The Computer Journal, Issue 30

Z-System Corner

© Jay Sage

Reproduced with permission of author and publisher.

After the tremendous effort I put into producing my last column and ZCPR33 before leaving on summer vacation, I was really burned out. I hardly touched the computer or even thought about programming all summer. This was quite remarkable, since for months I had practically lived in front of that computer screen and thought I would have severe withdrawal symptoms if I didn't get my usual daily fix. I was glad the last issue of TCJ came out a bit behind schedule, because three weeks ago I really couldn't think of anything to write about for this column.

Then the bug struck again, as I knew it would eventually. My mind exploded with ideas, and, as in the past, I again have far more things to talk about than I will have the energy or time to get down on paper before this article has to be sent in.

Echelon and I

I practically fainted when I saw the byline on my last column, and I wonder if both Echelon and my real employer did too! The byline read: "Jay Sage, Echelon, Inc." I guess Art Carlson misinterpreted my statement that I had joined the Echelon team. In no way am I an employee or official representative of Echelon. In real life I am still a physicist at MIT doing research on special analog(!!) devices and circuits for image processing and neural-like computing. Digital computing is still essentially a hobby.

The members of 'The Echelon Team' are independent programmers who cooperate with Echelon and each other to advance 8-bit, CP/M-compatible computing. Other members of that team, to mention only a few, are Bridger Mitchell, author of DateStamper, BackGrounder, and JetFind; Joe Wright, author of the auto-install Z-System and the BIOSs of most of the recent popular 8-bit computers (Ampro, SB-180, On!); and Al Hawley, sysop of Z-Node #2 (the only one to sign up for a node before I did) and author of the REVAS disassembler and the new ZAS.

Supporting 8-Bit Software Developers

Though team members have no personal stake in Echelon as employees or owners, they do benefit to some extent from royalty payments for their software that is sold. Nevertheless, before turning to the technical material for this issue, I would like to make an unabashed plea to all of you to support Echelon and the small number of other companies still developing 8-bit software by purchasing their products. They are the only hope for the sustained vitality of our 8-bit world. If we don't buy the products they offer, the creative programmers who have not already done so will have no choice but to abandon Z80 programming in favor of the more lucrative IBM-compatible market. Check the ads in The Computer Journal, and support the companies that advertise there. Do not make regular use of illegitimate copies of their software; buy your own.

Unfortunately, no one is getting rich on 8-bit software. I did not keep a record of the time I spent on ZCPR33, so I cannot calculate accurately the effective hourly rate, but a rough estimate indicates that babysitting would probably have been a better way to make money. In fact, my greatest financial reward probably came from the money not spent on babysitters as a result of staying home programming instead of going out!

While reflecting on these issues, I thought I had a brilliant idea - put together a transportable computer and actually hire myself out as a babysitter. That way I could get paid not only for the commercial products like ZCPR33 but for all the public-domain programs as well. Unfortunately, on more careful consideration, I noted some serious defects in this scheme. First or all, how many people would hire me to sit night after night until 2 or 4 in the morning? And who would want a babysitter who wouldn't notice the house burning down until the power went out on his computer screen?

Basically, I think it is fair to say that all of us (even those, like Joe Wright, who depend on it for their livelihood) are in the 8-bit programming business because we love it. Even for the nonprofessionals, like me, there are reasons why some financial compensation is important. While we may derive enormous enjoyment and excitement from programming, our families do not, but if it brings in a little extra spending money that the entire family can share, they are much more tolerant of the long hours spent in front of the CRT or on the telephone helping users with problems.

New Commercial Z-System Software

As a transition from my 'soap box' remarks above, I would like to begin the technical discussion with a review of some exciting developments in commercial Z-System software.

WordStar Release 4

The most exciting development in a long time is the appearance from MicroPro of WordStar Release 4, CP/M Edition. As far as I can remember, this is the first new CP/M product from a major software house since Turbo Pascal version 3 came out several years ago, and it is the only product ever from a major vendor that supports Z-System. I am thrilled at the official recognition this bestows on Z-System.

