Partition Resizer v. 1.1.2 ========================== (c) Zeleps 1994-95 ================== Program's Manual and Technical Information This file contains important information about the working and usage of Partition Resizer. Please read it carefully before using Partition Resizer. You MUST read README.1ST before running Partition Resizer. DISCLAIMER: This program is freeware. You may freely copy it and distribute it,as long as it remains unchanged and it is distributed with its original documents files. The author is not responsible for any damage caused by the use of this program. Using this program is completely at your own responsibility. 1- What is Partition Resizer 1.1 Why you need Partition Resizer 1.2 How does it work? 1.3 Is it safe? 1.4 Changes made to the previous version 2- Working with Partition Resizer 2.1 Safety precautions 2.2 Running the program: What does what 3- Technical matters 3.1 The partition system 3.2 The DOS partition structure 3.3 Is it all that easy? 4- Partitioning Techniques 4.1 Deciding what you want 4.2 Collecting many small partitions into a larger one 4.3 Splitting a partition in two 4.4 Playing with partitions' sizes and positions 5- Credits and addresses -------||------- 1- What is Partition Resizer 1.1 Why you need Partition Resizer For the last few years, more and more people search for more power in the new Operating Systems, like OS/2, Windows NT, or Linux, FreeBSD and other Unices. But trying to escape from the black and white world of DOS isn't always that easy, since most of the existing (and interesting) software is still running under DOS. So, most of these users try to keep a DOS partition along with the new OS's partition. This is not always easy, since repartitioning means loss of data, leading to the need of backup streamers, or worse (floppy backups). So, many users prefer to loose some of their data, while backing up the most valuable of it, in order to repartition their hard drive. I was in the same situation myself, so I thought: "Why not make a resizing program? This would save me time and money". So I made this little program, which does exactly that: It can change a FAT partition's size, and it can move this partition into the empty disk space, so that you can make space to create new partitions, or when you get many small partitions, you can grow them into one larger. Well, all you have to do is to read carefully all the instructions, and then use the program. In section 3, I have included many technical information that will help you not only understand how this program works, but to get some basic knowledge on the partitioning and FAT file system. But remember: This program does not intend to replace FDISK. It's rather complementary to it. You will need FDISK as soon as you make some empty disk space, in order to create the new partition. Partition Resizer does not remove partitions, and it does not create new ones. 1.2 How does it work? Well, this is the tricky part. The details are described in section 3, but I'll try to make a start from here. I assume that everybody understands the importance of partitioning. I also assume that most of you understand how partitioning is achieved. For those who don't, here it is: Partitioning is just a marking of territories on the physical disk's surface. At the first sector of the disk (sector 1, head 0, cylinder 0) resides an executable code block, which looks for the bootable partition, and runs it's boot sector code. That sector contains the information needed to divide the disk space into partitions. Later on, the boot sector (which is created by the OS's format utility) loads the OS Kernel (in DOS's case it loads the IO.SYS and the MSDOS.SYS), which continues the job by loading the information of all the partition structure into memory. Partition Resizer does two things: When moving, apart from moving the data, it changes the information contained in the partition structure. When resizing, it changes the file system's data, which is contained in the boot sector and the FAT area. Both partition and FAT FS structures are explained at section 3. 1.3 Is it safe? The previous version (1.1.1) of the program has worked flawlessly in all normal configurations. It has crashed once though, in the case of Geoffrey Coram (gjcor@mit.edu) who happened to have a configuration which had never been tested before: three primary DOS partitions on the same disk. This version works fine with this but you understand, even if the program has been tested with hundreds of partitions and partition combinations, and even if the program worked great till now, there might still be some incompatibilities or even bugs. In case you find a problem, please inform me ASAP. In order to avoid trouble, you should backup the most valuable data in your disk, those that cannot be replaced. If something happens, you'll only have to reinstall what's lost. Most of the partition combinations this program has been tested with are combinations created with DOS's FDISK, or compliant to it. If you have created a partition