Re: Leaving things on trains

Roland Perry wrote:

The only data that is ever seen in clear is the current record that is
held in memory. (Printouts were possible, of course.)

But what about file copies to a memory stick, or attachments to an
email? There are many ways to "export" data from a computer. I'd
actually find it impossible to use a PC where I couldn't email people an
attachment I'd found somewhere (almost always public documents). Even
the commented bit of an email is data you got from somewhere else, and
then send on expecting people to be able to see it.

Neither of those options was available on the machine. (It's almost 15
years ago that this system was specified.) There was a backup service
available via a dial-up connection. This was done nightly, with the data
transfer controlled centrally. This was also the method for updates.

As for claims about the security or otherwise of the encrypted data, the
company we bought the software from offered a financial guarantee in the
event that anyone cracked the encryption, and all I know is that for the
period prior to my company licencing the software, and during the time
we were using it, they never had to pay anything out.

That's not the point.

Au contraire, it's precisely the feature we were after when we licensed
the system.

What you are training people to do is to believe that *anything* they
do is protected, even if they only put a weak password on a
non-encrypted data file on some *other* PC.

I don't think so. Our users (who were charged for the use of their
machines) were constantly complaining that they couldn't use them to do
what *they* wanted to do, only what the company wanted them to do. We
always gave the same answer: that it was the chosen solution to keeping
their data safe. If we trained them into thinking anything it was that
the laptops they had were quite unlike any other machine they might have
access to.

Have you not seen people claiming that a data loss was OK because the
file was "password protected" when that can mean anything from almost
no protection to even less than that?

I have seen plenty of such claims. But some of them might have some
substance. Not everybody uses off-the-shelf software.

--
http://gallery120232.fotopic.net/p9633004.html
(45 028 at Derby, Aug 1982)
.

Relevant Pages

  • Re: [Lit.] Buffer overruns
    ... or are you talking about the pagein memory instead? ... but no additional memory protection ... Fiddling the storage keys for page protection could interfer ... since with virtual address space architecture fetch protection can be ...
    (sci.crypt)
  • Re: [Lit.] Buffer overruns
    ... > 360/67 had added virtual memory and features like segment sharing to ... > Fiddling the storage keys for page protection could interfer ...
    (sci.crypt)
  • Re: For the AdaOS folks
    ... > about how the structure of AmigaOS relates to microkernels ... > AmigaOS didn't have any protection and divided everything ... But the libraries weren't like those ... supports memory protection. ...
    (comp.lang.ada)
  • Re: Protect one single byte on Linux/x86 / sharing libraries / process mem. layout
    ... I did some research on the topic "memory protection" using google, ... PDE and PTE leave some space for managment information, ... The MMU generates a exception ...
    (comp.os.linux.development.system)
  • Linux Memory / Process Management on x86
    ... As you might have guessed from the topic "Linux Memory / Process ... I did some research on the topic "memory protection" using google, ... translated to a linear address by "ignoring" everything above bit 31. ... PDE and PTE leave some space for managment information, ...
    (comp.os.linux.misc)
Loading