ORVYL and WYLBUR

The names ORVYL and WYLBUR are often used interchangeably, but:

• ORVYL is a timesharing monitor that supports a file system, command language, program execution and debugging, and provides supervisor services.[1] The first version runs only on a 360/67, but later versions run on a System/370.
• WYLBUR is a text editor, word processor,[lower-alpha 1] job submission and retrieval, and e-mail program designed to work in conjunction with ORVYL or with IBM's OS/360, SVS, and MVS operating systems.[2]
• MILTEN is terminal control software used by both ORVYL and WYLBUR for start/stop terminals.[1]

WYLBUR is not a full standalone operating system in the mold of Dartmouth Time Sharing System (DTSS) or Unix. Instead it runs on top of an IBM batch operating system (OS/360, SVS, MVS). It takes the form of an editor with a Remote Job Entry system and thus has much the same relationship to the IBM operating systems as Emacs does to Unix. For these reasons WYLBUR is often thought of as a text editor rather than a time-sharing system. However, whereas Unix does not need Emacs to provide text editing services, IBM's operating systems originally needed WYLBUR. Later innovations such as IBM's Administrative Terminal System (ATS), Conversational Remote Batch Entry (CRBE), Conversational Remote Job Entry (CRJE) and Time Sharing Option (TSO) made WYLBUR less relevant for IBM users and gradually replaced it.

This article will use the full upper case spelling for commands and keywords. All references to characters and string assume an EBCDIC code page.

ORVYL and WYLBUR were used at the Stanford Linear Accelerator Center (SLAC), the European Organization for Nuclear Research (CERN), the U.S. National Institutes of Health (NIH), and many other sites. Retired from most sites in the late 1990s owing to concerns about Y2K issues, they remained in use at NIH until December 2009.[4] ORVYL and WYLBUR are still available as open source from Stanford.[5] There are also proprietary versions such as SuperWYlbur.

ORVYL and WYLBUR were much admired as shown by this excerpt from a 2004 article titled "Computing at CERN: the mainframe era":

[In 1976 the IBM S/370-168] also brought with it the MVS (Multiple Virtual Storage) operating system, with its pedantic Job Control Language, and it provided the opportunity for CERN to introduce WYLBUR, the well-loved, cleverly designed and friendly time-sharing system developed at SLAC, together with its beautifully handwritten and illustrated manual by John Ehrman. WYLBUR was a masterpiece of design, achieving miracles with little power (at the time) shared amongst many simultaneous users. It won friends with its accommodating character and began the exit of punch-card machinery as computer terminals were introduced across the lab.[6]

ORVYL and WYLBUR first became available in 1967–68, before TSS/360, TSO, or any other official time-sharing solution from IBM. This was roughly the same time that third-party time-sharing systems such as MTS became available and the under the radar development effort of CP-67 at IBM's own Cambridge Scientific Center took place. WYLBUR had the additional advantage that it could be used in conjunction with IBM's mainstream operating system, OS/360.

WYLBUR is a single-address-space system, unlike TSO. This conserved memory in the days when memory was precious. So even when TSO was available, organizations seeking to minimize memory use would often keep some or even a majority of their users on WYLBUR rather than letting them use the TSO interactive environment.

WYLBUR provides compressed Partitioned Data Sets (PDSs, aka libraries) to save disk space. In MVS, source code is typically stored as a sequence of card images (80 character lines). If a line contained only one or just a few characters, 80 characters were still used to store that line. Even when data, e.g., source code, are stored as variable blocked (VB), space could be wasted on strings of embedded blanks. WYLBUR implements stream-oriented storage of text in PDSs, (and sequential data sets) so that a one character line might only take 16 characters (line length, offset, chunk length, character) rather than 80 to store. WYLBUR, or an external program run via JCL, was used to convert files to and from the WYLBUR EDIT format.

Although TSO allows a user to do more than a locked-down WYLBUR system did, it is possible to write WYLBUR Exec scripts that execute batch jobs to perform functions that ordinarily would have required a TSO account, filling a batch job skeleton out with parameters, submitting the batch job, retrieving the output and displaying it on the screen.

WYLBUR has some security advantages over TSO, and some disadvantages. Advantages include:

• Being able to write rules to restrict user access to datasets other than those owned by them and stored under their prefix. This is analogous to a user's home directory on UNIX, and looks something like WYL.AV99.HCO, where AV99 is roughly analogous to the "group" and HCO the "user" within the group.
• Being fairer about resource use. WYLBUR doesn't implement commands such as TSO's alloc which can intentionally or unintentionally prevent others' access to data files for an extended period of time or use tremendous amounts of memory or CPU time. In this way, it minimizes the impact of any single user on all other users.
• Commands to set certain status parameters or "spy" on the commands being executed by other users were restricted to administrative users and could not be executed by regular users.

