[Zope] - [Zope] zip zap zope (long, rich ?) [Re: Zap]

Boris Borcic zorro@zipzap.ch
Mon, 18 Jan 1999 19:53:21 +0100


<STANDARD_DISCLAIMER>

This started as a short word prompted by the idea-announcement
of Zap, and ends up as a very long stream of half-baked and
not-quite-all topical ideas. Oh well, I guess I'll just have to
apologize to the non-readers, congratulate the readers, and
warn that I might repeat some of the imbedded ideas in
isolated and better presented form, depending on the feedback
this (doesn't) get (in this or other fora..hums).

</STANDARD_DISCLAIMER>

The more I think about Zope, the more I think it is best placed
to become "the" balance-tipping Open Software's Dynamics'
"killer" app.

(I name Open Software Dynamics the Zeitgeist in which
digicool's decision to opensoft itself, inscribes itself naturally)

I can dream about it in the small, I can dream about it in the big.

In my more enthousiastic and litterary moments, I can see
it as a cross between Roger Zelazny's Hellrides and The
Triumphal March of God's Feast Feeding Birds, straight
out of the Apocalypse, with our good Guido van Rossum
an onomatopoetic figurehead !

But in any case I need to find myself a niche as a source of
source code, if I want a legitimate claim as a significant member
of the Open Software Dynamics.

So i'll start dreaming in the "virtual small".

=========

First idea : Zope+Gadfly+Python 1.5.2+other_bits is really
the core of "Full Python".

Second idea : "Full Python" like for instance "Prolog",
is an artistic achievement opening the door to new *styles*
of programming.

Third Idea : "Full Python" should Serve Itself First, both
as a way to demonstrate itself, and as a way to govern
its development to as large a dominance that it will able
to achieve.

=========

So, to answer the matter of the Zap announcement :

The Zopinion project and the ZipZapZope project
===================================

The Zopinion project is to have Zope distributed standard,
with a possible setup to be used locally by private web
page authors maintaining contentful web page systems
using Zope, that later automatically unfold to the directory
of that end user's internet provider who's not supporting Zope.

Ideally, Zopinion could drive its own storing FTP sessions.

I believe I've already read that main idea of Zopinion,
stated in Zope's list.

The moral of Zopinion's name is that we think of it
as "a personal opinions server".

What's in Zopinion for Digicool, is that Zopinion will
push Zope's logo onto WWW sites faster than Zope
itself. Of course, Zopinion is far from exercizing all
the features of Zope, but such a notice should not
deter any real Open Software man.

What's in Zopinion for its intended users, is that
Zopinion is designed to help individuals with lots
of hairy-but-interconnected ideas and opinions, to
structure their output in such a way as to provide
lurkers with just the appropriate paragraphs and
bits.

Zopinion is thus the logic of web advertisement
technology given back, in the Open Software
Spirit, to the private ideas producer. Zopinion's
design criteria expresses the idea that mankind has
come to a special time, where the competition
for bandwidth and mindshare has become so
widespread and consistent that the best interest
of an intelligent and collaborative source of ideas,
like our Guido or Zope's Jim, has become to
*serve* his stream of designs with the best
dressings for others to copy, freely to the wide public,
rather than *hide* them for fear of seeing them stolen.

My personal use for Zopinion, is to maintain
Zorro's Own Bazaar of Condensing Vaporware (ZOBCV),
and my forthcoming "The Y programming language,
a short course".

[[More mundane Zopinion hype :

Zopinion could be augmented to take fully take over
and overflow Microsoft's Outlook Express, for reasonably
text and code oriented users... that is, manage mail,
news and mailing lists... I can see it running as a
permanent local proxy.]]

===

The ZipZapZope project is more ambitious and
"way over my head". ZipZapZope is (presumably digicool's)
solution to modularly managing Zope distributions,
and BaZaars.

ZipZapZope is to actual software as Zopinion to
condensing vaporware.

I envision two aspects/interfaces to ZipZapZope of
immediate relevance.

One would be a Zope-and-python-oriented hybrid of
Debian Linux RPM-FTP distribution and MS Office
CDROM installer, functioning over the Web. Ideally,
the background logic should automatically construct
an Up-to-standards Installer for (at least) the selected
items. Don't python libraries provide everything needed ?

Another would be, and I guess That One requires
user interface innovation, would be a facility for
turning, with least effort, mailing lists that are in
the state zope@zope.org is in, into "self-faqing"
branching archives and *persistent* mailthreads,
smoothly, with a mail-service logic partly serving
newbies with archived paragraphs.

