This is the multi-page printable view of this section. Click here to print.

Return to the regular view of this page.


More background

Interlisp-D was launched as a product from Xerox in 1980. It was a base from which AI applications were built, as well as a research project in its own right.

Interlisp evolved the notions of structure editing from an interactive “teletype” structure editor, a Display editor (DEdit) to SEDIT. Masterscope and the file manager included system-management tools that combined features of version control and build systems, with comprehensive cross referencing support.

The 1970s and 80s saw major advances in computing and Human-Computer interaction. Interlisp was one of the many experimental systems developed during these heady times. It represents and early example of a system with a GUI (Graphical User Interface), and IDE (Integrated Development Environment) and within the Xerox PARC environment, it was an early example of a network enabled machine supporting email, file sharing, printing and more. In essence it was a forerunner of today’s computing systems.


Medley Interlisp represents the last Xerox Interlisp release. Medley continued Xerox PARC’s experimentation with developing systems for D-machines, a family of computers (Dorado, Dolphin, Dandelion, Daybreak) descended. The project was later spun out to a company called Envos, which then turned into Venue.

Medley Interlisp User Interface.


Our aim is to allow Medley Interlisp to “live again”: usable on modern systems, sufficient to allow someone to develop some code and experience what it was like. You could think of this as a kind of “vintage software” project, to try to capture the sense of fluidity in the development cycle.

We hope to provide a platform for demonstration of early experiments of hypertext (Notecards), Desktop management (Rooms), Object-oriented programming (LOOPS), as well as Interlisp itself.


Primarily this is a matter of testing and debugging. Identified issues are documented in the Interlisp/medley repo.

We’re also trying to “modernize” some things; for example, we’re adding Unicode support for IO (Interlisp-D was built before Unicode, and supports an older Xerox encoding). As with most open source projects, what we’re able to accomplish depends on volunteers.

Medley was the last release of the Xerox Lisp environment, before the whole environment was renamed Medley. It was originally written in Interlisp (a dialect separate from the MACLISP/Common Lisp tradition, with its own ancestry), but later, Common Lisp also became part of the environment. Medley includes a WYSIWYG text editor (TEdit), email organizer (Lafite), performance tools (Spy) and many other libraries and user contributed code (from the 1980s).

The 1992 ACM Software System Award, to Daniel G. Bobrow, Richard R. Burton, L. Peter Deutsch, Ronald M. Kaplan, Larry Masinter, Warren Teitelman

… for their pioneering work in programming environments that integrated source-language debuggers, fully compatible integrated interpreter/compiler, automatic change management, structure-based editing, logging facilities, interactive graphics, and analysis/profiling tools in the Interlisp system.

What people are saying

More on organization and how you can help

1 - Ongoing Projects

We’re using the “projects” feature of GitHub to manage our work. There are (currently) 20 projects.

In addition, there are GitHub discussions for exploring and discussing possible directions.

An older list of ongoing projects (from 2021):

2 - Get Involved


The Interlisp organization on GitHub contains repositories for

  • Medley: Lisp code, build scripts and GitHub automation
  • maiko: The C-coded virtual machine
  • documentation (including this web site)

plus other related project components. Bugs, feature requests, and questions are best posted as GitHub issues in the Medley repository. Do check to see if your issue or question is answered there.

Participants in the project can join the GitHub Interlisp organization.


We have a weekly meeting on Mondays on Zoom. Meetings are to report accomplishments, status, plans, and problems, and discuss directions. Meetings are recorded, but recordings are not public, If you’d like to join in, let us know.

Monday, 1:30 pm ET / 10:30 am PT

Meeting Agendas and Notes

  • Recent meeting agendas and notes are here. Older agendas are here and here. Meeting recordings are made using Zoom “record to the cloud” and then download to google drive; but saving to a youtube private channel is a possibility.

EMail and access groups

There are two Google groups in use:

The archives are public. The LispCore group has access to meeting recordings, and usually gets more email (at least lately). Posts from non-members are moderated.

Other ways to discuss Interlisp or find out more

  • is a Google workspace; “” email addresses and aliases are available.

  • InterlispOrg Inc is a non-profit 501c3 corporation. We accept donations! Show your support!

  • Libera.Chat has a #Interlisp IRC Channel. A place for informal chat. Connect by pointing your IRC client to (TLS) and /join #interlisp. Say hello. There are also channels for #lisp (any dialect) and #common-lisp.

  • Twitter handle: interlisp8.

  • Direct email: In addition to the email groups you can contact by emailing us at

2.1 - 2021 Meeting Agendas and Notes

Agenda 13 Dec 2021

  • – getting to a (0.9) release
  • InterlispOrg Inc – 501c3 approval

Agenda 6 Dec 2021

  • Github projects?

Agenda 29 Nov 2021

Other Topics?

