Dear Madam or Sir!
MyDefrag hath greatly improved over JKDefrag.
Especially the blockmap update doesn't block at big files or the directory moving phase under FAT like before.
The Scripts are a great improvement, especially the hint to schedule monthly, weekly and daily Scripts in a sequence in this order with a minute difference at the same time to get the largest period version run at that time.
Thus MyDefrag can be run even on a server.
The log shows that MyDefrag reduces fragmentation to 2 or 3 pieces, my disks are half full, one NTFS, one FAT, fuller disks may result in more pieces, but not a bazillion.
Watch out for Exchange's files on a Windows server, they might be locked permanently, as was a habit for quite a number of programs including Adobe Reader. O&O Defrag regarded much more files locked than anyone else, you move even quite a number files Windows regards as locked like the Registry.
Here is my list of immobile files on the NTFS system disk:
«
C:\$LogFile
C:\$MFT::$BITMAP
C:\WINDOWS\Fonts\COUR.TT_
C:\$Extend\$UsnJrnl:$J:$DATA
C:\$MFTMirr
C:\$Bitmap
C:\pagefile.sys
»
The registry hives don't show up here. Sysinternals pagedefrg will have little to do. ;-)
Well, you'll have the Wintel Desktops and an occassional Wintel intranet server, but SAAS and WEB 2.0 including the Smartphones, most notably the iPhone and Google's Android, will gnaw at the Wintel Desktop/Server market share, note that Microsoft in the past put their web site onto BSD clones, quite a message, the own OS being plagued by malware due to its design which maketh the idea of orthogonality a nightmare, otherwise sacred in the sciences.
So please advertise MyDefrag only for Wintel Desktops and Intranet Server, Extranet Servers and other *NIX and other OS Servers like iOs and zOS etc. have to resort to other fragmentation policies like the deliberate controlled fragmentation initiated in the Berkeley Fast File System, the old 7thED *NIX File System shares its problems somewhat with FAT, note quite a number of BFFS' tricks went via OS/2's HPFS at least into early NTFS like the strage of tiny files in the MFT, that is transferred more closely into the direct inode storage of Linux' ext(x) series as the deliberate fragmentation of big files across multiple sector groups, at least all Reiser File Systems store tiny files in their B(+*)-trees' leaf nodes, others probably, too.
Thus a look into the BFFS paper, last revised in 1984, look for «ffs.ps», can be converted into PDF via GhostView/GhostScript, is a quite beneficial insight into the Computer Stone Age, those old problems have just been scaled a bit and newish features like interleaved RAM access and ZBR disks have been implemented at least as back as the late AEG-Telefunken TR440, look at
www.bitsavers.org there, you need fluent German to decipher this. Especially notable is the article in «System_Overview_Mar70.pdf» p. 45 in PDF pagination (document pagination starts at p. 101 at PDF p. 5) which is quite an early document on privacy breach by the interweaving of IT and Telecom long before the German Census Boycott in the Orwell Year 1984, with the ususal redactional delay, at the time of the first ARPANET, «IBM and the Holocaust» came quite a time after the boycott year, somehow the Hollerith machine trick, quite a crude variant of SQL and even dBASE's «locate/skip», must have seeped into the boycotters' minds.
ZFS, btrfs, and to some extent, Reiser4 with its «wandering Journal» rely on a Copy-on-Write strategy which would create massive fragmentation on older file systems since the newly written block will written to a new place and the metadate pointed to there. Old blocks and metadata will allocated to the current snapshot. Only the above mentioned BFFS strategies may cope with the problem, albeit under ZFS and btrfs, a volume may span more than one disk which maketh the task more difficult. Some help may come from ZFS' deduplication. This should take effect only with soft links that cross drive baoundaries, hard links cannot span a disk, so the stuff pointed to is deduplicated since Day One. Don't expect such modern stuff to emerge from Redmond.
So Computer Stone Age knowledge is far from obsolete unless you prefer yourself to reduce to a «WIMP mouse pusher», I am an old fossil of 1976 (see TR440) and made my money in servicing mission-critical embedded systems, still a realm of assembly low-level hacking, the Malware Guys just abusing that.
BTW, Windows is far from fulfilling professional needs.
Aye, from a short-sighted trade union view, this creates low-level minimum wage jobs at the level of a mechanical typist.
I think people should know what's going on under the hood, all that GUI-masquerading does only help Mr. Bad Guy from Upper Beluchistan. Who knows whether he is paid by the Computer security companies?
Kind regards
Norbert Grün (
gnor.gpl@googlemail.com)
