See also: IRC log
<Carmelo> Carmelo is here
<scribe> ACTION: Kai to continue integrating his tests WTF [recorded in http://www.w3.org/2008/12/02-mwts-minutes.html#action02] [PENDING]
Kai: still pending; would be good to get help from Robert
Wilhelm: I'll make sure he follows up
<dom_scribe> request to review testability of spec
dom: they're asking for review/help on testability of the spec
kai: spent quite a bit of time
with Marcos working with him on the specs and tests
... I think he's happy now - a bit concerned on the (un)stability of the spec, though.
wilhelm: any specific further action needed from us?
kai: nothing specific, just keep doing our work
<scribe> ACTION: Dmitri to review the DI guidelines [recorded in http://www.w3.org/2008/12/16-mwts-minutes.html#action04] [DONE]
<dom_scribe> Dmitri's comments
Dmitri: dom sent a reply to my
comments, and asked for specific contributions
... I'll work on it this week
<scribe> ACTION: Dmitri to make a proposal on dividing guidelines per test components [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action03]
<scribe> ACTION: Dom to review the di guidelines [recorded in http://www.w3.org/2008/11/18-mwts-minutes.html#action09] [DONE]
<dom_scribe> Dom's new version of guidelines
<dom_scribe> New version of the guidelines
<scribe> ACTION: Wilhelm to review new version of guidelines [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action05]
<scribe> ACTION: Carmelo to review new version of guidelines [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action06]
<scribe> ACTION: Kai to review new version of guidelines [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action07]
<scribe> ACTION: Kai to implement binary search tree to determin limits of memory test [recorded in http://www.w3.org/2009/01/13-mwts-minutes.html#action06] [DONE]
<dom_scribe> Kai's announcement of implementation
<dom_scribe> Memory test
Wilhelm: how I should run it? How do I know if I'm getting the proper results?
Kai: it starts with the highest
test; if that doesn't work, you go to lowest
... and then you go to middle steps
... one pb is that if you start too high, you may crash the
browser
... on my nokia device, I tried that, crashed the browser, and
the memory wasn't cleared
<Zakim> dom, you wanted to ask what 'what doesn't work' mean?
Kai: when it works, it gives you
a message "server limit exceeded" or something like that
... user agents with good memory capabilities can reach
that
... it's roughly for 2 Mo of data
Dom: maybe we could have
something more user friendly than "server limit exceeded" :)
?
... if we want to further advertise that test
Kai: could this be brought to another server to reduce the load on my server?
<scribe> ACTION: Dom to check if we can host the memory test on a W3C server [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action09]
<scribe> ACTION: Kai to make success message in memory test more userfriendly [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action10]
Kai: currently the generation of the data is kind of intensive
Dom: why is it generated on the fly?
Kai: it was the easiest way of
doing it :)
... I could probably be doing it more efficiently
... but in any case, this is likely to create load on whatever
server hosts it
... code is written in shell, should be fairly easy to
read
... Apprently, IE Mobile can't go over 1MB
<scribe> ACTION: Dom to run the memory test on his devices [recorded in http://www.w3.org/2009/01/06-mwts-minutes.html#action08] [PENDING]
<scribe> ACTION: Wilhelm to run the memory test on his devices [recorded in http://www.w3.org/2009/01/06-mwts-minutes.html#action07] [PENDING]
<scribe> ACTION: Kai to look into creating test cases for SSL and certificates support on mobile [recorded in http://www.w3.org/2009/01/13-mwts-minutes.html#action09] [PENDING]
Kai: I think I might have given
up on this - it is really difficult to come up with
something
... I would really like to make sure user agents allow to see
trusted roots
<scribe> ACTION: Kai to look into creating test cases for SSL and certificates support on mobile [recorded in http://www.w3.org/2009/01/13-mwts-minutes.html#action09] [DROPPED]
<scribe> ACTION: Kai to push forward with finding a replacement for the inputmode test [recorded in http://www.w3.org/2008/12/16-mwts-minutes.html#action07] [PENDING]
Kai: no progress on my side
<scribe> ACTION: WIhelm to look into replacement for inputmode test [recorded in http://www.w3.org/2009/01/06-mwts-minutes.html#action05] [PENDING]
Wilhelm: a couple of possible
interested areas:
... javascript timers - I'm writing a spec and tests around
them
<wilhelm> http://people.opera.com/wilhelmja/timers/tests/
Kai: is there any problem with timers?
Wilhelm: sometimes they're inaccurate, sometimes they don't work
<wilhelm> Notes: http://people.opera.com/wilhelmja/timers/data.html
Wilhelm: another idea would be
MathML through CSS
... not sure if that can be done with red/green square
... I'm also going through the DOM specs to see if I can find
something interesting there
Kai: I'll look in the timers stuff myself
Wilhelm: there is no spec for it yet
Kai: no spec from ECMAscript on this?
Wilhelm: nope; only ancient documentation from mozilla, and some stuff in html5
Kai: isn't MathML an XML-namespaced thing?
Wilhelm: it is, but the CSS profile for it is significantly simpler
<dom_scribe> What's measured improves, by Kai
Wilhelm: what should be our next post?
Kai: maybe talking about our search for a new test?
Wilhelm: sounds good
<scribe> ACTION: Kai to prepare a blog post [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action17]
Dom: Paul Nelson left a comment on the HTTPS test
<scribe> ACTION: Wilhelm to follow up on the blog comment re HTTPS [recorded in http://www.w3.org/2009/01/27-mwts-minutes.html#action18]
<scribe> ACTION: Wilhem to send Opera's feedback on OMA test suites [recorded in http://www.w3.org/2008/11/18-mwts-minutes.html#action16] [PENDING]
<scribe> ACTION: Dom to look back at OMA situation and report [recorded in http://www.w3.org/2008/11/18-mwts-minutes.html#action15] [PENDING]
Carmelo: will follow-up on NIST re-joining
<dom_scribe> AWTA 2009
Wilhelm: small but active
conference, with 25 participants
... around the Watir framework that can be used to automate
browser testing
... originally built for IE around COM components
... can be used to make browsers do anything
... various people are working on extending this to other
browsers
... we made a presentation on our Opera implementation
... can be done on any opera device
... it's made for being used over the wire
... There was discussions on the different frameworks
used
... and how web applications testers did their testing
... I think it's an important community to keep an eye on
... cross-browsers support for this could permit automate
testing for web standards conformance
dom: were there other browser vendors?
Wilhelm: nope, but someone has
implemented it for Safari through Applescript
... I think there is official support from the Chrome Team
Dom: any opportunity for getting other browsers to join that effort?
Wilhelm: no official support for
IE, but it "just works"
... the firewatir implementation for firefox is not officially
supported by Mozilla, but the extension mechanism allows to do
it fairly simply
dom: but is there any need to help there, or is just a matter of letting the community provide whatever is needed?
wilhelm: running code is mostly
what's needed at this point
... getting browser vendors to support this officially would be
great, but isn't strictly required
kai: was there any talk about pocket IE?
Wilhelm: no, it wasn't
... not sure what APIs they expose
Kai: I think it should be easy
Wilhelm: our implementations has
3 parts: the opera binary, to which you can communicate through
a documented protocol, a proxy that connects to the binary, and
tests scripts based
... we're adding support for Watir in our tests scripts
<dom_scribe> Automated testing that doesn’t hurt
<wilhelm> http://dev.opera.com/articles/view/opera-dragonfly-architecture/
Kai: support for Watir needs to be added in Opera binary?
Wilhelm: some can be added through javascript; some require new APIs in the binary, which we are adding
Next week (Feb 3rd)