  • mkvdate.c (was on agenda pushed to issue)
  • SDL
  • Releases (medley, maiko, docker)
  • Using GitHub projects?
  • Documentation: Book 3, release notes

Agenda 22 Nov 2021

  • Deployment
    • SDL status (Peter)
    • Docker build (Bill)
    • (Frank)
  • git/GitHub Medley branch commi push workflow (Ron)
    • Masterscope issue
  • Annual Report feedback (Larry)

Agendas 8 & 15 November 2021

We spent both meetings reviewing the Annual Report and Round Table

Agenda 1 November 2021 NOTE NEW MEETING TIME

We should put out a progress report of what we’ve done, new things or soon anticipated. First annual report. In the meeting let’s try Google Docs.

Agenda 25 October 2021

Too much to talk about – lots is happening all at once.

Agenda 18 Oct 2021

  • round table: Status & plans
  • README’s and release notes
  • Distributions

Agenda 11 Oct 2021

  • Documentation (reader’s guide, SK books)
  • Releases and automation
  • Help wanted bugs (#55 eval, MV return stack)

Agenda 27 September 2021

Lots of things to talk about, that might require some preparation

Let’s start with (5 min each max) round-table. Then 15 minutes each on plan for review

  • Documentation:
    • what we have (Wiki list, PDFs, site)
    • what we’re working on (SK)
    • What we need (Documentation issues, bibliography)
  • Subrs and opcode survey
  • Unicode plans (internal and external encodings)
  • Releases and distributions

Agenda 20 September 2021

Round-table. Save for last:

  • BACKGROUND-YIELD -> liapusers (for now)
  • Releases
  • no commit of loadups => need loadups-only release
  • (dribble files? need copyall?)
  • Interlisp/DOS

Agenda 13 September 21

  • Interlisp Org Inc status, Q&A
  • Zotero (Herb?)
  • Releases (Bill?)
  • welcome Peter
  • Envos files
  • Fonts progress
  • LispUsers

Agenda 6 September 2021

Labor Day (US Holiday); attendance uncertain. Let’s do a “round table” of brief status reports.

  • Interlisp.Org Inc – progress
  • Zotero bibliography
  • Releases
  • Documentation - FAQ?
  • Fonts (sigh)

Agenda 30 August 2021

  • loadups, releases, GitHub automation, Notecards configuration

Agenda 23 August 2021

before we dive into other topics, including:

  • *.TEDIT to PDF for all medley / envos repos
  • envos repo reorg / undo cr/lf transform
  • tests and benchmarks

Agenda 16 August 2021

  • Round table news
  • Tracking down bug 407

Agenda 9 August 2021

Agenda 2 August 2021

  • Testing – how to make test cases for changes
  • Git commit quality and separability
  • Recent PR merges
  • maintenance workflow
  • Simplify process of obtaining new releases (mjd)

Agenda 26 July 2021

  • DOSBox / uploads
  • Ron’s IO changes /Interlisp/medley/pull/348
  • Project goals & bylaws (TBA)

Agenda 19 July 2021

• 30 minutes topics important for filing for 501c3 articles of incorporation

  1. brainstorming a list of possible organizational goals. I imagine this as a “brainstorming” session where we collect ideas. It’s important to have a broad purpose statement to allow flexibility but specific enough to attract volunteers and donations. Consolidation, evaluation, and pruning will come later. The ideas should be expressed in terms suitable for filing for non-profit status. See Medley Interlisp Goals and Ongoing Projects for inspiration.
  2. Av list of ideas for the organization name.

I will record the session but it would be wonderful if someone could take notes (setting up a google document we all could connect to)

• 30 minutes round table: status, plans, blockers

Agenda 12 July 2021

I’d like to talk about overall project goals in relationship to Ongoing Projects

Agenda 5 July 2021

Ongoing Projects in reverse order

Agenda 28 Jun 2021

I’d like some feedback on Project Organization. My hope is to use this organization to drive the agenda.

Agenda 21 Jun 2021

Usual round-table status and plans

Agenda 14 Jun 2021

some topics to touch base on

  • Docker (Herb)
  • Notecards demo (Frank?)
  • Medley 2.0 on DOSBox (Arun?)
  • Space and versions and file name case (Bill? Larry)
  • EOL convention problems with DFASLs? (Ron?)
  • Meeting recordings available to LispCore for 2 weeks

Agenda 7 June 2021

  • Introductions (if invitees can make it)
  • Docker (Herb)
  • Topic “Notecards Demo”
  1. running old sysouts with DarkStar
  2. running old sysouls with Medley 2.0 on DosBox
  3. recompiling and loadup with modern medley
  • “International Keyboard Support”
  • “Zotero Interlisp bibliography”
  • Working with Fiverr
  • git-versions
  • GitHub space plan

Agenda 24 May 2021

Usual round table, hoping to include topics:

  • Docker plans (Herb)
  • Filter GitHub for big files; use GitHub LFS; or both (? Bill)
  • Running in Windows? (Steve)
  • Starting an Interlisp non-profit (Larry)
  • Stack overflow bug?

