Monday, May 21, 2007

The Great Defrag Shootout X: Diskeeper 2007 Professional and Pro Premier

This has been the most difficult review to write, because I feel like I'm swimming against the tide of public opinion, and I've had bad experiences with Diskeeper in the past. It's also the most expensive product reviewed. As time went on the review became easier as using the product became more loathsome.
It's also the first product I have ever encountered that is both paranoid and self-defeating. When you install it, it removes the built-in Windows Disk Defragmenter (WDD), and even replaces the icon in the "Start->Programs->Accessories->System Tools" menu with its own. I assume this is to ensure that it is used for disk maintenance, rather than WDD. But the sceptic in me tells me Diskeeper doesn't want you to see that it isn't much of an improvement from WDD; hence "paranoid". It's self-defeating because the help file basically admits it can't work properly when free disk space gets low. Consider this extract:

Getting More Free Space

If your volume is extremely full, there may not be sufficient free space to effectively defragment the files. In this case, here are several suggestions:
You can temporarily move some of the files off the volume, particularly large files. This temporary measure often allows Diskeeper the "working room" it needs to complete the defragmentation process. Also, this allows Diskeeper to defragment the free space on the volume, increasing the possibility that the temporarily-moved files can be moved back to the volume in a contiguous (or at least less fragmented) condition.
There are often a number of temporary files stored on your computer, and deleting these files can help you gain more free space. [snip]
You can empty the Windows Recycle Bin on the volume in question. By default, when you delete files on most Windows systems the files are not really deleted—they are instead moved into the Recycle Bin, so they still take up disk space. Emptying the Recycle Bin does delete the files, thus creating more room for Diskeeper to do its work.
While I agree that a simple program like CCleaner will reduce the clutter on your PC, the reason why I needed a program other than WDD was to sort out the tangled mess my large files were getting themselves into that WDD wasn't able to sort out. I fail to see why I should keep 15-20% of my drive unused simply because my defrag program can't deal with my large files. That's what Diskeeper expects.

