Interlisp project: Restore Interlisp-D to usability on modern OSes | Hacker News Reader

Interlisp project: Restore Interlisp-D to usability on modern OSes

interlisp.org

140 pointstosh24 comments

I missed the golden age of lisp, but i know this project for its legendary structural editor, SEdit[1], released the year I was born. i found it mentioned in an old paredit readme, something poignant like “paredit is only a shadow of what interlisp was: true structural editing”.

[1]:https://www.youtube.com/watch?v=2qsmF8HHskg

What I find fascinating is less the balancing and structural editing, but that these old lisps _strongly_ embraced the absence of source files. The environment was being edited, and not some text documents.

> The environment was being edited, and not some text documents.

That's all very well until you want to put your code in source control.

I don't see why the environment itself couldn't be adapted to have modern revision control.

This is in fact exactly how modern Smalltalk systems interact with source control.

Which Smalltalk environments and SQL databases have tackled already for a couple of decades.

Well, some older Lisps not only did support files, but also file versioning on operating systems allowing for that. You still have provisions for that in ANSI Common Lisp spec.

Where? I’d be curious to read it.

http://www.lispworks.com/documentation/HyperSpec/Body/f_pn_h...

https://en.wikipedia.org/wiki/Versioning_file_system

This is how Smalltalk is and was. Source-code is there but mostly for the purpose of restoring changes from the file change.log in case the "image" crashed.

Now, add the change.log file to the version control and you have a block-chain?

Aye, the thing is... I just haven't "clicked" with the ergonomics of Smalltalk systems. The language isn't too appealing, for starters; if feels too much like I'm using reverse polish notation, or a stack-based programming language. ;)

The golden age of Lisp is now.

You might not have been one of the lucky few people to have access to the institutional hardware running Lisp during the golden age of Lisp you're referring to.

To try to recapture the sense of fluidity in the development cycle that I've not encountered since.

Er. no. My comment on HN when this came up 6 years ago: [1]

There were some good ideas in Interlisp, ones that could be revived for other languages. One was the ability to select a block of code and have it pulled out and made a separate function. This was a safe transformation. All necessary variables were made parameters. Not hard to do in LISP. More useful for C++ refactoring.

[1] https://news.ycombinator.com/item?id=8751081

I think there are a few things like that in emacs lisp but not well distributed. Maybe people want to discover them for themselves.

From another comment in this thread:

> paredit is only a shadow of what interlisp was: true structural editing

I recently read, in another thread whic I can not find, that Visual Studio does this for at least C#.

The ‘Lexical Functional Grammar’ image mentioned in the article used to be available together with an emulator that ran on the Alpha under OSF/1. I remember this because it caused me to slightly improve the OSF compatibility layer of NetBSD¹ so I could run it on a surplused DEC 3000 that I'd picked up for $100.

¹https://mail-index.netbsd.org/netbsd-bugs/2001/05/30/0003.ht...

If curious see also

2013: https://news.ycombinator.com/item?id=5966328

Here's a ready-to-run image: http://autergy.com/medley.zip

I had an Xerox 1108 Lisp Machine. I used InterLisp-D for the first year and a half of ownership and then switched to their Common Lisp implementation.

InterLisp-D was great fun and I would enjoy having for nostalgia. That said, the modern Common Lisp ecosystem is awesome as is its community so I personally would not use InterLisp-D for large projects.

This is a great project. But it's only in the early planning phase so far, right? Or is there already some code on Github?

Looks like they have their hands on the code and are making changes, but haven't put the actual code on GH yet (maybe still working on getting permission?): https://github.com/Interlisp/medley/issues/2#issuecomment-66...

I'm on the fringes of this project, trying to shake enough free time loose to contribute to coding - I'm still employed beating code into submission for NASA.

What we have is access to the source code for the D-machine emulator that Interlisp-D/Medley Common Lisp ran on in the 1980's.

The emulator was quite usable on 1980-s hardware - I worked for the company that sold it in the late 80s and early 90s.

We are still deciding stuff like what exact open source license to use, and how much priority to give to the two main goals:

* Historical - can it run old system dumps with important systems like NoteCards, LOOPS, LFG...

* Modernization - integration with host OS (clipboard, file system, networking), increasing memory space, adding Unicode

The Teletype editor was a marvel. The whole program on one line (cond & & ..). In 1976 I remember I tried to invent something similar to natural languages. Finnish was very suited for this, with long words with plenty of affixes, but the base words are mostly short.

Reply on news.ycombinator.com