There are many sites offering comparisons of Registry cleaners - though most, it has to be said, are far from independent and their conclusions should be viewed with some degree of suspicion.
So, are they actually necessary? Personally, I'm with those that have their doubts. This is a quote from Wikipedia...
Quote:
| Most notably, critics say there is no reliable way for a third party program to know whether any particular key is invalid, redundant or neither. Poorly designed registry cleaners may not know for sure whether a key is still being used by Windows or what detrimental effects removing it may have. This has led to examples of registry cleaners causing loss of functionality and/or system instability. |
There seems to be view that the more 'problems' a program finds, the better it is. Again, that would depend on what the software writers have set as the search criteria. Some programs are much more aggressive than others, though whether they need to be is questionable. I get the feeling sometimes that the marketing of the product demands such fantastic results to supposedly raise it above others in the market.
It is true that uninstalled programs have a habit of leaving redundant entries behind which will eventually clutter the Registry, and cleaning it from time to time can only help. Whether it speeds up the operating system is debatable. It's also true that, as with a hard drive, data becomes fragmented - hence the appearance of products that will defrag the Registry. Again, in theory, that should speed things up, but probably only by a few milliseconds.
The majority of Registry defrag products run within the Windows environment which is hardly ideal when there are so many programs running and needing to access the Registry. The ideal option is to defrag before Windows loads, in the same way that chdsk does. My preferred option is to use PageDefrag which does just that - either at each boot up, or on a one-off basis.
I recently had to clean out a machine loaded with XP SP2 after they clicked on a nasty. With evilfantasy's help, it's now clean, but as it has never had even a defrag in 4 years, I thought I'd run a few tests comparing some Registry cleaners. The choice is based on what I had available so isn't in any way definitive, merely something to satisfy my curiosity.
I only compared four products, because to have an accurate comparison, I needed to restore each system from a backup image after each test, and that takes more time than I had available.
Products tested were Registry Mechanic, Ccleaner, Eusing and RegCure - each one in its latest version. The first stage was to simply run the scan options to see what the results were. The number of 'problems' found was as follows....
Registry Mechanic 1678
Ccleaner 963
Eusing 658
RegCure 858
Next stage was to run the fix option and restart. Results as follows....
Registry Mechanic
Refused to start reporting config/system error
Ccleaner (click here for tutorial)
Started up OK, most programs ran fine, but froze when IE was started.
Eusing
Started up and ran OK. Tried most programs without problem.
RegCure
Started up and ran OK. Again no problems running programs.
Trying to compare what each program had actually done would also have taken far too much of my time, but the wide variation in 'problems' found would tend to confirm a more aggressive set of algorithms in Registry Mechanic and Ccleaner, leading to subsequent system problems.
Comparing the last 2, Eusing and RegCure, they performed much the same, being less aggressive, but one is free and the other isn't! That's not to say that under different circumstances on another machine, the results would be the same. It would depend on the system configuration and what software was installed.
Hopefully, what the tests achieved lends weight to the argument that Registry cleaning software can cause more problems than it cures, so should always be used with caution.
Let me add that I have no interest, commercially or otherwise, in any of the products tested, nor should the test results be seen as favouring one product over another.
I'm simply reporting what I found.