Why doesn't Diskeeper completely consolidate the free space on my volume?
There are several factors that can prevent the free space on a disk volume from being defragmented:
  • A fragmented paging file.
  • A large number of directories on the volume.
  • On NTFS volumes, a portion of the free space on a volume is reserved by the operating system for the Master File Table (MFT).
It is important to know that having all of the free space in a single, contiguous piece provides very little (if any) performance benefit. Free space fragmented into hundreds of pieces will impact disk performance, but free space that’s in a few pieces should not have any effect on the performance of your disk.
So on the one hand Diskeeper doesn't make any major effort to sort out the free disk space, and on the other hand it states that if there isn't enough free disk space it can't (or won't) defrag large files. This is a self-defeating approach, and when it hits you you'll wish you never paid good money for this program. Just uninstall it and buy a decent program to begin with.
It's slow. Very slow. Not all the time, mind you, but in the case of my C: drive it took anything up to 20 minutes, and always longer than 10 minutes, just to analyse the drive; other programs reviewed took less time to defrag the drive. Removing certain crucial large compressed files fixed this, but I need those files! The image above shows the analyse screen after 11 minutes, timed using DS Clock. Not even WDD takes that long!
There seems to be a flaw in Diskeeper that prevents it from dealing correctly with large, compressed files, fragmented or not. Apart from slowing the analyse process down, once the analysis is complete it shows the wrong answer. This may explain the problems I encountered in actually getting the drive defragmented. In the picture above it is getting the size of the file wrong because it only occupies 1441MB, not 3488MB as shown, so its "Excess Allocation" error message is misleading. On another occasion it clearly showed some red fragmented files on the screen, but the list of fragmented files was completely empty.
I enabled the automatic background defragmentation process, and noticed the hard drive activity increased, especially when I wasn't busy. The user interface has a time line which shows how many fragments have been removed on a daily basis, and also the maximum and minimum number of fragments present on each drive. Presumably this is done to show how well Diskeeper is doing. Conversely it also shows how much Diskeeper isn't doing. The total of over 6000 fragments were generated through normal PC use in spite of their "revolutionary InvisiTasking™ technology" which is supposed to defragment the machine on the fly.
The program is fairly easy to use, and it has some nice interface and information features, but no way of being able to associate a given red area on the drive map to a particular file.
Diskeeper has a concept of "just enough" fragmentation of large files, where it will not use excess time and resources to further defragment a "large" file if there is no performance improvement. Even with this disabled, Diskeeper is not good at defragmenting large files. Here are 3 successive attempts at getting my hard drive tidy using the "Professional" version:
In each case Diskeeper has ignored some files, and not made any attempt at consolidating the free disk space. This is the same problem I experienced with Diskeeper version 7 and 8.
In desperation I uninstalled the $49.95 Professional version and tried the even more expensive Pro Premier ($99.95) version. It is the most expensive product in this category, and also the most underwhelming. I wanted to see if the "Intelligent File Access Acceleration Sequencing Technology" or I-FAAST™ feature would make a difference. It did (see below), but certainly not $50 worth!
The default setting in Diskeeper is to detect removable storage devices, so when I inserted my iPod Shuffle, DKSevice.exe crashed. After rebooting the PC and disabling removable drive detection, no further problems were encountered.
The "revolutionary InvisiTasking™ technology" doesn't work as well without the I-FAAST™ feature, and when it does work, it doesn't notice when you start using your machine again. I happened to double-click on an MP3 file it was busy with, and got a nasty error from WinAmp. A few minutes later the file played fine.
The picture above was taken on Sunday at 4pm, after I had removed the 2 large SQL data files, defragmented the drive without them, and then put them back. You can see how the compression process fragments the files.
Five hours later, with a bit of prodding by getting Contig to defrag all the other files on the drive, this is the result. I was unable to get this with Diskeeper 2007 Pro, only with Diskeeper 2007 Pro Premier, and only after specifically identifying the two files as being priority files for the I-FAAST system.
I guess that in terms of the Diskeeper hype-powered marketing pitch, that makes me a gullible "power user" who will happily part with $100 just to keep his disk tidy. I was suckered once, I'm not going to be suckered a second time. For $100 I could buy both Vopt 8 and PerfectDisk 8, with change to spare. Either one of them could do a better job, and faster too.
I also find the Diskeeper install process somewhat insulting. They create a "Diskeeper Corporation" program group in "Start->Programs", and then put a single icon in there. Why? No uninstall, no link to their web site, no help file, nothing. If there is only one icon, just do me a favour and put it in "Start->Programs" and save me some time, FFS!
I have left the best feature till last: the uninstall. Unlike version 8 which left residue that made WDD unworkable until I hacked it, this one does restore WDD when you uninstall Diskeeper. I recommend uninstalling it. I did. Twice. Even so, Diskeeper 2007 it left a desktop shortcut behind and a reference in the registry to AUTONTFS. Have these people no shame?

I have not decided what to review next: DiskMD didn't make it out of the starting block, since you have to buy it before you find out whther it's any good. In the meantime, I suggest you try my combination of freeware defrag utilities. Download it here. It's certainly better than Diskeeper.

Update: Monday 28th May 2007: It seems that Diskeeper is one of the only programs actually able to defrag the Master File Table (MFT). Because I converted my FAT32 partition to NTFS before I had reinstalled everything, the MFT doubled in size while I installed all the Office files and other applications. After trying PerfectDisk 8, JkDefrag, and Windows Disk Defragmenter, Diskeeper did the trick. When I uninstalled it I didn't get the errors reported above. I have contacted Raxco to find out why PD wasn't able to defrag the MFT, when it claims to be able to so. I did not receive a reply.
Update: Friday 8th June 2007: It would appear that Paragon Total Defrag 2007 can not only defragment the MFT, but also reduce its size, something no other defrag program can do.
Update: Wednesday 24th October 2008: DK 2008 has been released, and I'll be testing it for the next few weeks to see if it really lives up to its claims. I'm not optimistic but I hope to be proved wrong. See First Impressions.

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why | digg story | Benchmarks: DK2008 and DK2007

Sunday, May 20, 2007

10 things we hate about Apple

This is the article that the Editor of PC World quit his job over:-
The company formerly known as Apple Computer and now called simply Apple, Inc. is unique in many ways--including in its ability to drive even folks who admire it positively batty. It makes great products (usually), yet its secretiveness about them borders on paranoia, and its adoring fans can be incredibly irritating. Of course, its fans have to put up with some irritations, too: Simply being a member of the club still means you must endure unending jabs from the other side of the socio-political-techno aisle. But do they have to wear their suffering as a badge of honor?
Today, we -- that's us, Narasu and Alan, veteran Mac users both -- are going to get some stuff off our chests. We've enumerated ten things we hate about Apple (or its followers, or simply about the experience of using its products).

