[comp.virus] Restoring from backups

71435.1777@CompuServe.COM (Bob Bosen) (12/06/90)

In volume 191, Fridrik Skulason says:

>The problem with CLEAN (as well as my own, and all other disinfection
>programs) is that all the above viruses may corrupt the file they
>infect, making disinfection impossible... ...Worse still, occasionally
>.... this corruption cannot be detected by the disinfection software.
>Therefore, it is safer not to disinfect, but to replace with
>originals, when you are dealing with a Jerusalem-family virus. The
>disinfection works in 99% of cases, restoring the program almost to
>its original state, although the checksum in the header is usually
>wrong afterwards.

Although I agree with all of the above, I would like to try to
illuminate this issue from a different perspective. Two additional,
serious problems arise from choosing a "disinfection" approach instead
of restoring from backups:

1- It tends to encourage poor computing practice by advancing the
misperception that diligent and orderly backup and restoral procedures
are not necessary.

2- It damages the accountability of the software vendor. If the
software has a bug that causes problems for the user, the vendor will
be able to dodge responsibility and place the burden of proof on the
user if he can demonstrate that the software was modified by the
disinfection process. A simple checksum test will reveal modification
and place the presumption of blame on the user.

Restoration from known good backups always works. It is consistent with
good computing practice. And it does not complicate the issue of
responsibility.


Bob Bosen
Enigma Logic Inc.
2151 Salvio Street
Concord, California 94520
USA

tel: (415) 827-5707
FAX: (415) 827-2593
Internet: 71435.1777@COMPUSERVE.COM