Next Step : Zope's conquest of  CBZN, the Comprehensive
Bazaars over Zope Network. The delicate point with
That One is that we will start to need a facility for
(Zoped) Bazaars to Rate Each Other. This is a sensitive
matter because it pertains to Open Software's Dynamics
main alternative to common euro/dollar logic.

They are relatively tame perspectives and more
radically revolutionary ones. While it isn't always
the wisest, I have a tendency to contemplate the
more radically revolutionary ones front. It's not
always the wisest, because in many cases the
tamer perspectives are the right stepping stones
to the wild ones.

This said, designing for CBZN, I would determine
my inspiration by picking a design constraint that
pertains to a (yet) imaginary situation. A Kung-fu
gesture. I am really a Zelazny fan !

Keyword : snapshots or stripshots.

CBZN would "drive or participate" a "management"
of "its layer of internet" by mutual snapshot and
redistribution services that CBZN sites would provide
to ZipZapZope and Zopinion sites. The CBZN
main layer should be integrated to full ZipZapZope
distribution, I guess.

This design constraint brings vast amounts of
style to the design... it expresses a personal
taste, that I guess must be shared by many
Python and Open Software Lovers, that there
was, in the good old internet of the textual codes
days, something superior to the nowadays'
"multimediated internet and programming".

I mean 1) the mutual network relationship of CBZN sites
should be expressible in terms of this "snapshot" business,
 2) CBZN would specialize on *lossy* or *condensating*
snapshots, imposing constraints of *data*economy* to the
snapshots it takes, 3) CBZN would in particular "have in mind"
stripping snapshots of non-linguistic data, or at least, of binary,
image, or voice-like data that is not provided with a recognizable
or certified equivalent in written semantic (natural or artificial)
language.

3) Repeat : CBZN would be choosy on the way it mirrors data
automatically in just the right way, that makes Open Software
Sources Distributions it's notion of what best content is !

This suggest a couple more avenues.

One avenue is a new business, "imagating". Imagating
can be sold to big operators as a two-step advertisement
strategy, whose ultimate purpose it to permit the big operators
to assert their mindshare while showing they are inobstrusive.

"Imagation" is an *actually*free* service provided to CBZN-like
lossy Web Page snapshots producers and users : plug into
semantic descriptions of images contained in strip-shot
pages, relatively fitting images taken from Image Banks of
images *invisibly*watermarked* with some Big O's
watermark !

The second step comes after a while, when "Imagation"
has taken hold as an institution : Statistics will be made
that will show each Big O's Real Place in the World !

That's the advertisement we need to sell the Big O's !
Or we'll go to the Big O's competition !

Of course, "Imagation" logic can extend to other multimedia
snippets that just images, and, besides "retrofitting"
stripped pages, will of course serve Zopinion authors too
busy to waste their time scanning the images they'd
need to make their pages more sexy to readers.

(There could be a "reverse side" to imagation, the treatment
of illustrious images. Illustrious images, such as portraits of
illustrious persons, could be treated to a worldwide collaborative
network and friendly competition of "best legend design".

Yeah, this would fit the Triumphal Apocalyptic March view,
but I guess, in that case, it leaves me with a bit too much
of the Main Role to fullfill.

[Forthcoming : the YZ rules of blasphemy-classifications
 and the information Ramadan]
)

And of course, plotting packages able to turn a condensed
linguistic expression into a graphic automatically, will provide
basic imagation services.

The second mentionned avenue, is to secure Open Software's
Dynamic over names and identifiers. Mmh, while it would be too long
to detail independent justifications immediately, think of it as a
good idea to maintain a secured archive of Open Software Reference
identity, identifiers and litterate snippets library Registry, centered
on the works of Open Software's Dynamic's appropriate "conjugates 
set" of Guido and Jims as Figureheads and Implementations.

[One of the first starting tasks is of course for the Python
community to answer for itself the question : If Jims validly generalizes
Guido, who validly generalizes Jims ? We might say, Jims have
Guido-Number between 1 and 2, and we want to make a proper
census of people with Guido-Number between 1 and 2, "according
to us". For litterature on Guido-Number, see Erdos-Number.
Guido Himself is Unique with Guido-Number Zero.

For the Larger Communities the question will be rather : 
Who is like Guido ? Well, Larry, as matter of course, but then ?
Who is to Larry as Guido is to Guido's Friends ? Well, Tom and
Randall at least, but then ?

Well, let's not forget that Guido has a brother : Just !

And let's not forget that Guido's design has a preferred bedmate :
Tcl/Tk !