structure which is not compatible with DOS, this program might cause damage to your disk (you can never tell...). This typical partition structure, is described later on in this document (section 3). But reliability isn't everything. Stability is also a great factor. And this program is so stable that it can restore its work even after a sudden power shutdown! Normally, if the program was interrupted during the resizing or moving process, it would be impossible to access the partition's data again. This program stores every single step, so it always knows what the last action was. For speed and safety reasons I use the CMOS's bytes 1,3 and 5 (the alarm bytes) to store the step counter number. Because the contents of the cmos are preserved even when the power is off, the program can always find the last step and continue its work like nothing happened. If you don't believe it, test it. There is one thing that I couldn't prevent: Bad sectors. In case your disk has bad sectors, the program will not run, in order to protect your data. It will make a surface scan before it starts, but if you are certain that your disk is free of defects, you may skip it. It would be better if you had your disk surface scanned with a commercial program (like Norton Disk Doctor or PC Tools DiskFix) before you run Partition Resizer. And note this: The program can be *really* dangerous, if you don't follow ALL precautions described in the README.1ST file. Please read it carefully before using the program! 1.4 Changes made to the previous version -Fixed a major bug with more than one primary DOS partitions -Changed disk recognition routine -Documentation update 2- Working with Partition Resizer 2.1 Safety Precautions As with every program that messes with your data, you should be extra careful. It is not difficult to make a mistake that will cost you valuable data. So, you MUST follow carefully the instructions contained in the file README.1ST, in order to ensure data safety. It would be a good idea to print TROUBLE.DOC file in order to have it available when something wrong happens. 2.2 Running the program: What does what The program will first look for the disk characteristics, it will scan the partition information, and it will identify and check all the DOS partitions of the disk. After some basic checks, it will continue with the main menu of the program. Here you have 5 choices. The first is resizing a DOS partition or an extended partition. This option will resize a partition in order to make free space for a new partition to be created. It will also grow a partition if there is unallocated space available and the cluster size is big enough. You can also resize extended partitions. When you resize the partition, you may use the second option, move partition, to move the partition in the empty space. This way you will be able to change the order of the partitions in the disk. The third option is used to change a partition's cluster size. This will help you to make a small new partition able to grow up to any size you want. This option should be used ONLY on empty and formatted partitions, since it deletes all the data inside the partition. The fourth option will show you detailed information for every partition on the disk. This may sometimes help you find out which partition you want to move or resize. The last option will exit the program, and if you made any changes to the partition structure it will immediately reboot the machine, so that the new information will be loaded from the disk. Every choice but 5, will lead you to a menu where you will be asked to choose the partition which should be modified. Just enter the partition's number and press enter. Every partition description can have up to four flags. The first flag can be either P,L or E. These letters stand for Primary, Logical and Extended partition respectively. The flag * shows that the specified partition can be resized. Only partitions with * can be resized. The flag # shows which partitions can be moved. The flag ! shows that a partition is incompatible with the program, therefore it cannot be moved or resized. All partitions can be moved, but this is not always safe. I have tried moving FAT and HPFS partitions successfully, but I cannot be sure for every partition type. If you decide to move any other partition type but FAT, you will be proceeding at your own risk. After choosing the partition you want, you will be asked to change the new partition's size or move it. This can be done on the Cylinders map, with the +,-,*,/ keys. You can always press ESC to quit to main menu. If you decide that the new setting satisfy you, you may then press enter. You will be asked to confirm your action, so you should press 'y' here. Any other key will get you back to main menu. The screen will now clear, and you will be asked whether you want to perform a surface test or not. This is not necessary if you know that your disk is free of defects or if you performed a surface test recently, since it will take some time (from several seconds to 15 