PDA

View Full Version : Registry Integrity - II



xxmikexx
01-24-2008, 03:43 AM
This thread is the continuation of Registry Integrity - I, located here http://www.flightsim.com/vbfs/showthread.php?t=160311




Posted by angels355

Regarding, leave the registry alone, I'd say the notable exceptions would be the actions taken by Ad-Aware or Spybot Search & Destroy, I trust them to remove registry entries that are known threats of spyware or adware.



Posted by xxmikexx

Yo Loki,

I agree that Mark Russinovich is an expert. That's why he wrote "Registry Cleaner", or whatever his famous early registry health program is called. (This was before Microsoft bought him out.) Russinovich also wrote the famous "PagDefrag" which, in the general case, does not fully defrag the paging file, see the edit at the back of this memo.

I'm not saying that registry pollution is the source of all problems, I'm saying that, depending on how given applications use the registry, registry fragmentation etc can noticeably slow a system down, and that safe (repeat safe) registry management tools are a good idea. If you think differently, that's fine. You're very knowledgable, and when you speak, people ought to listen. I will defer every time to your knowledge about PC hardware, but not when it comes to the internal workings of system software.

I'm making known my position, and the reason for my position, for the benefit of those who might be interested in thinking situations through rather than simply taking somebody else's word for <whatever>. I'm also the person who recommended the crazy-sounding idea of alphanumeric sorting of disk drives where scenery happens to be resident as a way of helping to overcome FSX scenery microstuttering. You can read about that, and the reasoning behind it, and the experimental results, in a pair of articles of mine starting here ...

http://www.flightsim.com/cgi/kds?$=main/op-ed/ed395.htm

I guess my point is that analytical thinking wins every time and that, for analytical thinkers, appeals to higher authority mean zero. As I said, "Still, it moves".

Edit: Based on experiences subsequent to writing those articles I withdraw my recommendation regarding Ultimate Defrag. Today I use two products, not one -- PerfectDisk for offline defrag of $MFT and the swap file, etc., and O&O Defrag for the online Complete/Name alphanumeric folder/file sort. Again, people are free to believe what they want, I'm simply making the information available.




Posted by sandgate

loki
01-29-2008, 01:35 AM
I think you missed my reply to your post you copied above.


Mike,

Regarding your comment about Mark creating a registry cleaner, according to a post he made on his blog, he hasn't, and will never implement a registry cleaner. He does recognize that there are uses for them, but those are pretty specific.

Quote:
I haven't and never will implement a Registry cleaner since it's of little practical use on anything other than Win2K terminal servers and developing one that's both safe and effective requires a huge amount of application-specific knowledge.

10/7/2005 9:41:00 AM by Mark Russinovich
At this point I will just agree to disagree on this topic.

loki

I do agree that analytical thinking is the best way forward, and so far I haven't found anything that convinces me registry cleaning is worth the risk. And combined with my own experience, I advise people that ask me to not bother. Outside of fixing very specific problems, or someone wanting to really learn about it, the vast majority of people should just leave the registry alone.

If you can show me a study that shows a registry cleaner having a positive effect, please do. So far all I have found are "studies" from developers of registry cleaners and comments from people that use them regularly but no hard proof one way or the other that they actually benefitted. I have no problem with your defrag recommendations because it is pretty easy to verify that it helps FSX load faster. Just running a registry cleaner as part of system maintenance and having nothing go wrong so far doesn't convince me one bit.

xxmikexx
01-29-2008, 04:13 AM
That's fine, Loki. I'm not trying to "sell" anything. I'm simply telling people what I do and why I do it.

The purpose of my breaking out this material as a dedicated thread was simply to make things clear to the readership by consolidating into consecutive "posts" the material that had become scattered in the original thread.

It is still on my to-do list to continue the process of breaking up and then regrouping the rest of that thread.

xxxxxxxxxxxxxxxxxxxxxxxxxxx

Now ... Appeals to higher authority mean zero. If Russinovich knew what I know about the behavior of certain commonly encountered license protections schemes, he might well give a different professional opinion.

Rather than my giving a long pedantic lecture about the registry (because I've done enough writing for tonight, see Outer Marker), how about you inform the readership what the purpose of the registry is, how it is mechanised, where it "lives", how applications use it, and so on. And relate this to, for example, certain applications that continually check the integrity of their licenses several times per second.

(By the way, this from FS2crew News for 26 April, 2007 ... "This fix reduces the number of FSUIPC and license check calls ...")

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

P.S. I’m not claiming that I know for sure how the registry works. I do claim to know how I, an experienced system programmer, would have gone about it. (And the folks at Microsoft aren’t dummies so they’ll probably have approached the problem in much the same way.)

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

vdmeer2
01-29-2008, 07:34 PM
" .........and O&O Defrag for the online Complete/Name alphanumeric folder/file sort ...."

I have O&O, but where is the "on line" and how do you do "Name/alphanumeric folder/file sort"? All I seem to have is "Complete/Name".

Thanks for your help

Theo

xxmikexx
01-30-2008, 02:34 AM
Hi Theo,

Complete/Name is in fact the alphanumeric sort. My current thinking is that you should precede it with Complete/Space in order to give the Complete/Name algorithm as much contiguous space as possible in which to work.