Jerome Fine replies: Without spending a few hours checking every place that "invalid directory" is displayed from DIR.SAV, it is impossible to understand exactly what triggers the message. Also, I assume that you refer to using a hard disk drive or other device that can be file structured. However, I can say that the message is not caused when DIR.SAV is unable to complete a read operation, but when data is found that is different from what is expected. In addition: (a) I zeroed out all of block one on a file structured device - this did not cause a problem of any sort - which contains non-essential RT-11 information (b) I then zeroed out the number of segments (first word in block 6 - this did cause the "invalid directory" message - since a non RT-11 file structure likely can have invalid data as far a what is required for RT-11, it can be assumed that the RT-11 "invalid directory" message is probable with other valid file structures (c) The above result will obviously occur when a new scratch device with all zeros in at least the first 8 blocks since the first word in block 6 is zero. For example, under E11 it is possible to: MOUNT DU0: RAM:/SIZE:65536 (65536 is the number of blocks) If the user does not first INIT DU0:, then the DIR command will return the "invalid directory" message In short, if the "invalid directory" message appears, you can NOT assume that the device is a scratch pack. ALSO, Jay I have not yet figured out a suitable reply to your private question. If it takes me more than a week or two, please remind me again in case it has dropped off my radar list. I have some personal family items to finish first and they have priority. One other thing: I am attempting to assemble some x86 code for an Ersatz-11 DLL file. Since I am a dummy when it comes to even simple x86 stuff, I don't know where to find a suitable x86 assembler, let alone which one I should use. The Borland TASM.EXE file was suggested from Turbo C, but my old AT is not working. Can anyone help me find or provide a link to TASM and TLINK? I want to write code under RT-11 which will trigger code on a Pentium system that uses the x86 instruction BTR (Bit Test and Reset) so that I can clear a specified bit in a string of bits that is up to 4 GigiBits in length - so I can use that bit string as my work area. Since there are a very large number of bits to clear for the smaller primes (often more than 10,000,000) in a work area of that size, it makes a very efficient implementation if the DLL can clear all of them with a single request before returning control to the RT-11 program, probably as efficient as writing the code in any other manner. Also, just in case anyone knows, might it also be possible with the 64 bit CPUs to operate on a bit string larger than 4 GigiBits with the BTR instruction? My limited understanding of the manuals suggests that only a 32 bit address can be used - although it would obviously be possible with additional code to split the bit string into 2 portions. This question is mostly for the future since with most Pentium 4 systems, the maximum memory is only 4 GigiBytes and probably half of that will be required for other values in any case. Can anyone help? Sincerely yours, Jerome Fine -- If you attempted to send a reply and the original e-mail address has been discontinued due a high volume of junk e-mail, then the semi-permanent e-mail address can be obtained by replacing the four characters preceding the 'at' with the four digits of the current year.