W3C home > Mailing lists > Public > www-tag@w3.org > March 2006

Re: Rule of Least Power

From: Pat Hayes <phayes@ihmc.us>
Date: Wed, 8 Mar 2006 12:11:13 -0800
Message-Id: <p06230905c034e7c30887@[192.168.1.4]>
To: noah_mendelsohn@us.ibm.com
Cc: www-tag@w3.org

>Speaking for myself, and not necessarily for my co-editor Tim or the rest
>of the TAG:
>
>Roger Cutler wrote:
>
>>  Last month a reference to the "Rule of Least Power" appeared in
>>  some AC-Member emailing and, since I found it pretty
>>  interesting, I gave it some exposure in Chevron.  I regret to
>>  tell you that what response it got was of a sort that you would
>>  probably consider "unintended consequences".  That is, the
>>  principle was applied to support positions that were probably
>>  not really what you had in mind when you formulated it.  In
>>  particular, not a single comment that it provoked had anything
>>  whatsoever to do with declarative languages.
>
>[...]
>
>>  Frankly, I used to try to follow the Rule of Least Power and I gave up.
>
>Obviously, there are many ways for a given audience to misinterpret a
>finding like this, and I'm sorry it isn't proving more immediately
>effective for you.  It occurs to me, though, that one might make headway
>by starting with examples that are in some sense obvious and
>noncontroversial.   Let's say that someone at Chevron came up with a plan
>to do all of their Web pages completely in JavaScript.  I.e., even a
>completely static page would contain the minimal HTML framework to launch
>the JavaScript, and the script in turn would use the DOM to prepare and
>display the intended content.   To a browsing user, the pages would
>probably look fine.  If this were done with publicly accessible pages,
>however, search engines like Google and Yahoo would presumably not be able
>to find the pages very well.  That can't be good for your business.
>Similarly, internal search engines would not be able to index pages that
>you prepare for use on your Intranet.   It's also unlikely that you could
>effectively import such pages into popular word processors, something you
>can do with quite good fidelity if the content is in HTML.  So, I think
>it's easy to make the case that such pages would, at least in those
>senses, be less valuable than pages with similar content conveyed in HTML.
>

That indeed makes perfect sense. What I am now puzzled about, 
however, is that this isn't what the RuLePwr document seems to be 
saying. This example has to do with the extent to which information 
is accessible, even to processes that were not being considered when 
the information was being created: information buried in Javascript 
is less accessible than information in HTML. OK, good point; but the 
RuLePwr document talks about computing power in the 
theory-of-computing sense. It seems to be saying that the relevant 
dimension, along which one should strive to be minimal, is the Turing 
hierarchy. And these two dimensions - accessibility of information, 
and Turing computability -  seem on the face of it to have very 
little to do with one another.

When I read RuLePwr , it occurred to me that there is yet another 
dimension that seems to be relevant here, which is that of formal 
expressive power. This particular issue is very current in the 
semantic web literature, since it underlies the hierarchy of more and 
more expressive languages being considered (RDF. RDFS, OWL-DL, 
OWL-Full, RIFF and others). I was rather surprised that RuLePwr did 
not refer to this, since this seems to be more germane, and closer to 
the kind of example you give above, than the rather narrow issue of 
Turing computability. The two hierarchies of course do not coincide: 
very inexpressive languages are Turing-complete, while quite 
expressive descriptive formalisms, such as full first-order logic, 
have uncomputable models.

Im not sure how this comment could be helpful, but it might be an 
idea to consider re-wording the document, perhaps with some examples 
like this, to make your intention clearer. It isn't obvious what it 
even means to discuss the computability status of HTML, for example.

Pat Hayes


>This is a somewhat contrived example, in that few of us are tempted to
>build static pages entirely in JavaScript, but that's the point.  In many
>cases, the Rule of Least Power brings you to solutions that have
>commonsense value.  I think that if one looks to such obvious examples one
>may convince people that they are already acting on the Rule of Least
>Power.  If so, maybe they can start to understand how the Rule might guide
>them toward useful decisions that they might not have made as well without
>it.
>
>Thank you in any case for the feedback!
>
>--------------------------------------
>Noah Mendelsohn
>IBM Corporation
>One Rogers Street
>Cambridge, MA 02142
>1-617-693-4036
>--------------------------------------
>
>
>
>
>
>
>
>
>"Cutler, Roger (RogerCutler)" <RogerCutler@chevron.com>
>Sent by: www-tag-request@w3.org
>03/07/2006 10:36 AM
>
>         To:     www-tag@w3.org
>         cc:     (bcc: Noah Mendelsohn/Cambridge/IBM)
>         Subject:        Rule of Least Power
>
>
>Last month a reference to the "Rule of Least Power" appeared in some
>AC-Member emailing and, since I found it pretty interesting, I gave it
>some exposure in Chevron.  I regret to tell you that what response it got
>was of a sort that you would probably consider "unintended consequences".
>That is, the principle was applied to support positions that were probably
>not really what you had in mind when you formulated it.  In particular,
>not a single comment that it provoked had anything whatsoever to do with
>declarative languages.
>In general, it seems to me that the thrust of your intention is to
>encourage the use of declarative languages.  My personal experience with
>this, for what it is worth, has not been positive -- in the sense that I
>have not been able to get any substantial uptake among my user community
>other than a few isolated enthusiasts.  It seems that, in practice, the
>learning curve tends to be too high.  In addition, there is a tendency for
>the scope of what is being done to creep, and then you either start using
>some extremely obscure constructions into the declarative language or you
>give up and re-do it using a complete, procedural language.
>Frankly, I used to try to follow the Rule of Least Power and I gave up.


-- 
---------------------------------------------------------------------
IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Wednesday, 8 March 2006 20:11:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:39 GMT