gmane.comp.lang.smalltalk.fonc

Power consumption, dynamically linked vs bytecodes

David Girle — 2013-05-07T23:56:13

Some on this list may find this article interesting:

http://dunkels.com/adam/dunkels06runtime.pdf

concerning power consumption on tiny compute nodes.

david

Re: enchantMOON

Yoshiki Ohshima — 2013-04-26T02:40:05

Yup.  Certainly your opinion goes along the conventional wisdom.  (It
is not Scratch but their tile-based scripting system inspired by
Scratch.)

I have my own share of doubts and wonder how much improvement they
manage to put since January, when we saw a prototype working.  But it
is commendable that they make hardware and customized OS and drivers
just to support a single idea they want, which is to have a good
response for hand drawing and a customizable platform by the end-user.
 An they did manage to start selling it.

I'd say "let us see".

On Thu, Apr 25, 2013 at 4:47 PM, Josh Grams <josh-hJ4xZT2gPrxBDgjK7y7TUQ< at >public.gmane.org> wrote:

Re: enchantMOON

Josh Grams — 2013-04-25T23:47:44

Is there anything interesting about that other than that it appears to
have Scratch inside?  There doesn't seem to be anything on their website
other than a video, and from the video, the user interface looks to be
horrifyingly bad.

- It starts up with what looks like a moon waxing, but when it gets to
  be a full moon, you find out that it's *not* actually done loading,
  it's just displaying a moon spinning.  If you have a progress bar, it
  should either represent a fraction of progress or it should just give
  an indicator that the computer is busy and hasn't crashed and should
  very clearly *not* look like a fraction of anything.

- It seems to have a lot of interface glitz which just slows things down
  and serves no useful purpose whatever.  If I click a button, I don't
  want to watch a little explosion of growing dots, I just want it to go
  to the next screen.  It might be cute the first eight times, but after
  that it's just annoying.  For serious use, half a second to wait for
  some stupid a

Re: Report Card

Yoshiki Ohshima — 2013-04-25T18:30:21

Making our code public would have been ideal.  The problem was that we
don't know how the system should look like, and things have gone so
many iterations.  In a way, I regret that we couldn't do a good job
making code public, but in other ways, I'm glad we didn't as we did
not have to feel bad about letting people see what we already
abandoned.  (It did happen with IdSt, for example.)

Incidentally, we have a new research note up on our web site.  And the
"annuall" and final report is very close to be released.

(Speaking of new hardware, enchantMOON has gotten to the "pre-order"
stage.  http://enchantmoon.com/)

On Tue, Apr 23, 2013 at 10:05 PM, David Girle <davidgirle-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:



--

Re: “Turtles all the way down”: amazing 64KB demoscene product

Ken 'classmaker' Ritchie — 2013-04-25T07:21:56

Phenomenal technical craftsmanship and remarkably creative teamwork!

Thank you for spotlighting this gem!

Cheers!
--Ken ;-)

Ken Ritchie (Atlanta, GA)


On Apr 24, 2013, at 22:00, Josh Gargus <josh-AKL4+GqkIg4< at >public.gmane.org> wrote:

Re: Report Card

David Girle — 2013-04-24T05:05:13

I cannot address "report card".  I have greatly enjoyed the papers and
respect the aspiration to deliver a system in ~20kLOC by leveraging
algorithms and Moore's law.  A system, with size and elegance, that folks
could get their head around.

Your point on discussion of what has already been published is well made
... for instance I have not seen a lot of posts on people's experiences
with say Maru.  In my own case, I had a lot of learning (as I have no Lisp
experience) and little time (and will not, until my business sells).  Also,
I was hesitant to engage on a code base that was likely well behind what
was privately available, particularly for things like sockets/files.  Ian
was kind enough to respond to a couple of requests.

My hope is that VPRI will publish their report and ideally some artifact
code.  Having written a small system with an mbed target and dart server,
dart/SVG browser client, it is clear how unpleasant working with the mbed
device in C/C++ is compared to a modern language and accompanyi

Re: Why Mind Uploading could be horrible

Tristan Slominski — 2013-04-23T17:47:55


I agree. And that seems to be a subject of future research as well, the
whole "engineering to get stuff within tolerance" side of things. :D

What I find interesting is if, instead of replacing a part, we want to add
parts. Extra sensory module, extra neocortex-type module, etc. Then there
seems to be some engineering involved to determine when the Conscious Me
assimilates the module and "spreads itself out" over the assimilated
module.



