Rust will be great. Especially for the less-frequently-modified components like the atomspace.
I don't know if it would fit the bill as well for the other more volatile components. Opencog is not distributed as a binary and is rapidly modified. The primary development seems to be in the source code going in, and changes to that require lengthy recompile times. Making humans wait wastes their time and development potential. By having a highly-readable source code, we can make this development more rapid from development, experimentation, and execution perspectives. Julia can be static compiled if needed, it's not efficient from a file size perspective, but it's just as fast and I think rapid development and readable code offsets that easily, especially for developers just joining the project. Julia has phenomenal GPU support. It can leverage GPU memory and processors easily using native data structures. Allowing us to very easily scale out into high parallelization. Both for storage and processing. Julia's images library shows great potential for a future ML vision system as it can store entire images as a vector data structure, allowing us to apply very flexible analytics to the input information. ML libraries have already been built in Julia. https://github.com/FluxML/Flux.jl is the first that comes to mind, but others certainly exist. I don't want to be in denial. Fantastic code in rust would probably be more distributable and efficient than Julia, and if opencog ever went to a "release" type system it would be a great idea for all the components, but from my observations it seems Atomspace 2 will be the main "OK now we're serious" development area. It seems likely to me there will be an atomspace 3 once the science is perfected, but if rapid development is a priority for this stage of the project, it might be worth an evaluation. Can't wait to see the next evolution of Opencog! On Wednesday, September 16, 2020 at 7:28:02 PM UTC-4 Ben Goertzel wrote: > We have been seriously considering Rust for new distributed Atomspace > code, assuming the latter gets written... > > On Wed, Sep 16, 2020 at 4:16 PM Noah Bliss <[email protected]> wrote: > > > > To this end, has any thought been given to other implementation > languages? Julialang seems to have a very strong userbase in the research > field, is easy to pick up like Python, JIT so long compile times are > generally a non-issue, and it executes with near-C levels of performance. > > > > Just my 0.02, curious to see how this discussion develops. > > > > On Tuesday, September 15, 2020 at 8:51:52 PM UTC-4 Ben Goertzel wrote: > >> > >> Sorry to hear about your logistical hassles -- but good to hear from > >> you and know you're still around...! > >> > >> We are doing a from-the-round rethink and redesign ... whether this > >> leads to a from-the-ground rewrite or not is still an open question... > >> > >> As for the name of the hypothetical in-progress new version it's > >> "OpenCog Hyperon" for now... I am reserving "OpenCog Tachyon" for the > >> future closed-timelike-loop based version ;) > >> > >> > >> > >> > >> On Tue, Sep 15, 2020 at 4:47 PM Curtis Michael Faith > >> <[email protected]> wrote: > >> > > >> > The name and project need a ground up rewrite using lessons that > Kenny Kasajian knows better than anyone. I am trying to get Kenny to help > me build a team from the OpenCog project but I've been busy with trying to > stay alive as I lost my wallet in Manchester, NH and IDs and haven't been > able to open a new bank account since I left Portland 10 months ago. I have > checks piling up I can't cash and had to really get clevver to live... > >> > > >> > On Tue, Sep 15, 2020 at 7:43 PM Curtis Michael Faith < > [email protected]> wrote: > >> >> > >> >> So... the Open Cog project is run in academia and not really good > software land. AI is too specialized and HUGE you can't expect someone who > is successful to have time to do it right. > >> >> > >> >> I suggest you forget the name OpenCog. Forget the code. Forget Ben's > first prototypes. Think Think think > >> >> > >> >> we really all need to sit down together someplace in Europe or > Korea... perhaps > >> >> > >> >> On Mon, Sep 7, 2020 at 1:06 PM Ben Goertzel <[email protected]> > wrote: > >> >>> > >> >>> Hi all, > >> >>> > >> >>> If anyone wants to dig deeper into recent thinking by myself, Alexey > >> >>> Potapov and multiple colleagues on an in-progress (still very early > >> >>> stage) new OpenCog version, OpenCog Hyperon, we have put a bunch of > >> >>> docs on the wiki linked from here > >> >>> > >> >>> https://wiki.opencog.org/w/Hyperon#OpenCog_Hyperon > >> >>> > >> >>> Bear in mind these are mostly early-stage raw discussion documents > not > >> >>> refined specs or anything like that. We are sharing aspects of our > >> >>> thought-process-so-far with a view toward soliciting feedback and > >> >>> potentially participation... > >> >>> > >> >>> thanks > >> >>> ben > >> >>> > >> >>> -- > >> >>> Ben Goertzel, PhD > >> >>> http://goertzel.org > >> >>> > >> >>> “The only people for me are the mad ones, the ones who are mad to > >> >>> live, mad to talk, mad to be saved, desirous of everything at the > same > >> >>> time, the ones who never yawn or say a commonplace thing, but burn, > >> >>> burn, burn like fabulous yellow roman candles exploding like spiders > >> >>> across the stars.” -- Jack Kerouac > >> >>> > >> >>> -- > >> >>> You received this message because you are subscribed to the Google > Groups "opencog" group. > >> >>> To unsubscribe from this group and stop receiving emails from it, > send an email to [email protected]. > >> >>> To view this discussion on the web visit > https://groups.google.com/d/msgid/opencog/CACYTDBfrZjp7gE4qE7fUF0fLK%3DFvP3nM3hnrUA%2Be8eXTx8pJnQ%40mail.gmail.com > . > >> > > >> > -- > >> > You received this message because you are subscribed to the Google > Groups "opencog" group. > >> > To unsubscribe from this group and stop receiving emails from it, > send an email to [email protected]. > >> > To view this discussion on the web visit > https://groups.google.com/d/msgid/opencog/CAJzHpFr1Wjjvatktqq519aOarD05Mwt4kxv_JZgNGQt_gBeu2Q%40mail.gmail.com > . > >> > >> > >> > >> -- > >> Ben Goertzel, PhD > >> http://goertzel.org > >> > >> “The only people for me are the mad ones, the ones who are mad to > >> live, mad to talk, mad to be saved, desirous of everything at the same > >> time, the ones who never yawn or say a commonplace thing, but burn, > >> burn, burn like fabulous yellow roman candles exploding like spiders > >> across the stars.” -- Jack Kerouac > > > > -- > > You received this message because you are subscribed to the Google > Groups "opencog" group. > > To unsubscribe from this group and stop receiving emails from it, send > an email to [email protected]. > > To view this discussion on the web visit > https://groups.google.com/d/msgid/opencog/6517ab72-dcc2-4e3d-9403-6298458ae84en%40googlegroups.com > . > > > > -- > Ben Goertzel, PhD > http://goertzel.org > > “The only people for me are the mad ones, the ones who are mad to > live, mad to talk, mad to be saved, desirous of everything at the same > time, the ones who never yawn or say a commonplace thing, but burn, > burn, burn like fabulous yellow roman candles exploding like spiders > across the stars.” -- Jack Kerouac > -- You received this message because you are subscribed to the Google Groups "opencog" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/52c81a33-3e42-4ef4-879d-45846577658fn%40googlegroups.com.