For the season's kickoff meeting of the Boston Computer Society's CP/M Computers Group, we had representatives from MicroPro to introduce the new product. As excited as I was about Release 4, I was sure that this would be the end of the line, so I was quite surprised when the representatives talked about a release 5 for CP/M as well as for MS-DOS.

MicroPro speaks of itself now as the 'New' MicroPro, and, indeed, they sounded like a new MicroPro. They are extremely solicitous of user suggestions. Their upgrade policy is very generous (they'll happily accept the serial number from any older WordStar or NewWord), and the upgrade price of $89 for individuals and $79 for clubs is very attractive. Echelon's special offer at $195 for those who did not own WordStar or NewWord before is also quite reasonable for such a high quality product. I personally have not made much use of WordStar in the past, preferring my roll-my-own, do-it-my-way PMATE text editor, but I have already placed my order for WS4, if for no other reason than to show my support to MicroPro and to encourage them to stick with us 8-bitters.

Frankly, I am also quite eager to explore WS4's new Z features. Since my copy has not arrived yet, I cannot give you a first-hand report, but what I have been hearing from others is extremely positive. It apparently knows about the command search path and named directories of ZCPR3 and can run as a true shell. The documentation included with WordStar 4 is extraordinary, the equivalent at least of the customization package that the 'old' MicroPro used to charge an extra $500 for! No longer will the hobbyists have to ferret out all the patch points for the program (though I am sure there will be plenty of areas, nevertheless, to keep us entertained).


Another exciting development is release 3 of ZAS/ZLINK, Echelon's assembler/linker (and librarian) package. As many of you may know, most serious programmers - Echelon team members included - have had little but scorn for ZAS in the past. It was a strange and unreliable assembler.

But Echelon has now really made good on ZAS. They did it right this time. They did not go back to the original author and try once again to get him to fix it; instead they brought in the highly competent Al Hawley. Though I am sure it is still not perfect (what program is), it has correctly assembled all correct code that I have fed to it. And gone are its former irritating and unique idiosyncrasies, like square brackets instead of parentheses in arithmetic expressions. ZAS will now handle just about any code written in some semblance of standard assembly language. It supports a rich set of pseudo-ops, making it tolerant of common variants.

ZAS and ZLINK are also at long last honest Z-System tools, as befits an Echelon product. They recognize named-directory references for all files, and they communicate with the Z3 environment and message buffers. With an appropriate editor it is possible to build a code development system like that in Turbo Pascal. When ZAS encounters errors in assembly, it stores enough information about the first error in the environment that an editor can automatically locate the line with the error. Thus one can make an alias that bounces between the assembler and the editor in very convenient fashion. As an added bonus, the Echelon version of ZAS includes, as the one from Mitek always did, the option to generate in-line assembly code for Turbo Pascal.

If you have an older version of ZAS/ZLINK, you should definitely order the upgrade, priced at just $20 for the software alone or $30 with an updated manual. If for some reason you do not want to do that, you should at least destroy any copies of the older versions that you have lying around.


Now I would like to move forward in time and talk about a product that is in the works (though with the delay between when I write these columns and when they reach readers, it may be on sale by the time you read this). This product is New ZCOM, or NZCOM, Joe Wright's utterly spectacular follow-on to ZCOM. This product will make manually installed Z-Systems obsolete, because manually installed systems will now offer less performance than NZCOM systems.

First some history. In the summer of 1984 Joe Wright was reflecting on the difficulty of converting stock CP/M2.2 systems to ZCPR3 and wondering if an automatic method could not be developed. Rick Conn apparently opined that such a thing would not be possible. Joe did not ask me (we did not know each other at the time), but if he had, I would have told him the same thing - impossible! You know the old marines' saying: the difficult we do immediately; the impossible takes a little longer. Well, Joe didn't even take very long to do the impossible. In a matter of weeks he had a fully working version.

