It's tricky. I think most coders don't view things that way. Linus Torvalds, the god of linux, for instance runs Redhat linux. A middle of the road, definitely not 'hard core' distribution like Slackware or Debian.
Personally, this guy just sounds paranoid. That comment is about on the same level as this one I found from a while back, actually it's an excerpt from an OSNews (a real low grade magazine IMHO) article that was posted to slashdot a while back:
Sometimes they [lusers] even abuse the physical metaphor of tabbed browsing by opening multiple pages - not subpages of the same web site! - in multiple tabs of a browser window. I even know few people who never open more than one browser window, viewing all pages in tabs; I hope they do not try to glue a daily set of newspapers together before reading them...
http://www.osnews.com/story.php?news_id=7344
What's really bizarre about the OSNews excerpt is that lots of people find the physical metaphor of the desktop artificially constraining. You can't have symbolic links in a desktop, or mount another desk in an arbitrary drawer in your main desk.
Abuse the physical metaphor. What a dork. Reminds me of an entry from the Waste Books about nomenclature.
This brings up a more general epistemological problem -- how do I know for a fact that, say, grep is actually returning the results the syntax says it should deliver? Honestly, if I had to choose between trusting some algorithm and printing out a few pages of text, then highlighting all the lines on which a certain string of text appears, then retyping those lines to make a new text file, I think the choice is clear here.
That said, I agree with the general sentiment that Windows shouldn't "dumb down" so much -- but a lot of times, when those errors take place, there's really nothing that would be changed by more detailed knowledge of what's going on. Windows doesn't offer the details because it doesn't offer any tools (other than a reboot) for solving lock-up problems or other random errors.
You could read the code for the regex library, and the relevant part of grep's source. Maybe you wouldn't even need to read the regex library's code; I bet you could do that part inductively.
Besides, "grep" is fun to say.
[Insert obligatory comment about the trustworthiness of compilers and link to "Reflections on Trusting Trust"]
Oh yeah, that. Well, it doesn't change the fact that "grep" is fun to say.
Well, "gui" is fun to say. Especially if you pretend it's Italian.
you could boot into x and run grep, and then boot into an external drive which has access to x as a readable filesystem and grep, and then compare the two results, this would not prove the trustworthiness of grep, but it would show if grep in x had been compromised, an important step towards determining trustworthiness right?
hmm,this reminds me of a problem from today, i made a spider that crawled our site looking for xml schemas, which returned over 1600 schemas, this was problematic because a query against our UDDI registry as to the number of schemas returned slightly under 1500. So which was right? luckily I know enough to answer I was, I spent at least two days on that code, there's no way a 6 million dollar uddi registry is gonna outperform my quick hack.