Re: averse to lisp in base?
Hi, Tomasz Rola wrote on Sun, Aug 29, 2021 at 08:21:03PM +0200: > On Sun, Aug 29, 2021 at 03:27:27AM +0200, mayur...@kathe.in wrote: >> Would the core team consider including a minimalist lisp in the base? >> e.g. http://t3x.org/klisp/index.html [...] > If I would want to propose any Lisp into the base system, I would > point at s9fes (Scheme 9 from Empty Space), also by Nils Holm. As a general rule of thumb: if some program * is maintained by a person who is not an OpenBSD developer, * and/or is maintained in a repository outside of cvs.openbsd.org, * and is not required for building or running anything contained in the base system, then it is usually better maintained in ports than in base. For users, typing "doas pkg_add my_loveliest_lisp" once is negligible effort, so it really doesn't matter for them. For developers, regularly synching a piece of third-party software into base is quite tedious and wastes considerable working time, whereas keeping a port up to date causes orders of magnitude less work. For example, synching to base is a royal pain for essential stuff like clang and Perl and drm. It causes non-trivial work even for smaller stuff like unbound and nsd. In several caaes, it was such a nightmare that developers just gave up in those cases - for example, just think of nginx and sqlite. In any case, this thread is wildly off-topic on tech@. If you have additional questions related to this topic, please start a new thread on misc@, or post a port or a port update to ports@. Yours, Ingo
Re: averse to lisp in base?
On Sun, Aug 29, 2021 at 03:27:27AM +0200, mayur...@kathe.in wrote: > Would the core team consider including a minimalist lisp in the base? e.g. > http://t3x.org/klisp/index.html Hi. I am not the one to decide about such things but I have two cents to drop. I have skimmed the manual page for klisp and I suppose it does not offer interface to system calls. In essence, it seems to be of same usefullness as a Lisp interpreter written in, say, awk. If I would want to propose any Lisp into the base system, I would point at s9fes (Scheme 9 from Empty Space), also by Nils Holm. It is bigger than klisp, but at least gives some Unix/POSIX functions to use in the scripts. Also, it is Scheme, so in theory one can at least try to make use of some freely available Scheme code from the net. Much more useful in a Unix-like environment, I would say. http://t3x.org/s9fes/index.html But I am not proposing - I have no idea how many people would find this thing helpful in their life. As far as I know, there is not many (it any) scripts in Scheme that cater to Unix users. If one would like to have it on base, it is as easy as procuring shar archive which takes care of compiling and installing this stuff when needed and keeping it on some pendrive. -- Regards, Tomasz Rola -- ** A C programmer asked whether computer had Buddha's nature. ** ** As the answer, master did "rm -rif" on the programmer's home** ** directory. And then the C programmer became enlightened... ** ** ** ** Tomasz Rola mailto:tomasz_r...@bigfoot.com **
averse to lisp in base?
Would the core team consider including a minimalist lisp in the base? e.g. http://t3x.org/klisp/index.html Going further, openbsd might need to have artificial intelligence oriented functionality for tasks such as systems management assistance. Instead of the new-fanged "adaptive machine learning" approach, a "logic-based artificial intelligence" approach would be more suited, typically for scenarios which require intent-oriented commonsense reasoning which are better resolved using the BDI model. The author of "kilolisp" (above) is okay with re-licensing his code and having a lisp system via an embeddable library might go a long way, though work would be required to turn "kilolisp" from a general interpreter core to an embeddable library. Why do I suggest "kilolisp"? Primarily because it's purely symbolic and currently requires only 64K of memory. Thanks.