Notes 17 May 2021

These are rough notes from the meeting. I regularly record the meetings for private consumption.

  • Let me ask each of the attendees, your decision on the question of what to do: stop recording, available to LispCore members, available public, available on request. I want to respect everyone’s conditions, and would like an explicit ack. I’ll summarize before acting.
  • Reducing git repo size (Bill Stumbo) We decided “one change at a time” even if it causes multiple force push, and invalidates forked repos (There were a small number.) If the forkers can cope once, they can prepare for the second time). Bill will test his filter on a fork/copy and make sure it has the desired effect.
  • Steve was still having problems running WSL2 on Windows. Windows 10 Home vs. Windows 10 pro. We should resurrect the original instructions for running on Oracle Virtualbox or VmWare. Making progress on INTERLISP volume 2.
  • Bibliography progress (Abe) Abe showed off the Bibliography and advanced features on There’s a search feature but it’s buggy. (The implementation is in the GitHub Interlisp/ repository. I think Issues with the website and bibliography could just use Medley issues with tag “Documentation”.
  • Herb talked about the Docker and automation issues he was going to reverse inherit from Abe.
  • Character IO changes (Ron) Ron described his progress.
  • Versions from git (Larry) I demoed a ‘proof of concept’ for pulling old versions from git. I’ll keep you posted
  • Notecards Demo (Arun, Frank Halasz(?))
  • stack overflow multiple value return bug
    Nick and I talked over the fine details of who does what to the stack and why. The conjecture is this isn’t a new bug, just a fault that was rare.

John Vittal, Arun Welch, Abe Jellinek, Herb Jellinek, Larry Masinter, Bill Stumbo, Nick Briggs, Ron Kaplan, John Cowan, Steve Kaisler, Michele Denber

Agenda 10 May 2021

  • Round Table
  • Changing character encoding default (Ron)
  • The stack overflow debugging techniques (as a “case study”)

Agenda 3 May 2021

After trying other alternatives, I’m coming back to using the Wiki for agendas

  • it’s easy to find (not getting lost)
  • Anyone can update easily

This week, we will (hopefully) see John Vittal and Steve Kaisler (who joined us last Monday)

  • Introductions & usual ‘round table’. Topic suggestions
    • John Vittal welcome
    • Steve Kaisler’s Interlisp book and Volume II plans (Any success loading release?)
    • Bibliography (Abe)
    • EOL change in sources (Ron)
    • Lispusers (Herb?)
  • Other topics if there’s time
    • Loadups, releases
    • defaults for printing

Agenda 8 March 2021

  • As usual, what have you been up to? (Status and Plans)
  • Making releases:
  • Visibility
  • workflow on duplicate function removal (how to make each a PR)

Agemda 22 Feb 2021

Topics I have for tomorrow

  • Experimenting with loading all of sysfiles
  • Cleaning out “dead wood” in the Lisp code

regular features:

  • recent bug reports
  • round table

Agenda 15 Feb 21

  • Round table news & plans
  • Configuration vs. Documentation (Masterscope & DInfo, Match, fonts)
  • making the init and loadup review
  • reviewing the PR #187
  • debugging and analysis techniques

meeting recorded but available for 1 week on request

Agenda 25 Jan 2021

  • Recent documentation updates (READMEs, Wiki) and plans
  • Pinning CPU experiments
  • Common Lisp / Interlisp priorities

2.2 -

2022 Meeting Agendas and Notes

Every Monday 10:00am PT Zoom ID: 408 332 4290

Meeting recordings contact @masinter for access.

Agenda 2 May 2022

Agenda 11 April 2022

  • notices to other organizations (S Kaisler)
  • Web site organization (W.Marci? B.Stumbo)
  • round table

Agenda 28 March 2022

Review announcements and reactions. Time for broader announcements finally.

Agenda 21 March 2022

Larry’s out this week, but there was sentiment to meet anyway, so click the Edit button and edit this page to add your agenda items.

Agenda 14 March 2022

Let’s go through Documentation as well as open ‘Documentation’ issues and prioritize.

Agenda 7 March 2022

  • Issue review for release announcement Let’s quickly go through open issues labeled ‘online’ for a status update
  • SPN/Educopia consulting
  • Lispusers cleanup …

Agenda 28 Feb 2022

Agenda 21 Feb 2022

  • News (round table):
    • online (open issues)
    • GITFNS handling file moves
    • SPN and spreading the news
    • Branding and sponsors
  • Pick an issue: using “pick-issue” we will pick an issue and work on it together. If we can close the issue in 20 minutes, fine, but otherwise we will just write up what we learned. (later I’ll extend pick-issue to pick Lispusers packages too).

