Thanks Stephen,
On Tue, 25 Aug 2026, Stephen J. Turnbull wrote:
(although I hope you can move on to Mailman 3 if you're not already
there!)
my big issue about moving on is the apparent necessity of otherwise, for
me, COMPLETELY unnecessary use of containerization software I see
absolutely ZERO need for, and don't want on my boxes for these very simple
reasons: 1) the size of the necessary software (disk space); 2) the target
vectors possible for being cracked (unknowable), and; 3) it slows my
system down on doing what jobs it actually has to do and instead deal with
"virtualization" noise. ... I like to keep it as close to the hardware as
possible. But then I'm probably the only one on this list who has actually
written an entire real-time, multi-tasking OS solo... and have it still in
production for multiple decades. ... If I don't need it, WHY?!
However, I'm also a scientist: Give me a good argument to move on and I
will, I've just never heard one - or even an attempt at one - so far. It's
always the lame argument that the newer is always better which us
old-farts know is errant.
(Note to newbies; "hacking" is a positive thing, "cracking" is a negative
thing and what the media ignorantly conflates with hacking. I'm a hacker
and NOT a cracker! The term "cracking" is based on safe-cracking, as in
vaults with valuables, not some racial term.)
And sometimes it's us, so stay humble. ;-)
If I'm ever not, I don't mind being called out but do appreciate if it's
done privately! ;-) ... Uh... given my digression below, call me out
publicly if you want.
Then you mention this! -groan-
> the worst offender of which is Google in my experience.
Businesses that provide free stuff on an ongoing basis are invariably
antisocial, and untouchable because people love the free stuff. IIRC
Google invented the model of free stuff + surveillance capitalism, but
half of Silicon Valley piled right in. Somebody would have invented
it soon enough.
We just have to remind ourselves that Silicon Valley is a nest of
stochastic horrors, and invest in perimeter defense.
Don't I know it! I'm one of the ... uh ... asshats who ENABLED Google and
all this BS!
WARNING! Digression Alert!
... One day (Turing Award winner) Michael Stonebraker came into my lab
(380 Soda Hall, UC Berkeley - before it was reconfigured) with Sergey Brin
and Larry Page in tow, two recent Stanford grad students that had come and
asked for the dog-and-poney-show (at least months if not a year before
Google was founded).
Why? I had built the world's first (publicly known) terabyte-plus (maybe
1.4T) sized database. It was a geospatial database with graphical
interface with sub-second response time to browse the entire dataset,
though of course to access the larger corpus would take time. But not only
this, I had built the worlds first functional grid computing system
without even knowing that as we did it Ian Foster and company had coined
the term "grid computing."
Similarly to Waz and Jobs, Page, on learning this was technical exited the
room nearly immediately but Brin spent perhaps 7 hours with me, probed
and asked questions and I told him the how and why - it's much of what
follows.
Most easily notably: Just plot map contents via lat and long on your
raster display without worrying about projections because at this scale it
didn't matter. And, you "pre-cook" the layers of zoom - you don't
calculate them in the fly. This isn't for changing detail, just the
background map, and I told him how we got our data from the USGS and,
oddly enough the CIA which had the freely downloadable political
boundaries - which, by the way, proved to be "off to the right," as ALL
their data was provably incorrect, shifted a bit east by enough to matter
for our Earth Science applications. And there were also publicly available
datasets for rivers and streams, lakes, and transportation, all easily
downloadable from the USG and, increasingly, the rest of the world - the
latter largely based on satellite data that I, personally, focused on.
This background map is all Google Maps is anyway, though the derived
"directions" showing traffic congestion is an implementation of our
techniques, too.
Less visible but perhaps more importantly OUR work was a fully distributed
computing environment, our grid, which Ian Foster, ignorant of our work at
the time, later admitted we were first to functionally implement "grid
computing." With the modern developments it's important to know our work
became modern cloud computing in the hands of others, though they flipped
our paradigm. WE had fully distributed data AND computing - along
egalitarian thinking - and the cloud used our ideas to flip the idea
backwards and suck all the data to their own environ and just push client
computing to, well, clients. Instead of a fully distributed data model we
now have a concentrated one and instead of many distributed data centers
we have a few huge ones. (Don't give up your own computing abilities if
you want to avoid being at their mercy.)
