[Fwd: Re: SVG Tiny 1.2 spec coverage and OMA test suite]

Forwarded message 1

  • From: Chris Lilley <chris@w3.org>
  • Date: Wed, 31 Oct 2007 11:38:27 +0100
  • Subject: Re: SVG Tiny 1.2 spec coverage and OMA test suite
  • To: Dmitri Silaev <Dmitri.Silaev@Sun.COM>
  • Cc: schepers@w3.org, w3c-svg-wg@w3.org
  • Message-id: <1376633460.20071031113827@w3.org>
On Tuesday, October 30, 2007, 7:24:02 PM, Dmitri wrote:

DS> Hello Chris and Doug,

DS> I am an representative of the MWTS working group and we are working 
DS> together with OMA IOP browsing group on on helping them integrate the 
DS> SVG Tiny 1.2 test suite in their testing process. OMA browsing group 
DS> issued the draft of the test requirements for their test suite. This 
DS> draft could be found here:

DS> http://lists.w3.org/Archives/Member/member-mwts/2007Oct/att-0006/OMA-ETR-SVG_Mobile_Domain-V1_0-20070424-D.zip

Thanks for the email, Dmitri. The SVG WG had earlier extended an invitation to the OMA to work withthe SVG WG on the 1.2T test suite, but we did not get a response and I suspect the communication did not reach the right people - so, good to talk to you now.

Your requirements document has been received, and is on the agenda for the next face to face meeting, which happens next week in fact, in Cambridge MA.

DS> In short, they would like to test following areas:

DS> 1. Multi feature test cases.
DS> 2. Untested features.
DS> 3. Support for OMA use cases.
DS> 4. Potentially difficult areas.
DS> 5. Features new to SVG Tiny 1.2

We have some multi-feature test cases, although mostly we try to produce unit tests. We are certainly happy to add more.

As we become aware of untested features we add tests to cover them. We will be reviewing test coverage also at the f2f. If you have already started on a coverage analysis, we would be happy to see what you have so far and to work with you on combining with our work which is at
http://www.w3.org/Graphics/SVG/Group/wiki/TestCoverage

We would also be happy to work together to improve test coverage, and can supply you with documentation on how our tests are constructed.

Would OMA use cases be described in the "SVG in the Mobile Domain Requirements Document" document referenced from your communication? If so, do you have a direct link to it (rather than a link to the OMA home page).

We agree that the SMIL timing features can be tricky and require careful analysis. We have benefited in this area (and others) from advice from Dr Olaf Hofmann, and invited exert to the SVG WG.

In terms of features new to SVG Tiny 1.2, looking at the ones you mention, some are new to Tiny 1.2 9ie were not in Tiny 1.1) but have a long history in SVG (being ion SVG Basic 1.1 and SVG Full 1.1) so are well tested, since appropriate tests from the 1.1 Full test suite were converted to Tiny 1.2 tests:

?	Support for opacity, gradient
?	External element referencing

Others are indeed new in 1.2 and seeing their first use in Tiny 1.2:

?	Support for Multi line text and editable text
?	Support for multi media elements and their synchronization
?	Support for managing by discard and prefetch elements
?	Support for script execution and uDOM binding to the SVG DOM.
?	Introducing new concepts for event handling via listner and handlers 

Have you see the beta 1.2 test suite 9and associated implementation report) at
http://www.w3.org/Graphics/SVG/Test/

As a Sun employee you have W3C Member access so also please see our staging area for the test suite:
http://www.w3.org/Graphics/SVG/Group/repository/testsuite/1.2T/svg/
and the wiki page for test issues
http://www.w3.org/Graphics/SVG/Group/wiki/TestSuiteCommentsSVGMobile1.2
and the test suite instructions
http://www.w3.org/Graphics/SVG/Group/repository/docs/SVGTestSuite-howto.html



DS> Document also contain the list of requirements for their test suite 
DS> grouped by SVG features.

DS> Now OMA browsing group would like to understand how much of the SVG Tiny
DS> 1.2 Specification is covered by SVG tests produced by your team in 
DS> September, 2007. Our group helps them to gather information about 
DS> current coverage of specification and separate uncovered areas and features.

Please see the coverage page at
http://www.w3.org/Graphics/SVG/Group/wiki/TestCoverage
for our work towards that report.

DS> As test suite developers, the SVG working group may/shell prepare 
DS> coverage report for final version of test suite as explained in "Test 
DS> Development FAQ" section on page:
DS> http://www.w3.org/QA/WG/2005/01/test-faq#enough

Yes, we are aware of that requirement. Lofton Henderson, chair of the QA WG, was previously involved in the SVG 1.1 test suite work.

DS> The detailed report helps our group to understand current coverage and
DS> comments OMA browsing group requirements. Therefore we interest to get
DS> such report in future. Your group may produce such report quickly, but
DS> the lack of resource may delay the finalized documents. So, the part of
DS> this task need to be done by our group, we realize this. Therefore we 
DS> have the number of questions to understand the current state and future
DS> plans for this task. Could you please answer following questions (we 
DS> believe the coverage report for test suite going to be produced by SVG
DS> WG). The questions are based on "How many tests are enough?" of "Test 
DS> Development FAQ" (see link above).

DS> Is there any coverage report for SVG Tiny 1.2 Specification provided by
DS> SVG Tiny 1.2 test suite beta?
(see above)

DS> If no, is there a plan and preliminary date to create such report?

DS> Coverage measurement involves partitioning the specification. What is 
DS> the way for this partitioning your group choose? Feature, language 
DS> elements, logical sections, testable assertions and so on.

Logical sections (chapters and sections) and then testable assertions withing each. The convention for names of tests shows the chapter and section, see 
http://www.w3.org/Graphics/SVG/Group/repository/docs/SVGTestSuite-howto.html#namesT12

DS> What is the coverage goal for SVG Tiny 1.2 test suite.

We intend to cover each feature in the specification with at least a unit test. We aim to identify all testable assertions.

We do not aim at exhaustive combinatorial test (ie each value of each attribute in combination with each value of each other attribute).

DS> What the format of the coverage report you going to produce: mapping of
DS> tests to areas of the specification, more detailed report with binding
DS> of specification to tests and coverage numbers, etc?

Currently the links are from the test suite to the area of the specification that is primarily being tested. (Other areas are often tested too, for example all tests contain some text and most contain either paths, basic shapes or both with solid color fill. ).

DS> It will be helpful if you send any additional information about the test
DS> suite and coverage for SVG Tiny 1.2 Specification and related issues.

DS> Thanks in advance,

DS>      Dmitri Silaev
DS>      W3C MWTS Working Group


Many thanks for your mail and we look forward to working with you in the future.

-- 
 Chris Lilley                    mailto:chris@w3.org
 Interaction Domain Leader
 Co-Chair, W3C SVG Working Group
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG

Received on Wednesday, 31 October 2007 16:01:37 UTC