Agenda 14 Feb 2022

Happy Valentine’s day

  • New GitHub actions – how does it work?
  • LispUsers review
  • idea: 15 minute demos of Medley features (for later scripting and re-recording)

Agenda 7 Feb 2022

Agenda 31 Jan 2022

  • New user online experience (again) –
  • Web site review
  • round-table status and plans

Agenda 24 Jan 2022

Agenda 17 Jan 2022

Agenda 10 Jan 2022

Three topics; half hour each.

  1. Joining SPN (as a member, a sponsor, or some other relation). See SPN membership proposal and here for context; with Jessica Farrell and Eric Kaltman.
  2. Interlisp Online (hopefully Frank)
  3. New User Getting Started (Bill and Steve)

Agenda 3 Jan 2022

Happy New Year

  • Online status
    • new user setup (promptwindow, helptime, dwimwait, break window size & auto bt
    • payment set to use AWS $1k annual credit
  • SDL Status
  • Joining SPN (Visitors 10 Jan 2021)
  • People/skills to hire?
    • Videographer (plan, script, record, produce demos)
    • Historian (bibliographer, Interlisp history)
    • Lawyer (software preservation, Open source)

3 - Organizational Partners

Software Preservation Network

We joined SPN in 2021.

Universities: CSUCI

California State University Channel Islands

  • Professor Eric Kaltman

LFG (Lexical Functional Grammar) group

  • Set up to allow tax-deductable donations for student travel to conferences
  • LFG is a current Medley application maintained by Ron Kaplan

Computer History Museum

  • PARC is purportedly giving a grant to CHM of software collection, including lisp sources
  • hosts Smalltalk Zoo
  • offered to record history interview
  • SPN member

Internet Archive

  • MAME

Software Heritage

  • UNESCO sponsored activity to collect and archive software
  • Longer term focus claims are on reproducibility of software artifacts

3.1 - Educopia

Educopia is a non-profit organization hosting the Software Preservation Network and other software preservation groups.

The also offer additional resources on organizational principles.

  • Community Cultivation Framework: a framework for thinking through a community’s development.
  • Vision in Formation: “Articulating Your Community’s Purpose”: A community facilitator’s guide around starting those mission, vision, and values statements, with templates, including sample agendas, forms, and boilerplate email text.
  • “It Takes a Village”: Tools for planning sustainability for open source software initiatives serving cultural and scientific heritage organizations.
  • The Community Tool Box: templates and advice around community assessment, developing strategic plans and organizational structure, and building community.

3.2 - Software Preservation Network

The Medley Interlisp Project has joined (as an “implementation sponsor”) the Software Preservation Network. SPN is a coordinated, distributed effort to ensure long term access to software through community engagement, infrastructure support and knowledge generation. See Interlisp and SPN for more information.

SPN is hosted by the Educopia Institute.

They recommended the following resources:

  • Community Cultivation Framework: a framework for thinking through a community’s development.
  • Vision in Formation: “Articulating Your Community’s Purpose”: A community facilitator’s guide around starting those mission, vision, and values statements, with templates, including sample agendas, forms, and boilerplate email text.
  • “It Takes a Village”: Tools for planning sustainability for open source software initiatives serving cultural and scientific heritage organizations.
  • The Community Tool Box: templates and advice around community assessment, developing strategic plans and organizational structure, and building community.

3.3 - InterlispOrg

InterlispOrg Inc has beem formed as a non-profit corporation in California.

We have been approved by the IRS for 501(c)3 status, and the California FTB for tax-free status. We have an EIN number assigned. We have a bank account. We can accept donations that may be tax-deductable for the donor. The simplest way to do so is using the GitHub Sponsor mechanism.

The name of the organization is “InterlispOrg Inc” (no punctuation). The board of directors met and, among other things, approved bylaws. It is not a membership organization.

There are three directors on the Board of Directors:

  • Larry Masinter (President)
  • Ron Kaplan (Treasurer)
  • Herb Jellinek (Secretary)

The charitable purpose of the organization is Scientific and, in particular, Computer Science research. The principal activity is to “develop demonstrations of software ideas based on Interlisp”.

The reason for going through this rigamarole is to allow charitable donations, to sustain our Interlisp Online service, to hold copyright and licenses. There are also unanticipated benefits of discounts and service credits available to non-profits.

4 - What people are saying

Comments on Interlisp and this project

Via email

The Interlisp and LispCore Google Groups have open discussions; some positions on the project are here.


  • My encounter with Medley Interlisp by Paolo Amoroso Imagine someone let you into an alien spaceship they landed in your backyard, sat you at the controls, and encouraged you to fly the ship. This is the opportunity Medley Interlisp offers….