ZCOM had two drawbacks compared to a manually installed Z-System. First, it required an extra 0.5K of overhead. Secondly, and ultimately more seriously, it was not flexible. One had to accept a standard configuration. There was no choice of command processor options, number of named directories, size of RCP, and so on. Thus, ZCOM was the Z-System of choice only when a manual system could not be made, either for lack of skill or lack of BIOS source code.

Since the new computer I bought at work in 1984, a WaveMate Bullet, to my chagrin did not come with source for the BIOS, I tried briefly to figure out how to get ZCPR3 installed without BIOS modifications. I came up with an approach that might have worked, but before I got very far with its development, Echelon announced Z3-DOT-COM and, shortly thereafter, ZCOM. I bought them right away. After a short time, I figured out how they worked (and was amazed at Joe's cleverness). Then I began to implement the modifications I described in my last two columns, including ways to switch between different auto-install systems from within alias scripts and while inside shells.

Later, after I became acquainted with Joe, I told him about my ideas for an enhanced ZCOM. It seemed, however, that he was much too busy at the time with other products to give any attention to ZCOM, so I decided that my next project after ZCPR33 would be a program that I tentatively called Dyna-Z. This would be a dynamic Z-System, an operating system whose configuration could be changed on the fly.

Dyna-Z would be useful in several ways. One could normally run a system with all the standard ZCPR3 modules except an IOP (input/output package), giving one a good set of ZCPR3 features for normal operations. When one wanted to make use of an IOP, like Joe Wright's superb NuKey keyboard redefiner program, a load alias would first check to see if an IOP space was available in the system. If so, NuKey would be loaded. If not, the alias would switch the operating system to one that did include an IOP and then load NuKey. One would sacrifice the 1.5K of TPA (transient program area - the memory available for a program to work in) only while one needed the benefits of NuKey.

On the other hand, when one invoked a command such as WordStar 4 or a spreadsheet that required a large TPA, an alias for that command would first switch to a Z-System with no IOP or RCP, increasing the TPA by 3.5K compared to the full Z-System. One could even drop the FCP (0.5K) and/or the NDR (0.5K typically). When the memory-hungry program was finished running, the remainder of the alias script would reload the standard version of Z-System. One would hardly know that the operating system had been different while WordStar or the spreadsheet was running.

Thus Dyna-Z would not only overcome the major disadvantage of ZCOM but would also overcome the only intrinsic disadvantage of ZCPR3 - the loss of TPA space. A minimum Z-System would use only 0.75K plus the autoinstall overhead (reduced to a mere 0.25K), a very small sacrifice for all the benefits that Z-System offered. And in a real pinch for TPA, one could even drop out of Z-System temporarily and run standard CP/M with its maximum possible TPA. Using SUBMIT, even this could be done with a script!

I was delighted this summer when Joe Wright called on the phone to discuss his plans for New ZCOM. It was the first I knew that he had taken up the project, and I found that he had already made great progress. Many of the features I had planned for Dyna-Z were already implemented, and Joe was eager to incorporate the rest. Our partnership was born! And it is a perfect partnership for me - Joe is doing all the work.

I would have been excited enough just to see the Dyna-Z features in NZCOM, but Joe has done much more than that. He has made the process of building a system of your choice about as easy as it could possibly be. You simply edit NZBAS.LIB, a text file describing the system configuration you want, and assemble all the individual components (CCP, DOS, RCP, FCP, etc.) to Microsoft-format REL files. NZCOM.COM then generates a system auto-loader program for you automatically. It will even allow you to clone an existing system, accomplishing automatically all the complex processes I described in my last two columns.


Now I would like to switch back in time and describe a program that has been around for a while already but has not had the publicity I think it deserves. It suddenly struck me the other day just how often and in how many ways I use JetFind yet how few people probably are aware of its existence.