Tuesday, May 15, 2007

The Great Defrag Shootout IX: PerfectDisk Rx Suite 1.0 build 14

I was looking forward to reviewing PerfectDisk Rx Suite 1.0 by Raxco, since I have been using PerfectDisk Pro 8 for some time already, and found it to be pretty thorough, albeit stubborn and sluggish at times. I found out about the Rx Suite from the PerfectDisk blog. Going from PD8 to PDRx is like going from the sublime to the ridiculous. The less they market this product the longer Raxco will remain in business. It's that bad.
The first good point I noticed is that at $29.99 it's cheaper that PD8's $39.99. But it isn't worth the money: it's like PerfectDisk Lite with bugs, and it's a massive 25.8MB download. Almost every feature I tried didn't work properly. More on that later.
The second good point is that PDRx shows the name of the file being analysed or defragmented, which is one of the more annoying aspects of PD8. The interface is simple and easy to follow, and well organised.
The problems started when I did the stress tests: copying a 4GB file to a compressed folder, and then restoring the data to SQL Server, again in a compressed folder. I ran the "Quick Scan" of drive C: and PDRx told me there was nothing wrong, other than the free disk space was fragmented (see below). I could find no way of listing the names of the fragmented files, or how badly fragmented they were.
Even the hard drive display failed to show the massive files that are significantly fragmented: see top picture. I ran the defrag and after 10 minutes PDRx had finished, and displayed the following image. Compare it to the top image. I can't see any fundamental difference.
Now compare that to the "before" stats displayed by the built-in WinXP Defrag program:
Quite a difference. The "after" analysis was basically the same, indicating that PDRx either can't or won't defragment files that massively affect the performance of the machine.
Next I tried the feature that allows you to analyse the file usage on the hard drive. Out of curiosity I went to look at the Data folder for SQL Server, and discovered another bug: it gets the size of the compressed files wrong. The file it displays as being 380MB in size is actually 1.79GB compressed, and 4.37GB uncompressed.
I decided to test the registry cleanup, and was relieved to find the "Rollback" function worked, especially after the "cleanup" facility decided to delete references to dozens of .hlp files used by Office97 and other packages. What were they thinking?
A similar fate awaited my "cookies" - delete all or nothing. So the dozen or so sites I visit like Amazon, Yahoo Mail and BBC News where a permanent cookie is actually quite useful, all have their cookies removed and I have to answer a whole load of boneheaded questions when I return. I decided not to try the performance tuneup because the program wouldn't tell me which options are already set and which ones need to be changed.
There is no boot time defrag facility, unlike PD8, and the "Screen Saver" defrag option is extremely irritating because it carries on with the file it was busy with even after the screen saver stops. I get the feeling that the entire package has been cobbled together and rushed to market without proper testing or consideration of the user. It's one of those "we know best for you" packages that is more likely to stuff up your machine than actually fix anything significant. I was relieved to uninstall it and reclaim 20.5MB of wasted disk space. I will not be reinstalling PerfectDisk 8 Professional anytime soon either: I'm suddenly not so certain of the quality of the software. (Update: I have since relented and installed it again. I'm glad I did)
Update: Version 1.1 of PDRx will be coming out later this year.
If you want to keep your PC registry and hard drive clean, rather use CCleaner, NTREGOPT, JDisk Report, PageDefrag and JkDefrag. They are all freeware and I trust them more than anything else I have reviewed or used in the past. Even though PerfectDisk Rx Suite was bad, I'm dreading the next review: Diskeeper 2007 professional. I battled to uninstall DK8, so I hope they've fixed that issue, although I doubt it.