Most notably, however, this concept was crafted by Brin and company into
Hadoop, though you'll never read a single sentence of credit for Hadoop,
Google Maps, or anything else they used contributed by our work at
Berkeley. (The public knows about Xerox PARC's contributions to computing
but not ours; Apple admitted this, Brin and Page didn't.) And for those
that don't know, Hadoop is foundational to all modern search engines and
has even been instrumental in modern "AI." (I put "AI" in quotes to remind
people "AI" is a great marketing tool because it IMPLIES there's
intelligence present when there isn't, not even artificial, it's a
facsimile.)
On that latter point I feel we had impact there, too, because we created
the modern GIS systems (I could digress on that but won't) by fashioning
the already-created-by-others R-Tree indexing system for databases - it
does N-Dimensional indexes - and we plugged it in to Postgres, done by
Paul Brown on my team but Postgres was at the time a Stonebraker-adjacent
team and was at that moment creating what became PostgresSql, the last UCB
release of Postgres. We, of course, applied it to Geospatial. And, we
talked publicly about how our system learns, though our version of
computer learning used more intelligence than is used today, it just takes
longer...
Later many "AI" researchers came to our lab for a similar
dog-and-poney-show dialogue asking if we'd created a "neural net," as many
"AI" research teams did at the time. Instead I cited our R-tree adaptation
and said point blank that when you have a data processing need for some
form of index, a pointer, to what's needed that doesn't already exist you
invent the pointer for the type of data you have. ... This is NOT to say
we invented "AI" vectors but the multi-dimensionality of modern vector
space was right there in our originating use-case with R-Trees and the
impetus to create your own was also a kick-in-the-pants in the right
direction - exactly how things progressed later. When you have an
explicitly N-dimensional indexing space already extant... the ideas to do
modern "AI" were right there and we pointed the way.
Before signing off I feel compelled to clarify:
I had a team funded at 50 people for 7 years, Michael Stonebraker was the
Prof overseeing the team but did zero work on it - that was my job as the
lead scientist ("investigator") - and the team was never bigger than 47,
mostly less, but nearly all were our Earth Scientists (all the major
disciplines) working on our primary goal: We were the very first well
funded team to take on the goal of a proper multi-disciplinary study of
global climate change driven by global warming. The core of us computer
scientists was TINY and was really just three; me, Paul Brown and Keith
Sklower, with that order of precedence in intellectual contribution,
though Sklower did a fantastic job of a multi-dimensional toolset that let
us reach into a big blob of binary data and pull out whatever set of data
you wanted extraordinarily quickly and his contributions were very
significant.
We were originally funded at 5 years, our successes brought us to 7 and I
then got my own follow-on grants - without even UCB or Stonebraker being
cited in the grants (which proved to be a big deal).
Because I wasn't an academic (brought in from industry), we didn't get
published as Stonebraker was on his way out, to MIT, so most nobody knows
all this, so thanks for tolerating my sharing it.
Anyone who wants can contact me directly - challenge whatever, ask about
whatever... But DO note the two different e-addresses as noted below.
Richard
--
Richard Troy, Chief Scientist
Science Tools Corporation
510-717-6942
[email protected], http://ScienceTools.com/
[email protected] - as subscribed here
I pledge allegiance to We, The People, to mutual peace and harmony, and
to the natural Earth and biosphere upon which We, The People, depend.
------------------------------------------------------
Mailman-Users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
https://mail.python.org/mailman3/lists/mailman-users.python.org/
Mailman FAQ: http://wiki.list.org/x/AgA3
Security Policy: http://wiki.list.org/x/QIA9
Searchable Archives: https://www.mail-archive.com/[email protected]/
https://mail.python.org/archives/list/[email protected]/
Member address: [email protected]
