On Wednesday, 7 December 2011 at 09:51, Dominique Hazael-Massieux wrote:
> Le mercredi 07 décembre 2011 à 00:01 +0000, Marcos Caceres a écrit :
> > Although I think this document is quite informative, I again would
> > like to raise objections about lumping app cache and widgets together
> > for the same reasons I raised last time.
>
>
> Your last message on the thread last time made me think your objections
> had been lifted:
> http://lists.w3.org/Archives/Public/public-webapps/2011JulSep/1459.html
>
> But I guess I misunderstood it. I'm a bit at loss as to how to make
> progress on this.
I was in agreement when the disclaimer I proposed was included in your
document, because it said that this was just one (of many) application of the
technology (particularly as it relates to widgets, which are used in a bunch of
weird and wonderful places). Without that context, it sounds like widgets are
somehow competing (and badly losing) with AppCache. That's my reading, and
that's why I keep harping on about :)
>
> > However, I don't want to have that argument again: I just want to say
> > I think it's disingenuous (perhaps make it more clear at the top of
> > the document that the document represents mostly your personal
> > opinion?). I'm also concerned that the text that I contributed to the
> > document about the variety of applicability of the technologies has
> > been removed.
>
> I did remove it, indeed; listing all the things the document doesn't do
> didn't seem very helpful to the reader, and seemed redundant with the
> scoping statement of the document:
> "This document summarizes the various technologies developed in
> W3C that increase the power of Web applications, and how they
> apply more specifically to the mobile context."
Maybe I'm being too critical about this, but these technologies don't "increase
the power" of anything: they are the bits and pieces applications to be created
by humans - power of application comes from the brains that put these tools to
work, not from the tools themselves.
>
> > I'm also concerned at use of the terms "limited" and "very limited" to
> > label "current implementations" as being both subjective and
> > relativistic - and it implies that attempts to implement have ceased;
> > particularly next to "well deployed", "Largely deployed", "Growing",
> > and "Getting deployed". Either remove that column, or present some
> > data to which you can underpin each of the labels.
>
> I agree that the current data are somewhat subjective (and have amended
> the description of the column in the introduction accordingly).
>
> My sources have been:
> * my personal knowledge of what's available where, and what I've heard
> is coming soon
> * http://mobilehtml5.org/
> * caniuse.com (http://caniuse.com)
>
> Ideally, I would like a lot more of the data in that column to come from
> W3C test suite results, but since we're not there yet, I think
> subjective (but I'm hoping reasonably well informed) data are probably
> more helpful to the reader than no data at all.
I don't think that data is all that suitable: for instance, I know of a lot of
widget runtimes that implement the widget specs, but I don't include them in
the implementation report because they are not fully conforming (and because
those vendors have not asked to be included). I only include stuff that allows
me to meet the exit criteria for a particular specification: it would be a lot
of work for me (or anyone) to source that data by running an implementation
through a test suite.
> And as any other part of the document, I'm happy to get specific
> feedback on which of these assessments you think are not in line with
> the market.
>
I'm biased, so lets start with widgets. For instance, why would you say
"limited" instead of "growing"? I guess that is only true if you exclusively
look at the big Web Browsers. If that is the case, then that is a fair claim
(no new web browsers have implemented the widget spec). However, other software
has (e.g., a bunch of new WAC runtimes, PhoneGap, etc.).
--
Marcos Caceres
http://datadriven.com.au