On Tue, Apr 23, 2013 at 12:42 PM, John Nilsson <john-w3tKOdPXMGimg8Ev75riOw< at >public.gmane.org> wrote:

Re: Why Mind Uploading could be horrible

John Nilsson — 2013-04-23T17:42:31

It's not so much if a copy is the same as the real thing, but rather how do
you define the difference between an all-at-once copy with a simultaneous
destruction of the original and a piece by piece replacement of the parts?
It seems to me they only differ by the size of the part replaced.
BR
John
 Den 23 apr 2013 18:13 skrev "Tristan Slominski" <
tristan.slominski-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org>:

Re: Why Mind Uploading could be horrible

Tristan Slominski — 2013-04-23T16:13:06

With great trepidation, I will try to keep this to computing :D

It may revolve around the meaning of "uploading", but my problem with the
uploading approach, is that it makes a copy. Whether a copy is the same as
the real thing I feel is beyond the scope of a computing discussion in this
particular sense. I assert, that I am not interested in a copy of Me (in
legal style, I will use capitals for defined terms).

The next thing is the definition of Me. For the purpose of this, Me is
defined as the pattern of interaction of physical processes that happens
within the volume bound by my skin. I will further refine to a concept of
Sensory Me, which I will define as the pattern of interaction of physical
processes that happens within my nervous system. I will further refine to a
concept of the Conscious Me, which I will define as the "pattern of
interaction" from the definition of Sensory Me, and it is separate from the
"physical processes" of the same.

With the definition of Conscious Me in place, what I am int

Re: Why Mind Uploading could be horrible

Loup Vaillant-David — 2013-04-23T15:11:08

Personally, I can think of 2 objections:

 1. It may turn out that mind uploading doesn't actually transfer your
    mind in a new environment, but actually makes a *copy* of you,
    which will behave the same, but isn't actually you.  From the
    outside, it would make virtually no difference, but from the
    inside, you wouldn't get to live in the Matrix.

 2. There's those cool things called "privacy", and "free will" that
    can get seriously compromised if anyone but a saint ever get root
    access to the Matrix you live in.  And we have plenty of reasons
    to abuse such a system.  Like:

    - Boost productivity with happy slaves.  Just copy your best
      slaves, and kill the rest.  Or make them work 24/7 by killing
      them every 8 hours, and restarting a saved state. (I got the
      idea from Robin Hanson.)

      Combined with point (1), this is a killer: we will probably get
      to a point where meatbags are not competitive enough to feed
      themselves.  So, everyone dies soon, an

Re: Separating computation from the machine

Eugen Leitl — 2013-04-23T14:01:20


I'm curious why you think that it's wrong (?) and horrible.

Re: Report Card

Yoshiki Ohshima — 2013-04-22T17:04:14

On Fri, Apr 19, 2013 at 1:51 PM, Casey Ransberger
<casey.obrien.r-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

"Burden" is not the word I'd use, but as everybody would agree, the
S/N ratio has not been too high.  I personal think that we, the folks
at VPRI, need to take a bit of blame for that.  We could have written
more regularly what we are doing to steer the conversation.  (But
maybe using words like "foundation" and "new" in the title of the
mailing list did not help us to have reasonable conversations^^;)  At
the same time, we did publish research notes and memos now and then,
but very, very, little about them were brought up (Casey, you did
mention some of it, I appreciate it!); some questions and discussions
indicated that some people commented on our work without reading the
original proposal or annual reports.  That was slightly unfortunate.


I learned a lot, too.




--

Re: Universal language and system programming

John Carlson — 2013-04-22T16:47:57

I have this idea that impredicative systems create little universes that
exist separate from the larger universe, including recursive universes,
impenetrable universes et al.  But then that leads me to self-enveloping
universes.

"This moment contains all moments" -- CS Lewis
On Apr 22, 2013 4:33 AM, "Chris Warburton" <chriswarbo-gM/Ye1E23mwN+BqQ9rBEUg< at >public.gmane.org>
wrote:

Re: Report Card

Edward Wohlman — 2013-04-22T15:33:05

I'm a lurker, and an Alan Kay fanboi with vague ambitions of entering the
PLT/software engineering sphere of academia one day, and this discussion
has been truly fascinating.

I'm very glad there are people out there tackling these sorts of
conceptual problems, as so much of the industry seems to be so
wrapped up in the M$/Unix fanboi wars or milking the enterprise
using 30 year old technology, or the general public with the
latest social media web bubble.