So, Just in a sense "competes" with Larry as a Proper Way to
Generalize Guido, just as Tcl/Tk "competes" with Perl as a
Proper Way to Generalize Python, for the purpose of designing
Our Roster Of Best Friends !

Maybe this allows me, the apparently unjustified freedom in
this forum, to cite another person, who has little to do with python
and even with programming, as an adequate "pseudo-clone" of
Guido, for the purpose of Our Roster, and for the purpose of the
quality of the Sampling that we are Dreaming Of for Our Roster,
and this person is Mr. John Baez, of sci.physics.research and
"This Week's Finds in Mathematical Physics Fame".

Below comes the paragraph that was written first in these brackets,
and of whom all the above 6 paragraphs are an explicit extension.

The Difficult Part of That Work is : Mmmh the problem is the above
selection has a programmation that is well-centered in the Open
Software's Dynamic's Core, but that it appears to drive towards a
form of sample that is formally over-biassed as an example of the
ideal sample of persons that we would really want to unite as
Authority.]

To put it another way. Whatever the community you consider,
you may contemplate it with the view that you are exploring the
sets of the set of the persons forming the community, and doing
your democratic citizen's task by comparing and establishing
preferences between such imaginary sets, as to whether there
are a good representative subset of their whole, if it came to
take decisions or to negociate in the name of their whole.

Claim 1 : while there unavoidably exists variance between
the conclusions various citizens can draw from such an "imaginary
experiment", there nevertheless exists, in normal cases, a strong
"common conclusions" component, especially if you do not
limit yourself to one community's universe, but maintain the view
that all communities share in vaster embedded universes of
intelligent beings. (Which means you are actually classifying
your sampling-reluniverse pairs of choice)

Claim 2 : The Guido-Larry connection and the "Open Software's
Dynamic" characterization of "Our Final Sample", as of -10
paragraphs, imho points the way to "building a perfect sample",
but for a limited while well characterized community, while at
the same time appearing as a perversely biassed sample if
we want to view it as a sample of the wider community we
would like to scan for Our Wider Sampled Roster Of Best
Friends and Friendships.

"Perversely Biassed", in the sense that the "obvious" manner
to "naively extend the sample", by "lowering its filtration standards",
will achieve the reverse of the "obvious" expectation : it will make
the bias worse instead of better !

This said, the Guido-Larry connection provides well-defined benefits
and guidelines to the trust we want to give to the Open Software
Dynamic. Or at least a good figure for something I really think needs
to be implemented : an Open Software's Semi-Central Authority
of Identity, Identifiers, and Snippets, with an eye on protecting
Open Software's choices of names from abusive greed, and
another on providing basic support to syncretic scripting language
programming, by allowing coders to "mean what they want" in
mixed syntax and/or mixed language standard library allusions.

Still, all the above doesn't do Justice to the Other Great Dream
(OGD) I had for Python, that was to offer to Guido the best equivalent
of what he brought to us, the Python coders, by gently demonstrating
that we would not mind if he changed his projects, and decided to
evolve Python 2, not as a redesign "done right and pure" of Python
1.x, but as an integration of Python 1.5, with its special facilities,
in the programming language Oz 2, a marvel by itself, but that may
be in need of Guido and Python to really make it to the life it
deserves.

For instance - and this is really from someone who has no idea of
the implementation issues - I have the vague impression that Python
is sufficiently flexible to mimic good portions of algorithms expressed
in Oz 2, so that Python could "espouse" Oz 2 with the bonus of
permitting (with some restrictions) Oz 2 code to refer to the Python
Modules, and reciprocally Oz 2 could be made to "espouse" Python
by devising a Python-mimicking surface syntax for itself.

Note that Oz 2 has something quite close to Tkinter, and should
be able to provide Python with all that it yet lacks in Sexpyness
and related program transformation aptitudes.

Oh, and a last word : securing revenues has been a long-lasting
problem and will recur soon, and while I hesitate to draw from
Zope, more than the resources that I already draw for free, it may
just be the case that it is sensible to mention this issue with
the above diarrhea in perspective.

Best Regards,

Boris Borcic
---
Paragraph contents copylefted B.Borcic 1999 with special requirement that
the following assertion be appropriately presented whenever the issue
arises :

"Open software, serving weapons of designfully murderous potential, is a
contradiction in terms"

Forthcoming : a thoughtful criticism of the limitations of the
Bazaar/Cathedral
duality.

More ADN  at www.geneva-link.ch/~zorro
(pre-alpha packaging and presentation)