minutes, depending on the size of the partition). If you are uncertain, perform the surface test, since bad sectors can be dangerous to your data. If the program finds any bad sectors, it will quit immediately. If the surface test concludes normally, the program will start performing the requested changes. This may take from a few seconds to a few hours, so be prepared. If the program is terminated abnormally, p.e. in case of a power shutdown, don't worry, you can always reboot and run the program again. It will continue exactly from the point it stopped, so there is no worry. After it finishes, you may proceed with more changes, or exit the program, which will boot your machine. Sounds easy? Well, it should be. But if you are confused, you better check out what's on the next section, just to get an idea about the whole partitioning system. If you still have questions, you can get help from friends, or contact me via my Internet addresses (see the end of this document). If you don't have access to Internet, you can send me a letter (postcard preferred) and I will try to help you as much as I can. Don't forget to tell me details about your system, your partitions, and send me a copy of the PRESIZER.LOG file. 3- Technical Matters 3.1 The partition system The partition system is one of the most important things on the disk subsystem. It is a standard beyond File Systems and Operating Systems. The partition structure looks very much like the DOS directory structure. The root directory here is the root sector, the first sector of the disk. When the machine boots, bios loads this sector at address 7C00:0000, and jumps there. As you understand, there is code written at the beginning of that sector. This code searches the primary partition list to see which one is bootable, then loads its first sector (the boot sector) at the same address (7C00:0000) and it jumps there. After that, it is the operating system that takes control. Let's see some more details. Here is a detailed map of the root partition sector (sector 1, head 0, cylinder 0) (you may often see it as MBR, master boot record): Offset Description 0x000-0x0DA Boot code 0x1BE-0x1CD First Primary partition descriptor 0x1CE-0x1DD Second Primary partition descriptor 0x1DE-0x1ED Third Primary partition descriptor 0x1EE-0x1FD Fourth Primary partition descriptor 0x1FE-0x1FF System Marking word (0xAA55) Boot code is the executable code that was described above. The partition descriptors have the following form: Offset Description 0x0 Boot marking 0x1 Side \ 0x2 Sector & Cylinder 8-9 > Starting location 0x3 Cylinder 0-7 / 0x4 System Description 0x5 Side \ 0x6 Sector & Cylinder 8-9 > Ending location 0x7 Cylinder 0-7 / 0x8-0xB Relative sector 0xC-0xF Number of sectors in partition Boot marking is a byte value, which can be either 0 or the drive number (0x80). If it is 0x80 then the partition is the active partition of the disk (boot partition), and that's where the system boots from. The next 3 bytes contain the side, cylinder and sector where the partition begins. Sector and cylinder are stored in two bytes. Bits 0-7 of the cylinder value are stored in the second byte, while bits 8-9 are stored at the high bits of the first byte. The sector value is stored at bits 0-5 of the first byte. So, the word is bitmapped like this: FEDCBA98 76543210 CCCCCCCC CCSSSSSS 76543210 98543210 where C are Cylinder bits (the number below shows which cylinder bit is which) and S are Sector bits. Right after is the system byte, which is the identification byte for the partition. A value of 0 Means that the partition is not being used, while other values depend on the file system. DOS uses values 1,4 and 6 for FAT12, FAT16 and BigDOS partitions respectively. A value of 5 means Extended partition, which is explained later on. Next, there is the ending location (sector) of the partition, and later on, the relative sector value. This is a number that shows the position of the partition relative to the present sector. So, for primary partitions, it's the starting sector of the partition (Primary partitions are the partitions that are described in the root of the partition structure. As we will see later on, there are partition descriptors in other places on the disk as well, which are called logical drives). The last item of the descriptor, is the partition length (in sectors). In the root sector, there is space for 4 descriptors. Right after them, there is a word value of 0xAA55, which marks the sector as system sector. This value exists on every sector of the partition structure (including boot sectors) and if it doesn't exist, then the structure may probably be damaged. Now, concerning extended partitions, they are described as any other partition in the root sector, but their treatment is different. These partitions are specifically treated by the Operating Systems. They point to a sector which contains partition descriptors for other partitions and extended partitions, and so on. This is how we can have more than 4 partitions on a disk. Let's see a small graphic example: MBR 1st Ext 2nd Ext 3rd Ext ----------------------------------------------------- 1:p | 1:l | 1:l | 1:l | 2:p | 2:e-----------| 2:e-----------| 2:- | 3:e----| 3:- | 3:- | 3:- | 4:p | 4:- | 4:- | 4:- | ----------------------------------------------------- In this example, we see a map of a partition structure. MBR stands for Master Boot Record (root sector) and 1st, 2nd and 3rd Ext stand for 1st, 2nd and 3rd Extended partition nodes. 