JetFind, by Bridger Mitchell, is basically a text finding program. It is something like Irv Hoff's publicly released FIND.COM, which can search through a file for a specified text pattern. But it goes orders of magnitude beyond that.

To begin with, I should explain that JetFind operates in either of two modes: interactive or command mode. In interactive mode, the program is invoked alone, with no command tail. It will then prompt the user sequentially for each piece of information it needs. The user can then live inside JetFind, performing one search after another until he issues an exit command. Alternatively, a single search operation can be carried out by including all the necessary information on the command line.

Now for the capabilities of JetFind. First of all, JetFind is not limited to searching the text for only a single pattern at a time. It can search for multiple patterns, and each one can be either a simple text string or a regular expression (a UNIX concept). Let's take a simple case first. Suppose you want to find all lines that contain either "Smith" or "Jones". In interactive mode you would enter the patterns one at a time in response to the prompt. Just hitting carriage return would end pattern input. In command mode, you would enter for the search pattern the following expression:


The special character '|' represents 'or'. From command mode, of course, one cannot distinguish upper and lower case. To do that you must use interactive mode.

Now let's consider a more complex search that would make use of a regular expression. Suppose we want to find all labels in an assembly language program. We could use the following regular expression:


The first term in brackets means a character from the set of letters ranging from 'A' to 'Z'. The second term in brackets is the set including the digits from '0' to '9' also, i.e., an alphanumeric character. The asterisk means that the previous character specification may occur any number of times, including zero times (a '+' would require at least one occurrence). Finally the colon on the end represents the ':' character.

If labels have to be at the left margin, we could use the regular expression


A caret at the beginning of an expression indicates the beginning of a line. A mode control specification (explained later) can tell JetFind whether or not to ignore case. If other characters are allowed in labels, they could be listed inside the brackets as well.

There is not enough space here to give a complete description of JetFind's regular expression syntax. Suffice it to say that it can perform just about any search you would ever want to do.

A second major feature of JetFind is that it is not limited to searching single files; you can specify whole collections of files to search. You can give a list of ambiguous file specifications both for files to include in the search and files to exclude from the search. These files can all come from a single directory, or files from many directories can be included. For example, the file list


indicates all files in the named directory TEXT (yes, JetFind is fully ZCPR3-compatible) with a file type of D?C but not (that is the meaning of the '~' prefix) files whose name begins with 'A'. Note the '?' in the file type. The intention here is to search all DOC files. By including the '?' for the middle letter, files of type DQC (squeezed DOC files) and DZC (crunched DOC files) will be included as well. JetFind automatically uncompresses both squeezed and crunched files as it searches.

Moreover, JetFind is not limited to searching individual files. It can even search through members of libraries. If the first file name in the list of files is an LBR file, then the rest of the list is taken as a specification of the members of that library to be included in or excluded from the search. As with individual files, these files can be unsqueezed or uncrunched on the fly.

As if this were not enough, JetFind has about a dozen mode control options that define how it performs the search and what it does with the text identified by the search. Here are descriptions of just a few.

CThis option just counts and displays the number of matches without showing the matching lines of text.
NThe lines containing matching text will be numbered in the listing to make it easy to find them with an editor.
RmnThis specifies a display region ('m' and 'n' are each digits from 0 to 9). For each line containing a match to one of the patterns, the previous 'm' lines and following 'n' lines will also be included in the display to provide context.
ICase will be ignored so that 'a' and 'A' will be considered to match.
BBegin displaying the text as soon as the first match has been found.
VReverse the test and display only lines that do not match any of the specified patterns.
TType the file, extracting it from a library and/or uncompressing it as required.

With these and other options not listed above, JetFind can be made to perform many tasks besides searching, such as typing files, extracting files from libraries, splitting off parts of files, and displaying files in a directory or library.