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why | PerfectDisk Blog

Monday, May 14, 2007

The Great Defrag Shootout VIII: JkDefrag v3.8

Every now and then you come across a program that is so blindingly brilliant in its sheer simplicity that it takes you completely by surprise. JkDefrag v3.8 is one of those programs.
The first good point is that it's free (released under the GNU General Public License), and it uses very little disk space. The entire download is a skimpy 320kb, including documentation. It doesn't have a setup program (yet), just a few simple instructions. I was so impressed with the program that I used Inno Setup to make a setup program which is listed on my "defrag" page.
The second good point is that it has an intelligent screen saver. You can tell it to wait a predefined period (i.e. 6 hours) since the last defrag before trying again. The second brilliant feature is that once it has completed the defrag, you can tell it which other screen saver to run. So in my case I get it to run the Google Pack Screensaver, which is my favourite.
The third good point is that it has two defrag modes: fast or thorough. The default is fast, and it is as fast as Vopt8, and much faster than PerfectDisk. The thorough mode goes to the trouble of removing those tiny gaps between files that waste a lot of disk space, and took less time that PD normally takes. This is what DIRMS is supposed to do, and Vopt has a good go at tackling as well.
Another good point is the approach to temporary files. By default JkDefrag leaves 1% of the first part of your hard disk free, to be used by temporary files. The idea is that this would improve the performance of the system, but I haven't been able to measure it.
There are actually 4 programs in one: the graphical JkDefrag.exe, the command line JkDefragCmd.exe, the screen saver JkDefragScreenSaver.exe and JkDefragScreenSaver.scr. The command line version works great for me, because I can include it in a batch file I use to keep my hard drive in order.
When I ran my "torture test" involving the 4GB SQL file, JkDefrag performed well. The "fast" version decided that although one of the files was highly fragmented, all the pieces followed one another in a line, separated only by free disk space, so it left the file alone. The "thorough" option sorted the files out, and intelligently put the two main data files near the end of the disk.
There is no "interactive" mode where you can click on a disk cluster to see what files it contains. The graphical screen display uses screen pixels to represent the files. Unlike most other programs, this one uses the bottom left corner of the screen as the start of the hard drive. The colour scheme is also a little different to most, but the help file explains it all with elegant simplicity.
The only "missing" aspect of this program is that there are certain files it can't defragment because they are in use by the system. They recommend using Sysinternals PageDefrag v.32 to defragment these files, and the combination of the two works incredibly well. There is a third free utility, called NTREGOPT which can be run before rebooting, and this will compact the registry, after which PageDefrag sorts out any fragmentation. A "complete" install with all the necessary utilities is available on my Defrag page.
I can thoroughly recommend this program, both from its simplicity and its flexibility. No need to uninstall this one, although I'll have to disable the screen saver while I test the next product: PerfectDisk Rx Suite by Raxco.
Update: Version numbering may be a bit confusing: 3.8 was followed by 3.9, 3.10, 3.11 and so on. So 3.8 should really be 3.08.
Update: If imitation is the sincerest form of flattery, then Jeroen Kessels should feel flattered, because both Abexo Defragmenter Pro 5.0 and SpeedItUp FREE 4.0 clearly use the JkDefrag code, without acknowledging copyright or conforming to the GPL license.

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why | digg story | Benchmark of version 3.28 | download

