Everhart Collection The following are included ASNVD_64.mar and VDD64.zip - latest VD: virtual disks for VMS on alpha which use contiguous files as disks. These programs have added one handly feature: the asnvd program will do a packack to ready any disk that is being used for a VD: unit that is specified for /LBN treatment. Thus, mounting the underlying disk /foreign or /foreign/system is no longer needed. IDEZR.zip and IDEZRPC.zip - These two zip files are the same except that one was made to be unpacked from VMS, the other being ok to unpack on a PC or the like. They implement zrdriver (and zsdriver which is the same as zrdriver but allows ucb$l_maxbcnt to be 32768 instead of 8192 so you get less split transfers) and zridehost (and zsidehost which also allows larger nonsplit transfers). These utilities let ypu access SCSI devices that don't know much SCSI and/or that have blocksize other than 512 bytes. You connect a ZR unit e.g. $ sysman io connect zra0:/noada/driver=sys$system:zrdriver and then set command zridehost and get the zridehost program running in a process and give it the ZR: unit name and the name of the device where the sort-of-scsi disk lives. This can be on a SCSI<->IDE converter (like the author's) or wherever you like. The command is something like $ set command sys$sysetm:zridehost $ zridehost zra0: gkb3: where in the example gkb3: would be the GK: unit (can be a DK: unit too) to which the physical disk is connected. This can be automated, presuming zrdriver, zridehost.cld, and zridehost.exe are in sys$system: somewhere, with a command something like $ submit/param=(zra0:,gkb3:)zrddoany which will run things in batch (and incidentally load the ZR unit if need be). It presumes you have the privs to load drivers in the process. Build with @zridebld or @zsidebld and copy the .exe bits, the .cld files, and zrddoany.com if you like into sys$system. Zridehost uses io$_diagnose to read and write the disk, and requires only that the disk return its size and that read and write work. Thus it will work on most any SCSI disk. If the disk block size is not 512, the package will maintain an internal buffer and translate blocksize dynamically, so that the ZR: unit will appear as a normal 512-byte-block device regardless of what the actual underlying disk looks like. Works on VMS 7.3 or 7.3-1 alpha. Probably works on VAX also but build commands will need to be tweaked. Object files for the C programs are supplied so anyone can link the code. Glenn Everhart everhart@gce.com