We're not finished yet! JetFind also supports full input/output redirection. The output text that is shown on the screen can additionally be saved to a file, either in a new file or appended to an existing file. The set of patterns to search for can also come from a file. Thus we could have the command


This would search through all the ZFILER source code (ZF*.Z80) for the regular expressions contained in the file LABEL.EXP in directory ASM. The search would require whole-word matches ('W') and include line numbers with the matching lines ('N'). The output would be displayed on the screen and written to a new file called LABELS.LST in the current directory.

One final comment. JetFind does its work at incredible speed. Bridger Mitchell is an absolute master at wringing performance out of the operating system, using all kinds of tricks to speed up file operations. Hence the 'Jet' in the name. JetFind is available from Echelon or Echelon dealers for just $49.

New ZSIG Programs

Now I would like to turn to some exciting new ZSIG programs that have been released or are under development at this time.

LLDR - Library Loader

Paul Pomerleau - already well known to the community as the author of such widely used programs as VERROR (visual error handler), BALIAS (an alias editor), AFIND (alias finder), and the commercial LZED (little Z editor) - has released LLDR, a version of LDR with library support.

This program completely replaces LDR, the standard module loader program used to load new ENV, Z3T, RCP, FCP, NDR, and IOP operating system code segments. It does everything LDR did but adds one new, very important feature. It can load the modules from a library.

ENV and Z3T modules in particular are very short (one or two records), and it was very inefficient use of disk and directory space to have them sitting around as individual files. Now all the system files can be collected together in a single LBR file (or several, if you prefer). LLDR's syntax can be expressed in general form as follows:

LLDR [library[.LBR],]<list of modules to load>

If the optional library specification is omitted, then it performs just like LDR. If all the system modules are placed in a file called, for example, SYS.LBR, then one might invoke LLDR (renamed to LDR) as follows:


Besides the saving in file space and directory entries, there is another nice side benefit of using LLDR in this way - it is much faster. Since only one file (SYS.LBR) has to be opened by the operating system, there is only one directory search, and loading the entire collection of modules takes little more time than loading a single one did with the old LDR program. Thanks, Paul, for another nice program.

SALIAS - Screen Alias Editor

When I released VALIAS (visual alias editor) a couple of years ago, I wrote in the documentation that someone should please extend it to full-screen operation (it only supported insertion and deletion of complete lines). Paul Pomerleau's BALIAS allowed full WordStar-like editing of alias scripts, but it treated the entire multiple-command script as a single line. I much preferred the structured presentation of VALIAS, with each command on its own line. I suggested that VALIAS should be extended to automatically indent the lines to show the nesting of flow-control commands.

It has been a long wait, but finally the wait is over. Rob Friefeld, from the Los Angeles area (contact him on Al Hawley's Z-Node #2), has released SALIAS (screen alias editor), and what a beauty it is. You will no longer find VALIAS on any of my disks!

With SALIAS, alias scripts are displayed and edited rather as if they were WordStar text files. Each individual command is displayed on its own line, except that long lines can be continued on the next line by entering a line-continuation character (control-p followed by '+') at the beginning of the continuation line.

SALIAS works in two basic modes. One is the command mode, like that in VALIAS. In command mode, the status line at the bottom of the screen displays the following prompt:

CMD (Clear Edit Format Indent Load Mode Print Rename Save Undo eXit)

Entering 'C' will clear the script editing area. 'E' will enter full-screen editing mode. 'F' will reformat the script, converting it to upper case and placing each command on its own line, even if the user entered lines containing multiple commands separated by semicolons. The 'I' command is similar except that it indents the lines by three extra spaces for each level of IF nesting. Thus a script might appear as follows:

IF EQ $1 //
OR NU $1
IF NU $2

This format makes it very easy to see the relationship between the flow tests and to detect missing FIs.