'p' stands for primary partition, 'e' stands for extended partition and 'l' stands for logical drive. The 'e' partition in the MBR, is called main or root or primary extended partition. This is because the whole extended partition structure is depending on that partition. Now, the starting location of the primary extended partition points to the 1st extended partition. The space that the primary extended partition allocates, is reserved for logical drives. The 1st extended partition (and the rest as well) contain information about their respective logical drives, in a way identical to the MBR (described above), but without the boot code part. Every extended partition has 4 descriptors as well, which may contain logical drives, or extended partitions as well. So, the partition structure is a linked list structure, which can be as long as we like. (Every extended partition description sector has the 0xAA55 mark at its end, to identify it as a system sector). There are of course some specifications for partitions. Partitions must start at the beginning of a cylinder, and end at a cylinder's end. This is that starting location must always be Cyl: XXXX Head: (0 or 1) Sector 1 and the ending location must be Cyl: YYYY Head: k-1 Sector l, where k is the number of heads of the physical disk, and l the number of sectors per track of the physical disk. Also, if you delete one of the middle logical drives, (p.e. the 2nd Ext.) then the previous extended partition will be connected directly to the next, replacing the link. If you want to see the whole thing yourself, you can make your primary DOS partition smaller, and create some new logical drives just to play with them. You can get much help from a commercial Disk Editor, and you can of course play with Partition Resizer. 3.2 The DOS partition structure The MBR executable code will load the boot partition's boot record. This is the second step of the booting procedure. And this is where the whole thing differs from OS to OS. We will only look at DOS's proceedings for now, since the rest OS's are less documented and more complicated (although on the next version of Partition Resizer I'll manage to resize OS/2 partitions, and if I get lucky enough, why not NTFS...). The DOS partition is rather simple: The first sector is the boot sector, right after there are two copies of the FAT (File Allocation Table), right then we have 32 sectors for the root directory, and finally we have clusters 0 and 1 (unused) and later on the rest of the clusters. The Boot record, is an executable code block, which is loaded by the MBR, and it makes some preparation in order to load the IO.SYS file, which is the DOS kernel (if we can call it a kernel... ;-)). This is useless to us, since only the processor can understand machine code, but the boot sector also contains useful information. This information begins at offset 3 of the sector and contains the following: Offset Description 0x03-0x0A OEM ID: This is a small string written by the formatter. 0x0B-0x0C Bytes per sector: This is always 512 (it depends on the physical disks characteristics). If you have a disk with a different value, please don't hesitate to inform me. Partition Resizer WON'T WORK FOR DISKS WITH SECTOR SIZE OTHER THAN 512! 0x0D Sectors per Cluster: The cluster is the basic block of information on a FAT drive. A FAT drive cannot have more than 65527 clusters. This size can limit the maximum size a partition can reach after resizing. 0x0E-0x0F Reserved sectors at beginning: This is normally 1, the boot sector. After these sectors, begins the FAT area. 0x10 FAT copies: How many FAT copies are there. Normally there are two copies of FAT. 0x11-0x12 Root directory entries: This is normally 512. This shows the maximum number of files and directories that root directory can hold. This is because the root directory has a constant length (512 entries * 32 bytes/entry / 512 bytes/sector = 32 sectors) 0x13-0x14 Total sectors on disk (small): If we have less than 65536 sectors in the partition, this value contains the number. If it's more, then the number is stored in bytes 0x20-0x23. This is an entry that was left from the old DOS versions, when partitions could have up to 65536 sectors. 0x15 Media descriptor byte: This byte is always F8 for hard disks. 0x16-0x17 Sectors per FAT: This shows how many sectors does each FAT take up. This depends on how many clusters the partition has, and what is the FAT type (12bit/16bit). This can be from 1 to 255 sectors. 0x18-0x19 Sectors per track: Same as the physical disk's sectors per track value. 