Friday, May 11, 2007

The Great Defrag Shootout VII: O&O Defrag V8.6 Professional

It's difficult to know what to make of O&O Defrag V8.6 Professional. On the surface it has everything a defragmentation program should have, and more. But what do you make of a program that gives up on defragmenting a file simply because there isn't enough free space available? Apart from uninstalling it, that is.
The first thing that caught my attention was that Raxco (the publishers of PerfectDisk) do a comparison between their product and the O&O product. Methinks they protest too much. They should do a comparative review with Vopt, not O&O Defrag. But I digress.
The first good point is that the interface on installation is clear, well organised and incredibly useful. You can see exactly what is happening on your hard drive, and the pie chart is an interesting touch. Click on any block on the drive map and you'll see all the files it contains. The reports are clearly laid out and contain useful information, and you can go back a few days if required. I set mine to 5 days.
The second good point is that they have a defrag screen saver. Whenever your machine isn't being used you can get it to tidy up the disk. This is a brilliant idea, and although the graphics are somewhat cheesy (Why does the disk image need to rotate in 3D anyway?) the information that is displayed in the bottom left is helpful and thorough.
The down side of the screen saver is that it cancels any other manual defrag job you may have been running at the time, even if the job parameters are the same. O&O tech support confirmed that it's a "feature", not a bug.
The third good point is there is a boot-time defrag option, to get rid of those nasty fragmented files that you can't tackle when Windows is running. PerfectDisk only allows certain file types to be defragmented at boot time; Diskeeper and O&O Defrag will defrag any file.
The down side of this is that O&O Defrag simply "bails out" on any files it doesn't feel like defragmenting, i.e. if there isn't sufficient free disk space. It doesn't only do this during the boot time defrag (click on screen shot above), but at any other time as well (see top screen shot, taken after several attempts at defragmenting the hard drive). This is my main reason for uninstalling the product: why use a defrag program that can't defragment the files? I could understand if the file was huge and the available free space was tiny, but this is not the case. The mind boggles.
What's worse is that there isn't any predictable pattern to this behaviour. When I did my normal test of copying a large file into a compressed folder and then using the data to restore my SQL Server database, I landed up with one file that Windows had fragmented into 399000 fragments! Using the O&O Defrag Manager program I tried a "Stealth" defrag. Nothing. Then a "Space" defrag. Again nothing. Only the "Complete/Name" option made a difference, and it took 8 hours to do so, even though I had run it the day before. But the "Complete/Name" option fragmented a ZIP file that wasn't previously fragmented, and now the software doesn't have a clue how to defragment it again. It got sorted out on the reboot, but 6 other files it had similarly fragmeneted remain fragmented. Time to uninstall.
This product takes its time, even when you change from the default and tell it to use all available resources. I get the feeling that my normal disk activity is creating fragments faster that O&O Defrag can keep up. I'm not paying $49.95 for this package, even though it is well written and well put together. If it doesn't do its primary job, it just isn't worth it.
My next review package is a freeware product I found mentioned on Digg.com: JkDefrag 3.8.
Update: Version 10 has been released. I will probably review it later.

The Great Defrag Shootout: Part I | II | III | IV | V | VI | VII | VIII | IX | X | XI | XII | XIII | XIV | XV | XVI | XVII | XVIII | XIX | XX | XXI | XXII | XXIII | XXIV | XXV | XXVI | XXVII | XXVIII | XXIX| winner | all | why

Tuesday, May 08, 2007

Transport Minister Confirms eNaTIS Incompetence

Transport Minister Jeff Radebe should be commended for his "unreserved" apology for the inconvenience caused by the changeover from the NaTIS to eNaTIS system, but he hasn't explained the incompetence and bad planning that has led to a 6 week crisis at the licensing departments.
  • Why was the changeover done during the busiest month of the year, i.e shortly after year end in the month with the lowest number of working days, and therefore the longest queues?
  • Why was no provision made for offices to stay open longer and/or open on Saturdays? The offices were turning people away long before the office hours closed.
  • Why was it necessary for the old system to be switched off for an entire week, causing inconvenince and generating an immediate backlog?
  • What kind of bad program and database design has forced FOUR server upgrades in 4 weeks?
  • Why was it not possible to revert to the old system when the wheels fell off?
  • What procedures were in place to deal with the paperwork manually while the system was off-line? There appears to have been none.
  • Why has the Dept of Transport been lying to the public about the true status of the system for the last 6 weeks?
  • When will the Minister explain why the system is almost R100 million over budget (R408 million instead of R311 million) and three years late?
  • Why did the license renewal letters sent out in March not mention the planned shutdown of the entire system for the period 5th April to 13th April?
  • The fourth upgrade was announced on Friday, but only completed on Tuesday morning. Does it really take from 6pm on Friday to 8am on Tuesday to set up another server? That's 86 hours, or didn't they work on the weekend?
According to the SAPA report, "Motorists issued with traffic fines because the country's new electronic national traffic information system (eNaTIS) was offline may [possibly] have them cancelled, Transport Minister Jeff Radebe said on Tuesday.
Radebe apologised “unreservedly” for the inconvenience the implementation of eNaTIS had caused.
“Motorists already issued with traffic fines due to no fault of their own will have their fines reviewed on merit and considered for cancellation by the relevant traffic authorities,” Radebe said at a press conference on Pretoria.
Over the past few days a new server with bigger capacity had been installed. This and other work had led to the system 'operating in a stable manner' since it was brought back online on Tuesday. [We believe you, but how long are the queues?]
Radebe said more work would be done and the system should be running with stability on almost full capacity by next week.
However, he said, people who wanted to renew their licences could go to testing centres. Arrangements were [only now] being made for centres to operate on extended hours to reduce backlogs. [That's about 6 weeks too late, Mr Radebe!]
Backlogs, technical glitches and shaky, slow systems have been reported at testing stations around the country since the eNaTIS system was introduced last month."