W3C home > Mailing lists > Public > public-wai-ert@w3.org > November 2007

Re: Using Earl to store software performance test results

From: Shadi Abou-Zahra <shadi@w3.org>
Date: Tue, 06 Nov 2007 17:32:57 +0100
Message-ID: <47309739.2070402@w3.org>
To: Cristiano Longo <cristiano.longo@tvblob.com>
CC: public-wai-ert@w3.org

Hi Cristian,

Thank you for your comment. In fact, we have an open issue on making the 
outcome value of the test into an extensible list:
  - <http://www.w3.org/WAI/ER/EARL10/issues#outcomevalue>

We might not directly provide a way for expressing numerical results as 
our primary use case are pass/fail type of tests. However, it should at 
least become easier to create outcome subclass to express integers.

We hope to publish an updated Working Draft with these changes within 
the coming month and will look forward to your comments on it.

Regards,
   Shadi


Cristiano Longo wrote:
> 
> 
> 
> Dear all, 
> I need a 'framework' to store performance test results of various kind of 
> algorithms and software. I think that EARL is suitable for this purpose, but 
> such a test result is not "Binary"(passed or not passed), but a numerical 
> value. An example could be the convergence speed of an iterative algorithm, 
> or the error of a genetic algorithm after some iterations.
> 
> Sean B. Palmer suggested me to use the earl:info property to store such a 
> value, and i think it is a good idea. In order to have a common framework to 
> handle this situation, should be a good idea to create a subproperty of 
> earl:info with only numerical values as range, and to subclass 
> earl:TestResult.


-- 
Shadi Abou-Zahra - http://www.w3.org/People/shadi/ |
   WAI International Program Office Activity Lead   |
  W3C Evaluation & Repair Tools Working Group Chair |
Received on Tuesday, 6 November 2007 16:33:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:18:29 GMT