Link to tweet

Link to tweet

Link to tweet

Link to tweet

Link to tweet

Link to tweet

Link to tweet

4.1 - Individual Stories

Blake McBride

Imagine finding an old car from the 1960s. It’s been sitting in storage for decades, doesn’t run, and looks like hell. You remember driving the car when you were young and how much fun you had with it. It had a number of features you enjoyed back then that don’t exist anymore. You want to re-live the experience and be able to share it with future generations so that they can experience the same thing and learn from the lessons it taught.

You decide to fix the car up. Make it run well again. Make it look good again. It’ll be close to what it once was - a thing of beauty of that time.

The thing you don’t want to do to it is make it modern. You don’t want to add air conditioning to it, and you don’t want to add anti-lock brakes to it. You don’t want to do that for several reasons as follows:

a. You lose the whole value of bringing the old experience into the present.

b. You are trying to make the old car modern. Ultimately, you can’t compete with modern cars and all of your “enhancements” are just tack-on garbage.

I like InterLisp for what it was. It was a thing of beauty in its own right. I’m not looking to create another “modern” development environment out of it. Additionally, there are several top-notch Common Lisp systems available. Making yet another Common Lisp implementation is as utterly a waste of time as I can imagine. To this end, I think emphasizing and enhancing Common Lisp is doing more harm than good. Same for things like new fonts.

What we should be doing is:

1. Make sure it runs really well and is as portable as possible.

2. Emphasize InterLisp rather than Common Lisp.

3. Fix bugs and make it run as close to some spec that we chose.

4. Do not make enhancements that lie outside the spec.

5. Focus on making it look and run as it originally did.

John Cowan

When I went to summer camp in the 1970s, the director owned a 1932 Packard that we sometimes got to take rides in. But we could only do that on the camp property because the car was completely unsafe under any other conditions (for one thing, the starter didn’t work: you had to push the car). By now, riding in it at all would probably be either illegal or contrary to the camp’s insurance policies.

This isn’t directly on point for Interlisp, but it illustrates why you might need anti-lock brakes even if not air conditioning.

The Society for Creative Anachronism provides a medieval experience for its members and guests, but without the original filth and consequent disease. The simulation of Interlisp-D isn’t going to be perfect.

I’m not saying that it makes sense for the Medley development environment to support, say, Java programming. But Common Lisp is another matter.

Common Lisp is *part* of Medley (and its predecessors Lyric and Koto). The Interlisp specs have always been a posteriori: they tell you what the existing implementations have in common. That’s a good and valuable thing, but Medley is now the leading implementation (in the sense that OpenJVM is the leading Java implementation), and the spec is what Medley does (modulo OS interface bugs, of course. The Common Lisp standard is another matter. It is a priori: if you are writing or maintaining a Common Lisp implementation (at least one new implementation is currently underway) and it doesn’t agree with the current (1994) spec, it is buggy and you should fix it. If it were a question of adding Common Lisp support to Medley, I’d say “Forget about it!” But it isn’t. It’s just bad luck that most of the work was done between the 1984 and the 1994 standards. (No further standards are contemplated, so when it’s done, it’s done.)

At the very least, someone should run the ANSI test suite and see at which points it fails to be conformant. It’s known that some Common Lisp bugs and limitations were fixed since the Medley release, but nobody remembers which ones.

Note that I am not an Interlisp insider and have no axe to grind, except that Larry has said the purpose of this effort is to create a new group of Medley users/maintainers), and I think what will matter to them most is the programming tools. If they can develop on Medley and either deploy on Medley or on SBCL, that will be a big incentive.

Michele Denber

Well, I guess I might as well toss in my $0.02. I am not an Interlisp insider, just a user. I came to Lisp at Xerox back in the early 80’s. Coming from a Fortran/PL/I/APL/BCPL background, Interlisp-D was a revelation to me. The power of this environment was so amazing I couldn’t imagine ever going back to anything else.

IMNSHO, Interlisp-D needs just three things to be as good a development environment as anything around today: network connectivity, local device access (USB, printers, etc.), and color (just because everything is in color these days). All of these features were available back in the 80’s on the D-machines but were lost with the move to X.

The closest thing I’ve found to Interlisp-D in terms of quickly coding up some algorithm and getting it running is Matlab. But Interlisp-D as an IDE is far superior to Matlab.

Ron Kaplan

I was the one who was most interested in the resurrection, because of my desire to use it as a platform to explore new aspects of linguistic theory (LFG). Many years ago we (namely, John Maxwell in my group at Parc) translated the core algorithms from Interlisp to C, and that provided a high-speed commercial-grade implementation that became the basis for the Powerset spin-out company. But we took a 40-to-1 hit in coding productivity (with maybe a 1-to-40 efficiency gain), and it is virtually impossible for someone other than John to use the C implementation for experimental investigations.

