olive-list Digest, Vol 2, Issue 1

Benjamin Gilbert bgilbert at cs.cmu.edu
Wed Aug 28 17:05:24 EDT 2013

On 08/28/2013 04:08 PM, Peter Chan wrote:
> As you have already known, I am not a programmer and
> have limited technical knowledge. It took me 30 minutes to install
> Xenserver, XenClient and to create a virtual machine under it. Now, I
> have WordPerfect run under MS DOS 6.22; a physical (Dell Desktop used by
> Robert Creeley - one of our collection) to virtual machine (Windows XP
> SP3) and a Windows 7 instance of my office machine in my Xenserver. It
> seems to me that it will be much more difficult to implement the Olive
> programs - install VMNetX client in Debian/Fedora/Red Hat/Ubuntu and a
> webserver. It seems the Xenserver approach can reach more non-technical
> people like me. The more people contribute to preserving software, the
> better. What do you think?

Hi Peter,

At the end of the day, most users of virtualization software are not 
using it for preservation, and so it is not developed with preservation 
in mind.  Right now, DOS and early Windows are fairly well supported by 
most virtualization systems, but this may not be true in 20 years as 
those operating systems fall even further into disuse.  Even now, 
virtualization systems do not provide high-quality support for 
relatively obscure x86 operating systems such as OS/2, BeOS, SCO Unix, 
and ancient versions of Linux.  It will never be cost-effective for 
commercial systems to support these old platforms, so the only hope for 
long-term preservation is to create a highly-technical development team 
which can implement and maintain this support within an existing, 
open-source virtualization system.  This is the approach we have had in 
mind as we have developed Olive.

Of course you're correct that preservation of software shouldn't require 
detailed technical knowledge.  If that development team does its job 
well, the software it produces will be usable and useful for archivists 
of all kinds.  However, a lot of work will be required before this is 
possible.  In the short- to medium-term, I think it will be very 
difficult to do a good job of software preservation without a lot of 
specialized expertise.

So, to answer your question, I have no doubt that creating DOS and 
Windows VMs under the Xen product line is, today, much easier than 
creating them under VMNetX.  I have *serious* doubts about the archival 
value of those VMs in the long run.  Our current work is focused on 
developing a strong foundation for a system for long-term software 
preservation.  Ease of use will improve with time.

--Benjamin Gilbert

More information about the olive-list mailing list