0x1A-0x1B Sides: Same as the physical disk's head number. 0x1C-0x1F Special hidden sectors: This is how many sectors exist between the partition's description sector and the boot sector. Usually one track. 0x20-0x23 Big total number of sectors: If we have more than 65536 sectors in the partition, their number is written here. 0x24-0x25 Physical drive number: This is the physical drive number (c:0x80, d:0x81 etc.). 0x26 Extended boot record signature: This marks an extended boot record. If it is 29, the disk was formatted by DOS 4.0 or later. 0x27-0x2A Volume serial number: This is the partition's serial number. 0x2B-0x35 Volume label: This is the partition's label string. 0x36-0x3D FS ID: This is a string that identifies a partition as 12bit FAT or 16bit FAT. This is a detailed description of the boot record information. This will help you to understand more about the FAT file system. The boot sector is the first sector of a DOS partition. The number of reserved sectors (which is normally 1) shows how many sectors we have before the FAT area. So in most cases, the first FAT sector is sector 2 of the partition (from now on, every sector number will be taken relatively from the beginning of the partition). The first FAT, who's length is given in the boot sector, contains entries that mark the partition's space allocation. Every file has a small descriptor of 32 bytes, which resides in its directory area (not in FAT). The directory area is itself a file, which is described in its parent directory, and so on until we reach root directory, which is a constant area in the partition (we'll see that later on). A word value in the file descriptor, tells DOS which is the file's first cluster. FAT consists of word values (or 12bit values if it is a 12bit FAT system), which begin from the first sector, and continue until the last one. Every value represents a cluster, which one is depending on the value's offset from the FAT's beginning. So, the 3rd word of the FAT stands for the 3rd cluster of the partition, and so on. So when DOS knows a file's first cluster, it's looking it up to find the value that is stored in the FAT's respective position. That value points to the next cluster of the file, and so on, until an EOF marking of 0xFFFF is found (0xFFF for 12bit FAT). Empty clusters are marked with 0 and bad clusters with 0xFFF7. This is how FAT works. Now, the first two positions of the FAT are reserved, and they have an identification code which is 0xFFFFFFF8 (0xFFFFF8 for 12bit FAT). The rest FATs (2nd, 3rd etc.) are identical to the first one. They are exact copies of the first FAT, and they begin right after the end of the first FAT. Right after the last FAT, resides the root directory. This is normally 32 sectors long, and contains 32 byte entries which describe the root directory's files and directories. Right after the root directory, begins the first cluster (cluster 0) which is unused, and so is for cluster 1. Right after these two clusters, begins the user's space, where the files are stored. This is the whole story about DOS's file system. This file system was designed to be fast (it's very easy to find the unallocated space) but it has certain disadvantages: One is the very large amount of unused space, which is caused by the use of clusters. For example, if you have a 340MB disk, you have to use 8192 bytes cluster size (16 sectors). This means that, if you write a file that is 0 or several bytes long, you will consume space of 8K. This could be resolved if we could use more than 65536 clusters in a DOS partition, which could mean a 20bit fat or more, in order to have clusters of 1 sector. Since this is not possible, there is another solution: Why should I have one partition with 8K cluster size, and not 2 partitions with 4K cluster size? The loss is lowered dramatically, and it would save many megs. This is not though always possible, since changing cluster size without loosing the existing data is extremely complicated. But who knows, wait till you see the next version of Partition Resizer... 3.3 Is it all that easy? Well, actually it isn't. DOS partitions were designed to remain unchanged. So, I faced lots of problems while trying to successfully resize them. The worst was the FAT12 to FAT16 and vice versa conversion. Why is that needed? Because those #$%#$@#^!