To everyone out there rethinking the basics, please keep up the good work.
May we one day get to grip with the arch and start being able to construct
actual grand cathedrals rather than shaky pyramids!

Edward

On Mon, Apr 22, 2013 at 3:33 PM, Paul Odle <misodle-/E1597aS9LQAvxtiuMwx3w< at >public.gmane.org> wrote:

Re: Report Card

Paul Odle — 2013-04-22T14:33:36

Casey,

I agree with your sentiments here.

Just a lurker.

I have learned a lot.


I was motivated to attempt writing a compiler in OMETA for a pet project of mine.

It has all been very interesting and very enjoyable. Even the trolls were interesting and funny at times.

Waiting for the final VPRI report. And a binary build of FRANK maybe?

Thanks VPRI folks and all of you posters! 


Regards,
Paul Odle


________________________________
 From: Casey Ransberger <casey.obrien.r-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org>
To: Fundamentals of New Computing <fonc-uVco7kAcSAQ< at >public.gmane.org> 
Sent: Friday, April 19, 2013 3:51 PM
Subject: [fonc] Report Card
 

I wanted to send this message out after the final status report, but since that's indefinitely delayed (keep going!) I'm just going to do it now. 

Easy question: has keeping this dialogue open been useful to the folks at VPRI, or has it been more of a burden than anything else?

I can definitely say that it's been very good for me, in that I learned a hel

Re: Universal language and system programming

Chris Warburton — 2013-04-22T09:33:39


Self-referential logics are known as "impredicative"[1], but surely any
Universal programming language is a Universal language?

Having a Universal language only guarantees that we can represent every
(computable) thing; it tells us nothing about how difficult it is to
construct them[1].

Plus we can only ever construct models of things; no system can describe
pointers completely (for example), since we can write Goedel sentences
which involve pointers; a trivial example would be the halting problem
for pointer-manipulating programs.

[1] http://en.wikipedia.org/wiki/Impredicativity
[2] http://en.wikipedia.org/wiki/Turing_tarpit

Regards,
Chris

Re: Use case for graphical problem oriented widgets (POW,DSW)

John Nilsson — 2013-04-21T08:30:05

I know those do, and I would expect any application claiming to be a
spreadsheet to have that.
Google spreadsheet actually have an interesting hybrid with its "continue"
function that is automatically inserted to expand array results over a
range of cells.
BR
John
Den 21 apr 2013 10:00 skrev "John Carlson" <yottzumm-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org>:

Re: Use case for graphical problem oriented widgets (POW,DSW)

John Carlson — 2013-04-21T07:59:58

Excel is indeed interesting.  Can google docs/libreoffice/OpenOffice do
something similar?
On Apr 21, 2013 2:52 AM, "John Nilsson" <john-w3tKOdPXMGimg8Ev75riOw< at >public.gmane.org> wrote:

Re: Use case for graphical problem oriented widgets (POW,DSW)

John Nilsson — 2013-04-21T07:45:45

I think Excel is interesting in this regard. The common idiom in excel is
to employ user triggered code generation by fill formula to adapt the ui to
dynamically resized collections.
Not exactly automatic, but is it sufficiently trivial to group with or
trivial ui operations such as scroll or resize window that might also be
needed to not interfere with the display of the dynamic collection?
BR
John
Den 21 apr 2013 07:59 skrev "John Carlson" <yottzumm-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org>:

Re: Use case for graphical problem oriented widgets (POW,DSW)

John Carlson — 2013-04-21T05:59:04

If you want a more complex use case,  create a loop 10 times around the
collection add loop to insert a calculator into the collection.
On Apr 21, 2013 12:48 AM, "John Carlson" <yottzumm-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Use case for graphical problem oriented widgets (POW, DSW)

John Carlson — 2013-04-21T05:48:06

Here's a semipractical use case: add 1 to the display in each of a dynamic
collection of calculators (math domain widgets).  What can do this as
end-user programming?  It's fairly obvious that a textual language can do
this.  Can any graphical ones?  Can something like lively kernel do this by
demonstration?  How about excel?  With a dynamic collection?  What will
work on android jelly bean?  I'm away from my desktop right now.
On Apr 21, 2013 12:22 AM, "John Carlson" <yottzumm-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Looking for systems like this I found app-inventor activity starter on my
phone.  Has anyone tried this?
On Apr 21, 2013 12:14 AM, "John Carlson" <yottzumm-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote: