From: marcs <marcs@skyhunter.com>
Replying To: Tyler Close <tyler@waterken.com>
Date: Sun, 8 Dec 2002 15:38:33 -0700
Subject: Re: [e-lang] Naming Capability Systems

> The main point of this email will be to convince people to slow
> down and carefully consider their reasoning. Marc Stiegler has
> presented the list with a 4 day deadline for renaming a field with
> a 36 year history. He is seeking to do this through rough
> consensus on a mailing list. This is clearly impetuous.

I would agree that it was impetuous if I thought I was renaming the field in 4 
days. There are a couple of ways in which this is not true.  First, I don't 
view it as "renaming the field". I view it as drawing a crucial distinction 
inside the field, a crucial distinction that has been needed ever since 
ambient authority capabilities succeeded in corrupting the majority of the 
audience. markm tyler 

Secondly, just because I start using a bit of jargon (oh no! more jargon!  :-) 
in my marketing pitches does not by itself constitute a successful 
introduction of a naming convention, whether it is "renaming the field" or 
"drawing a distinction". As markm has observed to me, the point at which we 
are making a real commitment is the day when we start revising our web sites 
to introduce the distinction. Note that I haven't said I was going to revise 
even my web site yet :-) I view what I am about to do as "test marketing". tyler

> meet, or even acknowledge, this burden. So far, all of your
> arguments amount to believing that a new name is like magic pixie
> dust that pushes aside misconceptions without injuring the
> history.  You've done nothing to lend credence to this view.  So
> far, all you've got is wishful thinking.

I don't expect the name to be particularly magic. I expect it to give me an 
assist in keeping people's attention.  If I can get, once in a while, a member 
of the audience to say, "excuse me, but what is the difference between your 
object capability and ordinary capabilities?" I will view it as a big 
success. If someone asks this question, then I know that it makes sense go 
into more detail. If no one asks, I have more reason for believing that I 
should just skip the detail and move on to the next high-level point. For the 
audiences I am currently trying to educate, going into too much technical 
detail when not required is as fatal as failing to go into enough technical 
detail when needed.

> The problem is not one of distortion, but of a lack of
> understanding. Other people use the word "capability" to refer to
> the same actual systems that we do.  There is no confusion about
> "what" we're talking about. The problem is that these other people
> have been trained to not see the full value of these systems.

This sounds like distortion to me. A view of a system that does not see the 
full value of the system pretty much fits my definition of a distorted view. 

> You are injuring the history for no actual advantage.

You sound as if the history has not already been injured to the point of being 
scarred beyond recognition.  Object capabilities, from Actors through KeyKos 
through E and EROS, have a proud history. Ambient authority capability 
theories, from Lampson through Li Gong, have a history of infamy that reads 
like a journey from Pearl Harbor to the Bay of Pigs to Vietnam without any 
stops in between. The history of capabilities as a whole encompasses both 
these separate histories. This is not a proud history. This is a train wreck.

I also think it is informative just how separate these histories are.  Object 
capability system developers never learned anything from ambient authority 
theoreticians because the object capability folks already had better answers. 
Ambient authority theoreticians never learned anything from object capability 
developers because, if they had, they would have stopped talking about 
ambient authority systems. It seems to me that even from a historical 
perspective, these two threads of development deserve to be distinguished.


> This misses the point. We are not seeking to criticize the straw
> man. We are seeking to dismiss it.

We agree the goal is to dismiss it.  But to dismiss it, when it is lodged 
securely in the majority of minds, requires that it be criticized, as you 
yourself have done quite eloquently.


> In my mind, the key point that we need to get across is *why*
> Lampson was in error. The way in which Lampson was wrong is the
> crux of what we need to communicate. His errors elucidate all of
> the key advantages of capabilities.  That's the "Ah-hah!"
> 
> Changing names makes it much more difficult to achieve this
> "Ah-hah".

The claim that introducing a distinguishing adjective makes it more difficult 
to achieve the "Ah-hah" is as untested as my hope that introducing the  
distinguishing adjective will assist in inspiring people to at least ask the 
question, "what is different?" tyler

The most reliable way of getting people to have the "Ah-hah" experience I have 
yet seen is  (er, boy, this is going to sound big-headed) to give them the 
demo with CapDesk. But that requires getting them interested enough to sit 
for half an hour. There is a huge marketing campaign to get people to the 
point of being willing to sit for half an hour. One of the steps along the 
way for some people is to read an articulate discussion of why what they 
thought they knew is false. If, as I believe, a marketing campaign must begin 
with a single word that entices the reader to go ahead and read a paragraph, 
that might lead to reading a document, and then watching a demo, then picking 
a distinguishing adjective is a key strategic element. Actually, I suspect 
that people who have backgrounds in marketing would refer to this as 
"marketing 101", but since I have no such background, someone who does have 
such a background should refute this if appropriate. alan_karp

 
> Not many real arguments have been presented for renaming. I've
> tried to refute those I could find. I think the renaming effort is
> primarily founded in a wish for an easy way to make the 31 years
> of bogosity just go away. Unfortunately, it's just not that easy.

I'm not expecting a magic bullet. I'm just hoping for a little assist in 
overcoming an existing, far-too-well-proven, impediment. 

> 
> I suggest that Marc Stiegler list the arguments in favour of
> renaming and provide evidence that each is valid. Basing a
> decision like this on the changing emotions of participants in a
> mailing list is insufficient.

Providing evidence requires test marketing. That is what I propose to do. 

--marcs
_______________________________________________
e-lang mailing list
e-lang@mail.eros-os.org
http://www.eros-os.org/mailman/listinfo/e-lang