Entry of complex conditional aliases is greatly facilitated by the use of the tab key to indent the script as it is entered. The 'I' command will reindent the display according to the actual flow levels, even if the user made a mistake. The spaces (tabs) used to format the display are not part of the actual alias script. However, leading spaces entered by the user (to invoke extended command processor operation, for example) will be included in the script and are displayed in addition to those added by the indenter. The 'F' command will show the real contents of the script, automatically deleting the indentation tabs.

The 'L'oad command tells SALIAS to clear any existing script and to load an alias file for editing. If such an alias already exists, it will be read in. Otherwise it will be created. 'M' allows one to specify either a normal alias or a VALIAS-style recursive alias, one that clears the entire multiple command line buffer before it is run. In an earlier TCJ column I described how one would use this kind of alias for certain kinds of recursion.

The 'P'rint command will send the screen display of the alias script to the printer. 'R' will let one change the name assigned to the script being edited so that a script can be read in from one alias and written out to a new alias. 'S' saves the current script in a file with the current name. Undo will ignore any editing that has been performed on the alias and let the user start over with a fresh copy of what was read in from the file originally. 'X' will terminate operation of SALIAS (without any prompting).

SALIAS has an alternative mode of operation entirely from within the interactive edit mode. All of the functions that can be performed by commands in command mode, and some others as well, can be performed using control-character sequences directly from edit mode. These commands are as follows. In all cases, the first character (for example, ^K) is a control character. The second character can be a control character or a regular character in either upper or lower case.

^KSSave the alias under the current file name.
^KDDone editing - save the file and then clear the edit buffer.
^KXExit from SALIAS after saving the file.
^KNAssign a new name to the script.
^KQQuit without saving the script.
^KRRead in another alias script and append it to the commands currently in the edit buffer. This is very convenient for combining scripts from multiple aliases.
^KFReformat the alias, listing each command on its own line, removing any flow-control indentation, and converting to upper case.
^KIIndent the alias display to show flow control nesting.
^KUUndo changes that have been made to the script.
^KPPrint the script.
^KMToggle the mode of the alias between normal and recursive.

The editing commands follow the familiar WordStar pattern. Even ^QS (move to beginning of line), ^QD (move to end of line), and ^QY (delete to end of line) are recognized. The special form ^QZ will zap (clear) the entire script so you can start over. ^R and ^C move to the first and last lines of the script, respectively. ^QF allows searching for strings, and ^QA allows search-and-replace. ^L will repeat the last search or search-and-replace. ^V toggles between insert and overtype modes.

SALIAS is available in two versions. The longer version has imbedded help information that can be called up using ^J. In the shorter version, the help information has been omitted, and ^J instead is used to toggle the cursor between the beginning and the end of the current line.

I should mention a few other features of SALIAS. There is an additional status line at the top of the screen. It shows the name and version number of the program, the type of alias (normal or recursive), the number of characters free, and the current name for the alias.

The free-character value is calculated by subtracting the number of characters presently in the script from the number of characters allowed in the multiple command line buffer. This computation is not infallible. There are some parameter expressions, such as $D, that take up less room when expanded, so it is possible that SALIAS will refuse to let you save an alias that it thinks it is too long when in fact it is not. More likely, however, is that you will save an alias that has few enough characters for SALIAS to accept it but will become too long when the parameters are expanded. And even if this does not happen, you can run into trouble when the alias itself appears in a multiple command line expression, and not-yet-executed commands have to be appended to the alias script. These subtleties aside, having the display of free characters is helpful.

To summarize, in my opinion SALIAS makes all previous alias editors obsolete. You should be sure to pick it up from your local neighborhood Z-Node. If you do not have one, then join NAOG/ZSIG and order a disk from them.

VLU - Visual Library Utility

VLU is a utility we have all been wishing for - a screen-oriented library management utility. Its author is Michal Carson of Oklahoma City, OK. I originally suggested a library shell, like ZFILER but working on the contents of a library, but Michal pointed out that there is really no need for the visual library utility to be a shell, since shell functions like macros performed on pointed-to files will generally not be applicable to library members. So he built VLU as a non-shell utility that interfaces closely with ZFILER.