It was a real break when Nick had the time, about 3 years ago, to figure out what was causing it to crash on the Mac. I have probably been the only “user” for the last 3 years before this community effort got started. During that time I encountered a few bugs in the emulator that Nick was able to track down (I remember a fat/thin string bug, a bigending bug) and at least one that is still lurking (stack corruption with multiple-value returns). But it has been remarkably stable—I have run for weeks if not months in the same sysout without crashing.

Along the way, I fixed a few other compatibility bugs to make it work better in the Mac environment, and I also did some modernization extensions (interface to the clipboard, moving and reshaping windows by dragging the title bar and corners instead of the silly pop-up menus).

Another important modernization that I have worked on and has been discussed is to replace the XCCS character encodings and fonts with Unicode, so that text files at least can move back and forth between Medley and the OS. Especially because the XCCS fonts that we have are so incomplete. And a few other things that make it easy to interoperate with modern interface conventions as they evolved away from our intuitions in the 80s and 90s.

But none of this would change the core behavior of Interlisp, it would just reduce some of the barriers to entry and make it less confusing to move in and out.

I am not particularly interested in Common Lisp per se, except insofar as Common Lisp has added a collection of often useful and sometimes elegant new functions that I call from time to time. In terms of capabilities, Common Lisp and Interlisp tend to generalize on different dimensions and so a mix can sometimes be helpful.

However: the integration of Common Lisp into the Medley development environment is incomplete, and in some ways broken. Key attractive features of the environment simply don’t work on Common Lisp FUNCTIONS and MACROS that are contained in files. I would never actually do a DEFUN or DEFMACRO in my own code, although I certainly do call out to Common Lisp functions. Common Lisp feels basically like a reversion towards a non-residential system. (There are also some inconveniences in the way that BREAK works.)

My primary goal then would be to complete the Common Lisp/Medley integration so that the development experience is consistent for FNS and FUNCTIONS etc. Upgrading Common Lisp to this or that standard might also be an attractor (or less of a disincentive) for others to come into this particular implementation, but I would first want to ensure a consistent Medley development environment.

Larry Masinter

For me, the importance of Medley is as a development environment. People who used it (including myself) were much more productive. It spoiled me so much I’ve never been proficient in anything else.

I didn’t think people could get the sense of the development without a lengthy session, to use it “in anger” – trying to get something else to work while using medley as a tool.

I think showcasing “things people built in Interlisp” shows the power of the development style, especially for research prototypes.

I think with a volunteer effort like this, it is to be expected that people have different things they want to get out of it, and the project goals are necessarily a synthesis of the individual goals. We have common subgoals. I don’t think there are any serious conflicts except perhaps priorities.

Getting the system stable is very important.

Getting the envos tests to run is a subgoal

There is too much to do with available manpower, so getting people to join is necessary. The first hurdle of contributing is understanding how to get started

Paul McJones

I’ve had the feeling that I don’t understand “the big picture”, and how the discussions fit in. For example, support for the Unicode character set and for modern scalable fonts would make it nicer to do work like Ron’s computational linguistics in Interlisp, but don’t seem too necessary to support people using Interlisp to understand its programming environment. Similarly, getting Interlisp to interoperate with git repositories seems like a distraction for someone who is trying to understand how Interlisp’s programming environment worked back in the 1970s and 1980s.

Here is my attempt to think about a “stripped down” path to preserving Interlisp history in a fashion allowing it to be used and studied:

1. Create one or more snapshot releases of Interlisp (a suitable virtual machine, a file system with sysouts, sources, and compiled files, and a matching set of documents (reference manual plus getting started guide) that will run on one or more modern platforms and that are complete enough to use the standard Interlisp tools to edit, compile, debug, and run an application. The snapshots would be distributed as compressed tar files and could be downloaded and run without any use of git or GitHub. If one of these could be adapted to run in a web browser, it might become the dominant way for “outsiders” to sample Interlisp.

2. Create a series of source snapshots corresponding to various interesting points in the Interlisp timeline for which complete or partial sources are available, for example:

Parc/BBN PDP-10

Parc D machines

Envos (D machines?)

Venue (Solaris)

Each snapshot would be a compressed tar file, and, where available, would include multiple versions of various source files, as maintained in the traditional Interlisp development environment. These snapshots would be archived (for example at CHM, and possibly software They would also be the starting point for creating the executable snapshot releases of #1.

3. The one component of Interlisp that seems compatible with git file-based development is the maiko virtual machine, although pragmatically I think the main activity would just be fixing show-stopper bugs and adapting it to new platforms.

But for the Lisp source files, it seems that informal mechanisms used in the 1980s to coordinate within a small group of developers are still the best way.

