[cap-talk] Potting the web-calculus in a paragraph

[Ian G]

Jed at Webstart wrote:
> At 05:22 PM 12/8/2005, Sandro Magi wrote:

>> It seems to me that technologies are generally second-class citizens 
>> next to brands. People buy into products: the Waterken server can do X 
>> and this other server (which also implements the web-calculus) is 
>> "compatible" with it, I can do ASP.NET programming in Microsoft.NET or 
>> Mono, etc.
>>
>>> As to whether it is effective even for communication to developers,
>>> well, I'm not sure there.  Ian is a developer and is put off by it.
>>
>>
>> Indeed, but I think he is more put-off by the lack of introductory 
>> material appealing to his viewpoint, rather than the name itself.
> 
> 
> I got the impression it was the name itself.  Specifically the 
> "Calculus" that seems a bit, well, stuffy.  Not a down
> to earth engineering term (remember, my background is mostly academic, 
> so I tend to lean in that direction).

Sure, but it wasn't really the name itself, a name
is just a name.  More, the choice of the name
was indicative of the core leaning - the notion
that it is an academically inspired project.  I
think we're all experienced enough to avoid getting
hung up on a name.

>> Initially, he just didn't get the "calculus" connection. When the 
>> connection was explained, he wasn't impressed.
> 
> 
> I think because (I did follow the discussion) the basic "Calculus" 
> notion being sold has morphed so many
> times over the years and shown up in so many forms that calling it a 
> "Calculus" and harkening back
> to the original Lambda Calculus is a bit like trying to sell something 
> by saying it can do anything
> a Turing machine can do and is therefore really general.

Right, certainly 'calculus' has either no meaning
or pretty vague meaning.  And, when poking at the
real meaning behind the name, what came out was ...
a message of "if you only understood FP and lambda
calculus, it would all become clear..."  Not even
"we took the X from the tradition of Y and stuck
it in here."

So ivory tower is a good characterisation.

Especially, this can be seen in comparison to the
widewords.com site;  which does something similar
without any notion whatsoever at academic foundation.

Which feels better?  Easily, widewords, because we
don't have to wade through stuff we don't and cannot
and will never understand.

> People, not 
> just users but also engineering people,
> really don't care about that stuff and are somewhat put off by it.  (Ian?)

People (both users and engineers, yes) are put off
by it, but the reasons are not just politeness, there
are some economic underpinnings.  I'll try and write
about them, but be aware that this is a vexing topic
and I've puzzled for years now why the caps school
so impressively "puts me off."

The first thing to understand is that many of these
things have been tried before, and will be tried
again, all with minor variations.  E.g., Java is
a minor variation on caps, or is it the other way
around?  So anyone who is looking at this stuff is
looking to relate it to what has been done before,
for both self-learning purposes and for analysis
of the project.

We are used to commerce people
like Pelle ignoring their competitors and waving
their hands like windmills and talking hype in an
effort to get people to use the tech, so we expect
commercial people to give us no help in finding out
the alternates.  Helping us is not their job, selling
us is.  But no such "pass" is extended to academic
world, instead, we actually expect to find lists of
other work and careful explanations of how it relates
and builds.

So when the caps community talks vertically about their
roots into foundations and theory, and does not reach
horizontally out to similar paths, it quickly takes
on the insulated and out-of-touch appearance that
we call the ivory tower.  This is a persistant
problem with all of the caps field, but is perhaps
more blatant in the web-calculus project.

When we (or I) say we are put off by it, what we
are saying is that in our experience, the lack of
integration and lack of reach means that we can
make an economic judgement - it's too insular to
learn and work with the alternate experiences out
there, so it will be overtaken by some other faster
learner.  Hence, no point in adopting it, but there
remains point in understanding the wheres and whys
of the model, as long as it is not to hard to enter.

(This is at the core of why I spend time on this
list.  It's a competitor list to my stuff, but I
try to relate it and document it on the websites,
so as to meet the academic expectation - horizontal
reach as well as vertical experience.)

Which is all to say, this is not just about the
name.  It's about the caps culture, and how the
web-calculus project epitomised that culture.

> I'm content with "Secret URL" or SURL at this point.  I'm bouncing those 
> possibilities off
> non-technical people to see what sorts of reactions I get.

(excellent!)

> Here I'd like to hear from Pelle.  I'll contact him directly if need be.

(you should.  Any serious engineer has no time for
these threads ;-)

>>> There again I think the tension between
>>> a more academic oriented term and what one might call a more
>>> engineering oriented term is legitimate.
>>
>>
>> I think having the formalism as a foundation is important.
> 
> 
> We agree there.

I am not so sure.  What is the formalism that Pelle used?

(I don't want to push Pelle more than as a current example,
which you are all familiar with...  I personally am not
sure he's hit on the winner yet, but that's not relevant
here.  Another like system is https://loom.cc/gold/ and
another approach that is highly comparable is the Ricardian
contract which is a hash of a document.)

I can suggest Pelle's style - he follows a market process
and grabs whatever tools he needs to do that.  he happens
to have grabbed a caps approach but that's only because
he looked for the simplest way to do a doc.  His background
is in crypto and payment systems and the like, and there
we've all known that the easiest way to build these things
is with a big random number, or a hash of an object.  So
we are all squeezing around the hash or the private key
concept to get the maximum flexibility out of a net access,
we've been doing this for years.

And what we all ride on is experience - sometimes each
others, sometimes our own - and our understanding of how
these things interact up and down the software and user
stacks.

Compared to that experience, the formalism of caps looks
like a rather small part - I count it as one alternative
design in one layer in my 7 layer model of FC (rights,
title supplied thanks to MarkM);  it might conceivably
inform the next lowest layer (Software engineering) but
that's as far as it could go.  There's a huge space that
caps simply doesn't address.

> And the names used need to appeal to engineers, not researchers.  That's
> my only point regarding the "Web Calculus".  That level of name is 
> really not a
> concern to me.  I hope Tyler is able to sell it.  Selling a name with 
> that sort of
> heritage seems to me good for our profession - even if it does rub a few 
> people
> the wrong way in the process ;-)

Sure, the name isn't the big thing.  It's how you move
forward and sell it outside an audience that doesn't
understand the name.  Talking about the roots of the name
isn't going to help that, especially if your eyes light
up and reveal that you actually believe in the foundations
more than the user.  Then you are asking the person to join
your religion, which works spectacularly well sometimes,
but most times is better left off the agenda.

iang