@ who designed the FAT file system, thought it would be smart to determine the FAT type from the partition's clusters number... runtime! This means that when you boot your machine, DOS calculates the number of clusters in a partition, and if it is less than 4088, it's assuming the FAT to be 12bit, or if it's greater or equal that 4088 it takes it to be 16bit! Well, I couldn't find a good reason to this technique, but it caused much trouble, since if I grow a partition over 4088 clusters, or shrink it to less than 4088, I have to change the FAT from 12 to 16 or vice versa. Another thing that troubled me was the fact that PC Tools DiskFix found errors in the partition structure while everything seemed to be OK. Later on, I discovered that PC Tools give a false alarm, if there is empty space in the extended partition before the first logical drive. This means that, if you have two logical drives in the extended partition, and you delete the first one, then PC Tools will warn you without a reason. This is not because of my program, since the same thing happens with FDISK too. Still more problems appeared, when I fixed all the above, since PC Tools consider the cluster limit to be 4080 clusters, and not 4088. So, while DOS and CHKDSK found the partition to be OK with 4084 clusters and 12bit FAT, PC Tools DiskFix considered it as 16bit, and it appeared to be messed up. So, I avoid to make a partition to have clusters number between 4080 and 4088, just to avoid similar problems with other programs. So, normally the partition should be resized correctly. But in case your diagnostics program finds FAT errors or anything, don't jump to conclusions. Try the good old CHKDSK. And if you still doubt your partition's integrity, try a third diagnostics program to get rid of all your doubts. 4- Partitioning Techniques 4.1 Deciding what you want Well, OK, you've got the program. But now what happens? Where will you create the proper space for a new partition? And which partition is the one you want to shrink or grow? How should you place the partitions in the empty space to order them as you like? How will you grow a partition which refuses to grow more? First of all, you have to decide what you want. Take a look at the partition list that FDISK or Partition Resizer provide you. Decide which partitions) you need to resize. You may recognise them by their size, or serial numbers. Then you'll have to think what the new sizes will be. Don't forget that resizing a logical drive doesn't make space for primary partitions, but only for logical drives. You will have to move the partitions inside the extended partition, and then resize the extended partition itself in order to make space for a new primary partition. The same is for logical drives. Think of the extended partition like a balloon in a box. You cannot get space in the box if you don't deflate the balloon, and you cannot inflate the balloon if you don't empty some space in the box. Remember that what's inside the extended partition cannot get out of it, and what's outside cannot get inside either. Now, here are some useful techniques to use in certain situations. Just take a look to see if they meet your needs. 4.2 Collecting many small partitions into a larger one Sometimes, you get your disk divided into three or four partitions, which is not a useful thing. Until now, the only solution was to backup all your data, delete the whole partition structure, and then create a large partition to put your data inside. Partition Resizer can resolve the problem without any backing up. It's much faster and easier. Just do the following: a. Decide which partition will be the one to remain. Normally, you should choose a primary partition. The problem is if it can grow enough. Decide what is the size the partition will finally grow to, and check if your primary partition can grow to that size. If it can't, go to step b, otherwise read step c. b. If you want to create a partition that can grow enough for your needs, you will have to empty your primary partition. To do this, you have to shrink it to its minimum, while growing a logical drive. This can be done by shrinking the primary partition, expanding downwards the extended partition, moving the first logical drive to the beginning of the extended partition, and growing it to the max. Then, exit the program, move as much files as you can from the primary partition to the grown logical drive, defrag the primary partition, and repeat step b until the primary partition is empty. When you empty it, use the program to change its cluster size that matches your needs. You can provide the program with the desired size, which is the size you want the partition to have when you finish the whole process. Partition Resizer will then suggest you a cluster size to apply to the partition, which is the optimum size for the given desired partition size. If this was your boot partition, you will have to make it a system partition again, so run SYS to do that. c. Now that your primary partition is able to grow enough, start moving data into it. Try to get data from small partitions, so that you can finish with them early and get rid of them. If you filled up the primary partition and there's still data in the partition which you are emptying, move data from the emptying partition to other partitions, in order to empty it faster. When the partition is empty, you can delete it with FDISK. If you cannot empty it, then defrag it, shrink it to its minimum, and move it to the end of the empty space created, grow the previous partition to reclaim the empty space created, exit the program, and move the rest of the data remaining into the grown partition. Repeat this until you empty the partition, and then delete it with FDISK. d. One of the small partitions is