Nick Briggs

I’m mostly interested in getting the underlying emulator code working reliably. Unfortunately, there is little-to-no documentation available for the (C language) code itself, or description of the technical design and implementation details behind it. This makes it critically important to me to have the code history preserved, in its entirety, so that when I find something that appears to be a problem I can look at all the revisions that have been made to determine if there was an error introduced recently, or exposed by, for example, operating system or compiler changes, or is an endian dependency that was never resolved (for which we need to have both the original big-endian and newer little-endian versions). Since there is very tight integration with the lowest levels of the Lisp sources, I’m also interested in ensuring that we keep the history of the Lisp code available too – preferably in a way that permits easy comparisons of previous versions, and such that we can track the source in a running sysout back to the Lisp source files that we are keeping.

Arun Welch

I started working part-time for XAIS when I complained to John that the TCP/IP code was broken and he responded with “yeah, but all the AR’s reporting problems are coming from you and all the patches to fix those AR’s come from you so why don’t I pay you to support it since I don’t have any resources available to do that anyway”…

From there I supported the networking code, maintained the PCL port, and assorted other applications layered on top of the basic system as the ownership transitioned from XAIS to Venue to Envos.

I’m interested in getting some of the applications that were built on Interlisp/Medley working in the current environment, including:

  1. ROOMS
  2. NoteCards

Bob Bane

I would like to resurrect and extend the Common Lisp development part of Medley Making Common Lisp a fully supported part of Medley was an amazing achievement back in the day; I did a bunch of work in that environment and really liked the combination of Common Lisp and managed source files.

William Maddox

I saw the recording of your demo at the NYC Lisp group…. I figured I’d download it and give it a whirl over the weekend. I’m interested in seeing what a residential Smalltalk-80-like Lisp system feels like. I’ve used Allegro CL and various file-oriented Lisp and Scheme implementations over the years, but Interlisp-D was a bit different. I also have a “thing” for self-contained systems that can run on bare metal. I think Wirth’s FPGA Oberon is really, really cool, but it’s really a bit of an educational toy, and the language is a dead-end. A real Lisp system would be so much neater. My interest, if I got involved with any development, would be getting a bare-metal port of maiko running. Recently, a fellow combined a bare-metal support library with an ST80 virtual machine in C (both pre-existing open-source) and got an ST80 image running on bare-metal Raspberry Pi – no Linux in sight. In reality, though, a stripped-down embedded Linux would do the same thing, more or less. I really think to experience a Lisp machine, however, it needs to boot directly into Lisp. Running any sort of emulator “in a box” on a modern desktop, with a web browser, Facebook, and all those other distractions is just cheating. ;)

Stephen Kaisler

Back in the 1980s, I had written a book on Interlisp: The Language and Its Usage, which was published by John Wiley & Sons. The book has been out of print for years.

One of Xerox Electrical Optical Systems salespersons had told me they had purchased some copies to give to customers of their early D machines.

I had used Interlisp in several variations: Interlisp/360 (an implementation from Israel), Interlisp-D on a Xerox 1100 that I had at CIA, Interlisp-D on a Xerox 1186 that I had as a DARPA Program Manager, and Interlisp-VAX.

I began a subsequent volume entitled Interlisp-D: The Interactive Programming Environment, which had the same structure, format, and approach as the first book. It was submitted to John Wiley & Sons, who had initially agreed to publish, but then demurred after Xerox apparently stopped selling D-machines and Common Lisp came to the fore. The volume currently has 13 chapters and about 500 pages with numerous examples that I had tested on both my Xerox 1186 and on Interlisp VAX (the ISI version). It was written using Microsoft Word.

So, it languished for many years as I had hoped to see Interlisp-D be revived. Now, it appears that is the case with the Medley Interlisp project.

I am interested in participating in the project, re-editing the book with its examples, adding to it as appropriate, and adding it as another source of documentation to the Interlisp-D revival.

I wonder if you are also going to include LOOPS as part of this revision. I took the LOOPS course at PARC in 1983 with Peter Fisher of XEOS as my partner.

I am looking at the possibility of converting the first volume from PDF to Word so it can be further edited. Not sure if this is possible depending on how the PDF was created. If I can do this, I may split that volume into IA and IB to make it more manageable.

As I got more into editing the second volume, I realized it was going to be very big. So, I have split it into Volumes II and III:

  • Volume II: The Interactive Programming environment
  • Volume III: Tools and utilities like TEdit, DEdi, etc.

Once I start working on LOOPS, it may become a fourth volume. Others may arise as we get further into the project, but that is a ways off.

