Steve Hayes wrote:
On Sat, 9 Jan 2010 18:58:53 -0500, "Daave" wrote:

Steve Hayes wrote:
On Thu, 7 Jan 2010 23:28:46 -0500, "David H. Lipman"
wrote:

You don't. There is no need to clean the Registry. It is a myth
to sell snake oil. Very often these so-called Registry Cleaners
are malware.

So it's OK for the registry to grow and grow, with redundant and
outdated entries?

It depends upon the operating system.

For a system like Windows XP, the answer is yes. Nobody had ever
offered convincing evidence that these outdated entries slow down
performance.

Thanks.

I thoguht that in this ng, unless otherwise stated, that WERE talking
about Windows XP operating system.

Good point.

The problem is that with earlier operating systems (e.g., Windows 98),
registry cleaning actually *did* produce a certain amount of improvement
in performance. And many people who have had positive experiences with
these registry cleaning utilities had them all those years ago when they
ran those OSes, yours truly included. In fact, the utility that I
preferred was Jouni Vuorio's RegCleaner 4.3. (This is interesting
because the program recommended by Kim Kommando is written by him, too!)

But operating systems from XP onward are designed so differently, and
seemingly countless orphan entries in the registry interestingly cause
no appreciable difference in performance. This is why so many
experienced people caution against the use of these programs. There is
no noticeable benefit and there is a small chance that significant
damage may occur. Yes, although rare, there have been instances reported
in these very newsgroups where people have been unable to boot into
Windows after running these cleaners!

Adding to the confusion is the large number of scams one can find
throughout the Internet. Many of these scams are malware disguised in a
registry cleaning package. A friend of mine fell for one of these scams
about a year ago, panicked, and wound up using his credit card to make
himself $50 poorer!

For advanced people who always have an up-to-date image or clone of
their system hard drive, using the _non-scam_ registry cleaners is not
an issue because even if the rare situation of a non-bootable system
occurs, they're covered. And some people like to play around and clean
house, attempting to rid their registries of as many useless entries (or
entries *perceived* of as useless!) as possible. Some of these people
(hello, Twayne) will insist that the perceived threat of cleaning a
registry is overblown.

But this brings us back to Square One: With systems like Windows XP,
these leftover registry entries simply do not affect performance in any
appreciable way. No one has *ever* offered actual evidence to support
this claim. The closest (and it's not close at all!) I have found is
anecdotal evidence like the following:

I've never noticed a perfrmance boost on my own machines but on
occasion
I have seen it help in customer's machines. I don't look for it either
as a
rule because it's not my purpose in running such a program. Even then
you
have to be purposely looking for it though, since an A-B comparison
can't be
made.

(from a post made by the aforementioned Twayne)

Just because someone on some newsgroup says something like "on occasion
I have seen it help in customer's machines" doesn't mean this is actual
evidence! Human beings are funny creatures and imagination can be a
powerful thing. That is why I always look for actual evidence. All one
would need to do is design an experiment that *would* allow for an A-B
comparison.

And if I reinstall the software, will it just overwrite the old
entries, so that they don't interfere with the new installation?

It depends on the software.

If the uninstall and/or reinstall instructions are well-written, old
entries will be written over. Then again, many times old leftover
entries will remain. The point is that these old leftover entries just
sit there doing nothing 99.9% of the time, thereby not affecting
performance. In certain situations, there might be an issue, but it
would be an actual issue like the *inability* to install a newer version
of the same program, *not* a general performance issue. And those
specific issues can be addressed by either a specific removal tool
(Norton and McAfee have these on their sites for download) or by using
Regedit.

That being said, a program like the one written by Jouni Vuorio can have
some benefit in locating *specific* problematic entries quicker. But to
expect that just by running it for no reason will improve performance is
to be let down because it just won't happen.

You might find the following interesting:

http://www.whatthetech.com/2007/11/2...istry-cleaner/