now deleted. You can now move all the other logical drives to the end of the empty space, and resize the extended partition in order to make more space to grow the primary partition. Then choose another partition and repeat step c until you empty and delete all the logical drives. With the last one, you can delete the extended partition too. And presto! You can now grow the primary partition to fill all the empty space in the disk. 4.3 Splitting a partition in two This technique is much alike the previous one. Suppose you have one primary partition, and you want to split it in two partitions. This is faster and easier than the previous procedure, just do the following: a. Defrag the primary partition, and then shrink it to the minimum. This way, you will create empty space for the new partition. b. Using FDISK create an extended partition, and then a logical drive in it. Format the logical drive, and run Partition Resizer. You will now have to decide the final size of the new partition, in order to change its cluster size while it is still empty. When you decide the final size, choose the third option of Partition Resizer's main menu, and give as desired size the size you decided. The program will suggest you a cluster size that is optimum for the partition, and you should apply that size. c. When you finish creating and modifying the new partition, exit the program, and move the data you want in the new partition. Then, shrink the primary partition again (defraging is necessary here), extend the extended partition's starting position towards the beginning of the disk, and then move the logical drive to the beginning of the extended partition. Grow the logical drive to the maximum, and repeat this step, until you reach the desired size. 4.4 Playing with partitions' sizes and positions And now, some general advice about partition resizing: a. When shrinking a partition, always defrag it first. This will leave space at the end of the partition, which will be freed by the program. The less space you have at the end of the partition, the less the shrinking will be. When defraging take special care for the unmovable files: They must be moved too, so change their attributes before defraging the disk. System files are a special case of unmovable files. You must be extra careful with these. Check the TROUBLE.DOC to see more info about the system files. b. Although Partition Resizer can move primary partitions, you will rarely need to do this. You should always leave your DOS primary partitions in their original place, unless you absolutely need to move them. This will not harm the partition, but it might leave unused space before the primary partition which will never be used by DOS's FDISK. Warning: if you move a primary DOS partition, you might not be able to boot from that partition! c. If you want to install a new operating system, you may have to delete much of your data in the DOS partition, since you cannot move that data in the new partition. So, you will only have to shrink your DOS partition once. d. Always follow the safety rules described in README.1ST. Playing with your data can be REALLY dangerous sometimes, so be very careful. 5- Credits and addresses First of all, I would like to thank everyone that believed in my project (that is ABSOLUTELY NOBODY!). I would like to thank Laertis (Billy Anagnostou) for sharing his work with me, since he was the first to disassemble MBR code and Boot record code, and gave me the basic knowledge on partitioning. I would also like to thank Raymond Skevakis for testing my program on his disk before I even finish it. Special thanks go to Dr. Hans Michael Kvasnicka (akl02@rs1.rrz.uni-koeln.de) for additional SCSI testing and for his advice on SCSI support. Special thanks also go to Jorge (George Anastassakis) for beta testing version 1.1.0 on his ruined system, and to Nodens (Takis Kalogeratos) for letting me run the program on his SCSI disk, which helped me a lot. Many thanks go to Daniel Pun (dpun@direct.ca), who gave me a first hint about the problem with the OS/2-FAT installations. Special thanks and appologies to Geoffrey "tough luck" Coram :-) (gjcoram@mit.edu) for discovering the multiple primary DOS partition bug. I also thank you, for spending the time to read this document and for using my program. I hope that the technical info included in here will seriously help many of you to create your own programs. If you find this program useful, you can always send me a postcard to the following address: John Lagonikas 17 Lycourgou Str. 16675 Glyfada ATHENS, GREECE or send me a message at the Internet address: zeleps@hol.gr Of course, if you have any question you'd like to ask me about the program or the technical info, don't hesitate to send me a message. I'll be glad to help you (if I can of course). Anyway, if you used the program and you liked it, please send me a message with your opinion and your suggestions for the program. Request: If anyone has any info on any other file system than FAT, please contact me. Thank you. ------||------