W3C home > Mailing lists > Public > www-qa@w3.org > May 2002

Re: Testable assertion tagging for W3C specifications

From: Alex Rousskov <rousskov@measurement-factory.com>
Date: Thu, 30 May 2002 11:01:08 -0600 (MDT)
To: David Marston/Cambridge/IBM <david_marston@us.ibm.com>
cc: www-qa@w3.org, Scott Boag/Cambridge/IBM <scott_boag@us.ibm.com>
Message-ID: <Pine.BSF.4.10.10205301044110.31432-100000@measurement-factory.com>

On Thu, 30 May 2002, David Marston/Cambridge/IBM wrote:

> >Cut-and-pasting is what we do today. That is not enough for my
> >applications because:
> >  - [...other reasons omitted...]
> >  - we want to provide reverse index where a reader can
> >    read the specs and get "one-click" access to
> >    test cases that correspond to the sentences she is reading
> >...
> 
> And how exactly do you propose to do that? 

There are probably many cool ways using state-of-the-art pop-up style
sheets, javascript, and such. Here is one idea that will work even
with a Lynx browser:

	The test suite renders the specs so that the first
	word, the first sentence, or the first line that relate
	to a test case have a footnote or citation mark. That mark
	is also an <a href> link to the corresponding test case
	description.

> Would your proposed solution work if the test cases were written
> by multiple authors?

Sure, test case authorship does not matter as long as the test suite
knows about (maintains) the test cases in question. There can be many
footnote marks attached to one word (using the Lynx example above).

The same is true for any test case processing/cataloging system, not
necessarily a test suite. For example, a WG can use this technique to
provide reverse links to independent test suites covering a given
assertion in the specs.

> How much burden is being imposed on those people who wish to cite
> the specs?
 
Not sure what you mean here. If I write a test case, I will need to
supply a citation address. That is not much of a burden compared to
cut-and-pasting a piece of text given a reasonable addressing scheme
(e.g., I may have to cut-and-paste a few times, or look at the XML
tags with unique IDs). In general, the overall burden should be much
less than writing the test case itself.

> Based on your comments up to this point in time, the only way I
> see that you could accomplish this is by checking whether there is
> any byte overlap between the byte ranges in the citations and
> whatever byte range has been highlighted by "the reader" in the
> above scenario. That sounds like a slow process.

Slow?! The overhead of run-time "footnote marking" above is probably
within 10% of the time it takes to render an average spec. And in many
cases (like the WG catalogue example above), all marking can be done
off-line. And it does not take much for the reader to click on a
footnote mark if they need to see the test case description. Am I
missing your point?

Thanks,

Alex.
Received on Thursday, 30 May 2002 13:01:52 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:40:29 UTC