Guenther Goerz

I got familiar with Interlisp when I was working on an implementation of a parser for my PhD thesis work in the late 1970 and 1980s. It started with a little cooperation with Martin Kay (and also Ron Kaplan) from PARC at a computational linguistics summer school in Pisa. Although I had a tenured researcher position at my university, I did not have any means to acquire one of Xerox’ LISP workstations, which were sold in Germany by Siemens for something like 90000 Euros (today’s equivalent) - so I used Interlisp on a Siemens mainframe to which I had access. The Interlisp version was a descendant from the IBM implementation - both machines had basically the same hardware architecture but different operating systems (which was not a problem for the rather self-contained Interlisp environment). After spending some time with IBM’s LILIOG project at IBM in Stuttgart and getting my first professorship at the University of Hamburg, I came back to Erlangen in 1991. So, in the 1990s I used the Fuji Xerox emulation of Medley on Sun workstations and later on Linux machines. Additionally, we also acquired a license for the DOS version of Medley 2.0 from envos/Venue. What I used for some time was the LFG implementation in Medley - which I got from Ron Kaplan in Linux - in Suse Linux, running on my Mac(s) with the Parallels Virtual machine, mainly for teaching purposes. But I was also interested in various features of the programming environment… Recently, just for fun, I tried to get Medley up and running under DOSbox on Mac OS X. I was also wondering whether there is a DOS emulator for the ipad - and there is one, iDOS, which works pretty well! So, despite the small screen, Medley 2.0 runs amazingly well on the ipad too.

I must add that we (our chair at FAU Computer Science) first acquired a Medley 2.0 license for the Sun workstations through Xerox France first and the DOS license later. We used it for our first steps in the Behaim Globe project (also trying out NoteCards and implementing our first ontology with CLOS) and the actual portal has still some of the code deep inside and furthermore for NLP, in particular, the LFG grammar development environment. Since I was a simple user, my expertise is pretty small; I was never involved in anything close to system development. Later on, I switched to Scheme for teaching and research for reasons of clarity and simplicity and also for theoretical reasons - I never liked Common LISP very much. But of course, there was never such a comfortable programming environment for Scheme.

After my retirement in 2012, I focused on the development of our VRE Wisski in cooperation with two museums (, which turned out to be quite successful - in the meantime, there are several hundred applications. Furthermore, I helped to get Digital Humanities at our university up and running - they are starting a new department quite now - and I was involved in some DH projects at the Max Planck Institute for the History of Science in Berlin and more recently at the Bibliotheca Hertziana (Max Planck Institute for Art History) in Rome in the field of “cosmography” (which was the correct term in the Renaissance), i.e. modelling and annotating historical texts and maps (

A while ago I tried - just for curiosity - whether the old DOS version of Medley 2.0 would also run in the DOSBox under Mac OS X. It worked really well, and so I became even more curious whether there would be something like DOSbox for iOS - and yes, there is iDOS, basically designed to run old computer games. Now Medley 2 runs on my ipad… Actually, I don’t have any further plans, and I didn’t realize what happened after Medley 2.0. So, from your presentation at lispNYC I learned that there seems to be something like a new version 3.0. Sounds tempting… So, I would be glad if we could keep our contact active.

J Moore

I checked out That’s an interesting project. I am glad you’re doing it. Too often the first implementations of good ideas are overshadowed by evolutionary improvements. (In that spirit, a few years ago I reimplemented the first Boyer-Moore theorem prover – the one we built in Edinburgh in POP-2. But I implemented it in ACL2, i.e., applicative Common Lisp. But it proves the same theorems the same way as we did in 1972-3.)

As for the VM, I don’t have much context to add beyond what I wrote in the Acknowledgements of CSL-76-5, which I’m sure you’ve read if you’re interested in that question. I got my PhD from Edinburgh in November, 1973, and arrived at PARC in January, 1974. Danny Bobrow had taken a sabbatical from PARC (BBN?) in ‘73 and spent it in Edinburgh, where he convinced me to join PARC. The Acknowledgements says Warren Teitelman asked me to write the VM document in January, 1975. But I actually suspect it was earlier than that, perhaps as early as January, 1974 (i.e., I got the year wrong!). The reason is that the Acknowledgement says we spent 2 years writing it, but CSL-76-5 came out in September, 1976, so I suspect Warren approached me in the a few months after I joined PARC. For the record, I don’t remember doing anything at PARC besides (a) re-coding the Edinburgh theorem prover in Interlisp as a warm up exercise and to continue my research on theorem proving, (b) helping Simonyi with the representation of text in Bravo by implementing the underlying data structures in Interlisp and responding to Charles’ requests for new features, (c) writing the VM, (d) inventing the fast string searching algorithm with Boyer and fighting the lawyers at Xerox to publish it (ultimately by threatening to remove my name from the publication and let Boyer author it), and (e) fighting with Ben Wegbreit and Jerry Elkind over the ``right’’ way to approach verification and ultimately leaving. So I suspect that my VM work was my first and main official task at PARC.

Other than that, I think the Acknowledgements say it all, accurately. The line there that says I think my friends in the formal verification community will be displeased was not speculation. Tony Hoare wrote me a letter upon seeing the document in which he complimented me on the precision and urged me to get back to formal methods. Oh well.