Contributing to OpenCog

[Mark Waser]

So . . . . let me weigh in . . . .

 

I’ve been a professional developer for 37+ years and am constantly using open source software for both work and personal endeavors. 

 

The *worst* thing for me is when the publicly available repo cannot be immediately compiled and run due to missing/incomplete instructions or errors.

 

While this is nearly unavoidable once in a blue moon, it is certainly unacceptable when it happens on a regular basis.  It shows an incredible lack of professionalism -- much less a tremendous lack of respect for the time and effort of others.  At work, it is a termination offense.

 

I’ve read this mailing list for years.  It seems that the repo being broken is far, far more common than not.

 

The fact that the project has so many moving pieces is seriously daunting.  The fact that you can’t even start analyzing it without repairing it is a total non-starter for me – and I suspect, most others, particularly professionals who *know* how software projects must be run to ensure progress and success.

 

I had great hopes for OpenCog once – but I seems that it has done nothing but spin in circles except for a few (outstanding) one-programmer pieces (that haven’t picked up the traction they deserve because they’ve been anchored down by the rest).  I don’t see that changing without a serious upgrade in the engineering management of the project.

[Ben Goertzel]

We haven't done a great job of keeping OpenCog friendly and usable as
a general opensource tool for the world at large, nor at making it
approachable to developers outside our close community...

Pretty much, a close group of people have been using OpenCog for
various applications and contributing to the generic OpenCog codebase
along the way.... The code has been kept open source and some effort
has been made to keep the code documented and to keep a wiki up to
date, but the project hasn't really been run or maintained in a way
oriented toward fostering a developer community.... This has not
been due to a lack of interest in fostering a developer community, but
just due to the core people involved being perennially swamped with
other stuff...

Indeed, nobody has been actively managing OpenCog as a general OSS
project. Folks have been maintaining the codebase and repos (Linas,
Amen, Nil, etc.) but some mess has been tolerated there so long as it
doesn't stop our applications from working...

I'm certainly not opposed to project management in general -- e.g. we
have just hired a new software project manager for Hanson Robotics;
and my new SingularityNET project has pretty structured project
management. It's just a fact that so far there has not been funding
to hire a good project manager for OpenCog as a general OSS project,
and nobody with the right mindset and skillset has emerged to do this
on a volunteer basis (and it's not a very easy job)...

-- ben

> --
> 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 opencog+u...@googlegroups.com.
> To post to this group, send email to ope...@googlegroups.com.
> Visit this group at https://groups.google.com/group/opencog.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/opencog/000001d33f68%248e9bf310%24abd3d930%24%40gmail.com.
> For more options, visit https://groups.google.com/d/optout.



--
Ben Goertzel, PhD
http://goertzel.org

"I am God! I am nothing, I'm play, I am freedom, I am life. I am the
boundary, I am the peak." -- Alexander Scriabin

[Curtis Michael Faith]

On Saturday, October 7, 2017 at 8:34:06 PM UTC+8, Mark Waser wrote:

The *worst* thing for me is when the publicly available repo cannot be immediately compiled and run due to missing/incomplete instructions or errors.

 

While this is nearly unavoidable once in a blue moon, it is certainly unacceptable when it happens on a regular basis.  It shows an incredible lack of professionalism -- much less a tremendous lack of respect for the time and effort of others.  At work, it is a termination offense.

 

I’ve read this mailing list for years.  It seems that the repo being broken is far, far more common than not.

I found this equally frustrating at the beginning. That is until, I asked the same questions and found the whys behind the reasons. Also, having worked a few months recently in the same room as most of the core OpenCog Hong Kong team I can say that Ben is aware of the shortcomings of OpenCog at any point in time, and the high-level steps required to make things right. We've been talking about some of these issues for several years. I've seen a rationale ongoing effort and several attempts that worked for a while and fizzled because there wasn't someone in those roles.

Building something hard is a game of stamina more than one of being right in all ways at the beginning. To keep going you need to keep the bills paid and people happy.

So you can look at OpenCog for what it is today, and has been... a research platform for a small cadre of AGI researchers working on world-class-hard problems that solves the basic architectural problems of AI suitable for hardcore programmers with some time on their hands... 

Or you can think about what OpenCog will be... after a transition to the next level up with production level quality, documentation, and product management in service to the larger world community building benevolent intelligence.

I've been choosing to jump in and help with the ongoing upgrade. It sounds like you have a lot of experience here. Want to help us build the team to take OpenCog to the next level? Point good people Ben's way and he'll recognize them...

There is no doubt we can improve the interfaces and greatly improve the experience for anyone approaching OpenCog anew. I've got some points on these issues I'll continue in the other part of this thread...

[Linas Vepstas]

On Sat, Oct 7, 2017 at 3:34 PM, Mark Waser <mark....@gmail.com> wrote:

 

I’ve read this mailing list for years.  It seems that the repo being broken is far, far more common than not.

Perhaps it seems that way, but its false.  In actual fact, over the course of the last ten years, the repo has been broken about a dozen times, never for more than a day or two, and of those breakages, perhaps only one has been reported on  the mailing list.

The reports you see on the mailing list are always "user errors". The one from two days ago is infuriatingly egregious.  The user apparently didn't know what gcc was. What can you do with people like that?  From their point of view, its all very broken, but that is not the reality.

The only viable audience for opencog, at this time, are developers with extensive knowledge & interest in AI.  College students, even grad students, encountering Linux and programming for the first time in their lives, and struggling with it, do not count.  A class or two in java programming does not making one a programmer.

Working with software is like playing the piano. Anyone can do it.  You just hit the keys, and sounds come out.

--linas

 

--

"The problem is not that artificial intelligence will get too smart and take over the world," computer scientist Pedro Domingos writes, "the problem is that it's too stupid and already has."