How faithful of a clone are we talking?

[cb88]

Do you happen to have the Dynix handbook if not it might be handy to have ? I have a copy and there are quite a few examples of what screens could look like in Dynix there... many libraries had custom screens I think also since after all it was programmable.

Dynix was built on top of PICK OS which is a multivalue database operating system and PIC BASIC is used in it for some things too but that is about all I know... it would be cool to actually find an old dynix system intact and investigate that.

Dynix was the first public computer system I ever used! I even checked out books from our inter library system via modem as a kid using our Dynix system.

[cb88]

If one wanted to go about it "properly" maybe starting from https://mmbasic.com/ (or perhas http://dunkels.com/adam/ubasic/ as it is BSD licensed) and adding the features of PICK OS and PICK BASIC to it... and you would eventually end up with a Dynix alike. And it would probalby still run on retro hardware since mmbasic itself only needs about 16k of ram to start with.

I think the main reason to even do anything like that would be just for fun, and also because it would acutally work better as a real multi value DB .... rather than a relational database.

[p3r7]

Hi there!

I got started w/ book Dynix : a guide for librarians and systems managers. I don't know if your handbook is this or something else.

My aim was not to be 1:1 faithfull (that would be hard wo/ access to a running system or source files) nor run on older systems.

I only wanted to reproduce the end user consultation interface and maybe the librarian pages (document declaration).

One aim was to make it interact with instance of more "modern" ILS/OPAC, APIs (I've experimented w/ goodread's which sadly got closed recently) and software (Calibre).

From the few information I managed to find, the original Dynix is said to be some kind of a nightmare to install / configure.

I initially started experimenting w/ Python + curses because I was familiar w/ those. Still I'm not happy with current state of things as I've used classes / objects for pages which is IMHO a hassle. Centralizing state in a per user session context (+ eventually per search) would be leaner and more efficient (a la React Redux / Clojure). Hence I haven't touched to this project for a while, procrastinating this refactor.

One thing I managed to implement is a translation layer from Dynix search query language (which, if I understand correctly, is called the "Recall" syntax) to SQL (source).

Regarding using PICK or a PICK-like database, I'm not positive that it would be as performant as modern relational databases such as MariaDB or even SQLite. It would certainly more resource efficient though, as it was conceived to run on older hardware.

[p3r7]

Also, there seems to be a few libraries still running on Dynix.

I've also found this list, but most (if not all) of those instances seem to be down.

[cb88]

A few years ago I was able to telnet into a Dynix instance in Texas I think it was but its gone now.

I doubt there are still any Dynix systems online anywhere :/ , that said the librarytechnology.org system is wrong about a lot of librarys for instance it says ours ran Galaxy from 1993-2000 or so, but that isn't true, it ran Dynix for most of the 90s at the very least, this is corroborated by the local college listing also mentioning Dynix from 1994-2000 and I could check out books from any library in the county over Dialcat/DialPac. The local college was also running some accounting software on PrimeOS up untill early 2010s! That system was likely emulated though, I doubt the original Prime minicomputer was still in use.

I think they just assumed the prior system was Galaxy due to migrating to Polaris after Dynix.

This gives some history on how the system developed in my area, I definitely used Dynix during hte DAC era I seem to remember there being two separate systems for state level library queries and local county ones but my memory is very hazy about this as I was only about 8-10, book loans within county at least were completely free and would be mailed to the library and saved for me to pick up:
https://www.nccommunitycolleges.edu/sites/default/files/library-services/history-of-cclinc.pdf

[cb88]

Also might be possible to use ScarletDME .... since that is a Pick/Basic environment. Its not really developed but then... its not like Pick/Basic has changed in decades either. The owner of the git repo below appears to be active at least if not making making major changes.

https://github.com/geneb/ScarletDME

[p3r7]

Thanks for all this insight.

It can be frustratingly hard to find this kind of information.

I'm not 100% I'll go the multi-value DB / PICK OS routes given how few time I can dedicate to this side project, but will happily read more about those.