VLU can be invoked in two ways. Without any file name specified on the command line, it tries to open a library whose name is the same as the file name stored in the Z3ENV system file 2. This system file is where ZFILER keeps the name of the file it is currently pointing to. If one has a ZFILER macro called, for example, 'V' with the simple script 'VLU', invoking this macro while pointing to a library (or to another file with the same name as an LBR file in that directory), the library will be opened for work by VLU. Alternatively, one can specify the name of the file on the command line, in which case this name takes precedence over any name in system file 2.

Once VLU is running, the display contains two fields. The upper field is like that in ZFILER. It displays the names of the files in the currently logged directory. The lower field is similar in appearance but shows the names of the member files in an open library file. If no library file is currently open, this field is blank.

As in ZFILER, there is a file pointer that can be moved around using standard control characters or, if defined, cursor keys. In VLU, the escape character toggles the cursor between the upper (file) and lower (LBR member) fields of file names.

Many operations can be performed on either set of files. Files can be tagged and untagged, and two wildcard tagging functions are provided. 'GT' group tags all files, while 'W' allows a wildcard file specification to determine which files get tagged. Individual files or groups of tagged files can be viewed, crunched (but not squeezed), or uncompressed (either uncrunched or unsqueezed, as appropriate). The 'F' command will show the size of an individual file or library member at the cursor. For library members, the size is shown in records as well as kilobytes. Additionally, tagged individual files can be deleted or renamed, and tagged library members can be extracted, with or without decompression.

The special 'GB' or Group-Build function of VLU allows tagged individual files to be built into a new library. As I write this article, the details of this feature have not been fully worked out, but it will be possible for the user to indicate which of the tagged files should be crunched according to an automatic algorithm (which will crunch them only if that results in a smaller file) and which should be put into the library in their existing form.

Single individual files will accept the command 'O' to open a library with that name if it exists and the command 'C' to close the currently open library. They can also be renamed using the 'R' command.

There are several miscellaneous commands. '/' will toggle between a built-in help display and the file name display. 'X' is used to exit from VLU; 'J' is used to jump to a file; '*' is used to retag files that were tagged before some group operation was performed; and 'Q' refreshes the screen display.

Future versions of VLU will include some features not yet in the present one. A printing capability will be added to complement the 'view' functions. Presently, VLU cannot add files to an already existing library, though it allows the user to specify the number of elements to accommodate in the library directory so that another tool such as LPUT can be used to add more files later. VLU has an 'L' command to log into a new directory. By opening a library in one directory and then changing to another, one can extract files to a directory other than the one containing the library. At present, however, the 'L' command does not recognize a file mask as ZFILER does to restrict the files included in the display. Even in its initial release form VLU is a very welcome addition to the toolbox of Z utilities, and I extend thanks from all of us to Michal Carson.

Subject for Next Time

As I promised last time, this column has taken a less technical tack, though I feel that it has covered important and valuable material. For next time, however, I expect to return to a more detailed technical discussion. In the past few weeks, I took up the task of rewriting and expanding ZEX, the ZCPR3 in-memory batch execution utility. This has been my first detailed look at a program of this type - a resident system extension (RSX), a program that takes over and replaces functions of the operating system, in this case the BIOS. I have learned a great deal and have made what I think are some spectacular improvements to the way ZEX works and additions to what it can do. Next time I will share with you what I have been doing and what I have learned.

[This article was originally published in issue 30 of The Computer Journal, P.O. Box 12, South Plainfield, NJ 07080-0012 and is reproduced with the permission of the author and the publisher. Further reproduction for non-commercial purposes is authorized. This copyright notice must be retained. (c) Copyright 1988, 1991 Socrates Press and respective authors]