RE: spec verbosity and algorithms

Alas, I think all of these have to go through submitting them as
bugs.

What I've found useful in turning an algorithmic description into
a more concise one is a two-step process:

1) "How would I test this?": Write the test cases (or at least
the test plan for what the test cases should test) for how you 
would determine if an implementation matched the description.

2) "What are the test cases testing?" Looking at the test cases,
it's often easier to write a specification that isn't in 
algorithmic terms which describes what they are testing.

As we get down the road in the testing task force, I think it
may help simplifying the spec this way.

Larry
--
http://larry.masinter.net


-----Original Message-----
From: public-html-request@w3.org [mailto:public-html-request@w3.org] On Behalf Of Julian Reschke
Sent: Friday, January 22, 2010 9:10 AM
To: Boris Zbarsky
Cc: public-html@w3.org
Subject: Re: spec verbosity and algorithms

Boris Zbarsky wrote:
> On 1/22/10 11:43 AM, Julian Reschke wrote:
>> This seems to be equivalent with
>>
>> "Two origins A and B are said to be the same if they are of identical
>> type (opaque identifier or tuple), and all of their components have
>> identical values.".
> 
> It's not.  Your phrasing makes all opaque identifier origins (which have 
> no "components") the same.

Of course the intent was to consider that a component.

> I do agree that less verbose is nice, if it can be done without losing 
> correctness and clarity, though.

Indeed.

Received on Friday, 22 January 2010 19:47:34 UTC