate where such behavior comes from and how frequent it is?
Javascript disabled and a site relies on it? "Hey, this site would probably work better if you enabled Javascript, but you have it disabled. Would you like to enable Javascript again? Or perhaps just for this page?"
SSL/TLS disabled? "Hey, so TLS is disabled, but I need it in order to show you this web page. You want me to enable it for you?"
The overarching theme here is not that there are too many options, but that there are too many poorly documented options with poorly documented consequences. Fixing that problem would give users the best of both worlds: flexibility and ease-of-use.