A suggestion for going forward from the RMS/FSF debate

[Christopher Dimech]

Some had contacted me about it.  Could have sent response off the list.


> Sent: Monday, April 19, 2021 at 1:05 AM
> From: "Richard Kenner" <kenner@vlsi1.ultra.nyu.edu>
> To: dimech@gmx.com
> Cc: gcc@gcc.gnu.org, siddhesh@gotplt.org, ville.voutilainen@gmail.com
> Subject: Re: A suggestion for going forward from the RMS/FSF debate
>
> > It is an argument against the idea that LLVM is the default way that
> > people choose.
>
> I don't think that anybody made the argument that LLVM is the "default"
> in any sense.  What's being given here are reasons why some people
> prefer LLVM over GCC.
>
> > In those places, they don't trust Microsoft or anybody that provides
> > software products that are difficult or impossible to review.  Free
> > software is not prohibited, since the government has access to the
> > source code.  Any tool that comes compiled is not acceptable there.
>
> For a compiler, of course, you need a compiled version of it to start
> with.  If you use that same compiler to build itself, having the
> source code does *not* protect you from malware, as Ken Thompson
> showed back in 1984.  Even if you take the stance that you'll compile
> GCC with LLVM and vice versa, you still have the risk that both of the
> binaries have been compromised in this way.

There are tools that look for code that is not supposed to be there.
But people get sloppy and it's a lot of bother.  That's been my
experience.