Disadvantages related to security included:

• WYLBUR is a single-address-space system. That means that if a user can figure out how to access raw bytes in the address space, they can potentially access information they do not own. For example, there once existed a program written by two college students in the WYLBUR Exec scripting language which could dig the password of the most recently logged on user out of WYLBUR's memory.
• Because the WYLBUR process runs under the system account assigned to WYLBUR, one is completely dependent on its enforcement of dataset access protections according to the rules set up in WYLBUR. Enforcement of the access rules could be completely disabled by an administrative user, for system maintenance purposes, who might not remember to re-enable them.
• WYLBUR implements disk quotas, with an interesting twist: any system user could give away all or part of their quota to other users. This functionality could be combined with typical course-related student accounts that went away at the end of every semester, and computer-savvy student staff who had non-expiring accounts with low disk quotas, in a manner not always anticipated by university staff.
• In systems running the ACF2 security package, a user with accounts in both TSO and WYLBUR that are tied to the same account name could reset the contents of their WYLBUR account's security record interactively from within TSO. This could be used to turn a regular WYLBUR user into an administrative WYLBUR user, increase its disk quota, etc.
• At least through the 1960s, the WYLBUR security rules were not enforced for batch jobs running on the same system. So, utilities such as IEHLIST and IEBGENER could be used to discover, read, and modify files belonging to other WYLBUR users unless you password protected those files, which was operationally awkward.

Wylbur had a special edit format for Wylbur data sets, which are compressed and have a line number and revision flag for each line. In addition Wylbur supports standard FB and VB datasets. A Wylbur user normally specifies a default volume, which may be the special word CATLG. requesting a search of the catalogue for an existing dataset and requesting that a new dataset be cataloged. Wylbur has the ability to convert line numbers between edit and IBM data sets, either as scaled integers or with an explicit decimal point.

Wylbur provides a line editor that works with temporary data sets, similar to buffers in other editors. At any point in time one of the temporary data sets is designated as default. Wylbur maintains a current line pointer for each temporary data set. The user may specify an explicit working data set on a command; if he omits it, then the default temporary data set is used as the working data set.

The unit of operation is a set of lines (associative range) and individual lines are identified with a line number in the range 0.0 to 99999.999; leading zeros in the integer part and trailing zeros in the fractional part may be omitted. The user can specify a line number in any of the following ways

1. Absolute line number: ddddd.ddd
2. FIRST: first line in the working data set
3. CURRENT: the current line for the working data set
4. LAST: last line in the working data set
5. END: a target for copies, following the last line in the working data set
6. relative: line+ordinal or line -ordinal.
7. macro variable containing a line

Wylbur libraries have a nonstandard format, however it allows the user to export to native OS files with integer sequence numbers and to import native OS files with integer sequence and rescale the line number by a factor of 1000.

A range can be specified as a combination of

1. An explicit range, e.g., 5.3-2/7.4+3
2. A pattern, e.g., 'X' DIGITS*3
3. A pattern with a column range, e.g., 'BAL' 10/15
4. An ordinal, e.g., 3RD 'BAL' 10/15
5. An ordinal relative to a match, e.g., 3RD AFTER 'BAL' 10/15
6. A sequence of lines following a match, e.g., EVERY 3RD AFTER 'BAL' 10/15
7. A Boolean operation, e.g., 'PAGE' INTERSECTION COMPLEMENT 'NUMBER'
8. A specification in parentheses prefixed by SET, e.g., 'PAGE' INTERSECTION SET (COMPLEMENT 'NUMBER')

A pattern is similar to a regular expression, but the syntax is closer to that of SNOBOL than to that of Unix or Perl, there is no backtracking and only the NIH Wylbur has capture of subpatterns. A pattern may be:

Sample commands

CHANGE 73/80 TO X10+10 IN EVERY 3RD LINE OF 'ABC' 1/3 in 1.5-2/3.7+2
COPY 'BAL' 10/15 TO EMD
CHANGE 73/80 TO '' IN EVERY 3RD LINE START 2 OF 'BAL' 10/15
CHANGE '(' DIGITS*3 ')' TO SUBSTRING 2/4

The specification of base+increment means that the replacement text on the first line is base and is incremented on subsequent lines, so that X10+10 replaces the matched text on the first line with X10, on the second line with X20 and on the third line with X30. The specification of SUBSTRING 2/4 means columns 2-4 of the matched string; note that this is less flexible than captures.

Various organizations developed enhanced versions of Wylbur. These included

• National Institutes of Health[7][8][9][10]
• Online Business Systems which was acquired by ACS (Affiliated Computer Services Inc).
• Optimum Systems Inc., sold to Electronic Data Systems and later spun off as SuperWylbur® Systems, Inc.[11][12][13][14][15]
• RAND Corporation:
  • WYLBUR Command Facilities 1975
    • This manual describes two related facilities that extend WYLBUR in the direction of a programming language that supports structured text entry and text manipulation applications.
      1. The first facility is an extension to WYLBUR's command vocabulary that gives WYLBUR many of the capabilities associated with traditional programming languages, The extension was written by Paul Andersen.
      2. The second facility is a batch preprocessor that permits the WYLBUR programmer to develop WYLBUR command programs in a language similar to PL/I. The preprocessor was developed by David J. Smith.
  • WYLBUR Learner's Guide
  • WYLBUR Reference Manual

• Time-sharing system evolution

• Roger Fajman; John Borgelt (May 1973). "WYLBUR: an interactive text editing and remote job entry system". Communications of the ACM. 16 (5): 314–322. doi:10.1145/362041.362234. ISSN 0001-0782. S2CID 6892186.
• ORVYL and WYLBUR software download from Stanford University
• Usenet notes on WYLBUR
• WYLBUR Command Facilities, The Rand Computation Center R-1555/10, The Rand Corporation 1975