Re: Are Level 0 features included in Level 1 Spec?

Hi Bryan, CG,

Although I know the charter says what it says, I'm not sure I see it as useful to capture "the current state" - which as Bryan says is in any case is piecemeal - and in any case is moving by the day. So I think that Level 0 is more usefully construed as "what is the minimum level of functionality that qualifies a device/browser combination as being capable of supporting a minimal 'modern mobile web app'". i.e. the baseline is not some average, or intersection of supported functionality of real browsers of "today", but something more abstract, but grounded on use cases.

For tracker's benefit here are the actions I am responding to (as well as replying to Bryan):

ACTION-2 Write a short document on how to get an L0 out and what it might mean
ACTION-4 Start a discussion on "what is the meaning of mobile web applications?"

The reason these two things belong together, from my perspective, is that you have to know what a minimal Web app means in order to say what functionality is required to support it.

Answering the question "what is a mobile Web app?" might be a question similar to asking "what is a Web site?" and indeed "what is the Web?" AFAIK these questions are thought not to require answers because the Web is existential and self-defining. You know what it is when you perceive it  - and rather like I don't need a definition of cheese to know when I am tasting it, those questions don't require formal answers. So trying to answer the question "what is a modern mobile Web app?" might be said to belong in the category "Fools rush in where angels fear to tread". But here goes.

We made a stab at this in the days of BPWG, way back when, in "Mobile Web Applications Best Practices" [1] under "Web application" [2] which says:

>> For the purposes of this document, the term "Web application" refers to a Web page (XHTML or a variant thereof + CSS) or collection of Web pages delivered over HTTP which use server-side or client-side processing (e.g. JavaScript) to provide an "application-like" experience within a Web browser. Web applications are distinct from simple Web content (the focus of BP1) in that they include locally executable elements of interactivity and persistent state.

While the focus of this document is the Best Practices that apply to applications running in a Web browser, in many cases these recommendations are equally applicable to other kinds of Web run-time, such as the W3C work on Web Widgets [WIDGETS] and also in a number of vendor-specific initiatives. <<

[1] http://www.w3.org/TR/mwabp/ 
[2] http://www.w3.org/TR/mwabp/#webapp-defined

So that's a starting point. What's different now to what was said then? i.e. what's different now to what the perception of what a mobile Web app was - back in the day - that means we now say "modern" mobile Web app. My suggestion is that we don't mean that it is some consensus view of what the "current state" is, but more - like Level 1 - necessary functionality to support some class of [possibly rather trivial] applications. I think that Matt Kelly's ACTION-3 "Circulate his research on types of apps requiring types of features" directly refers to this point. Let's put a pin in taking this discussion further till we hear from Matt.

To summarise then, I propose that we reconsider what Level 0 means by defining it as the set of features that is required to run some (smallish) set of app use cases out of the selection that Matt has defined (as amended by consensus of this group, of course). Could be that it means little more than what MWABP meant back in the days of yore. That would be a bit contradictory to the word "modern" but personally I can live with that.

At the F2F, we decoupled the definition of functionality of a level from the availability of comprehensive test suites for the level. We decided that Level 1 - being aspirational and some of the features being as yet unstandardised - is defined independently of a comprehensive test suite for it.

Seems likely to me that some, at least, of the more difficult test suite cases will actually fall into Level 0 - HTTP for example.

Anyway, I hope that this is not too whimsical, I hope that the group discusses this proposal and find consensus around it, and I hope that Bryan/AT&T think that this is is kind of spec they're on for editing. It should be noted that Bryan was the original editor of MWABP.

Final note, a (facetious) way of looking at this approach is that it is like saying:

"What is the functional difference between a Level 0 compliant device/browser combination and a piece of cheese?"

thanks
Jo

On 20 Jul 2012, at 08:06, SULLIVAN, BRYAN L wrote:

> Jo,
>  
> To me, having led development of a spec similar to the Level 0 spec (the WAC Core Specification), the reason to continue work on this spec and the test/compliance program it drives is that it does not yet achieve the stated intent from the wiki as “a de facto spec, aiming to describe the current state of the Mobile Web Platform. It is based off of market shares of the default browsers on deployed smartphone and tablet handsets. For the purpose of simplicity, this was roughly identified as the intersection in feature set of the Android 2.2 Froyo and iOS5 default browsers but with standard syntax.”).
>  
> The set of features in the spec are not universally supported by Android (even up to ICS) and iOS5. And I would challenge the limitation of the “current state” to Android and iOS. Even within Android, in various cases the spec goes well beyond what is reliably supported across the range of ICS-based devices (and it gets even spottier as you move back to Froyo).  I think in order to create an actual “defacto baseline” - which would actually be useful - we need more focus on Level 0 and validation in a reasonably broad set of launched devices.
>  
> So in order to actually produce a yardstick measuring devices against the current state, we support the completion of this work and the test assets that will back it. This will help both in setting the starting point more accurately for Level 1, and in catching those devices that for whatever reason fail in some aspect or other against Level 0 (which will still occur, as e.g. in Android not *everything* comes free with Webkit, as any OEM will tell you).
>  
> Thanks,
> Bryan Sullivan
>  
> From: Jo Rabin [mailto:jo@linguafranca.org] 
> Sent: Wednesday, July 18, 2012 11:07 PM
> To: Core Mobile
> Subject: Re: Are Level 0 features included in Level 1 Spec?
>  
> Thanks Bryan
>  
> That is a very helpful offer. The group needs to decide what the point of such a specification might be and what relationship it might have with a testing component. I know that there is some trepidation in the group about returning to a "meta" level of unactionable discussion on this topic. So it's clear to me that the original purpose won't wash and that if there is to be such a document then it needs to have a restated purpose. It might benefit from a name change to make that clear. I think that we would need to be focused and time bounded in the extreme in producing any such document. I'll say more on this in my responses to my ACTION-2 and ACTION-4.
>  
> I have a planned chairs call today with Robin, we'll discuss taking on AT&T's kind offer.
>  
> Meanwhile my apologies for there being no meeting summary as yet - I'm finding that summarising the 2 day meeting is taking more than 2 days. "I would have written less if I had had more time" comes to mind.
>  
> Thanks
> Jo
> 
> On 18 Jul 2012, at 23:43, SULLIVAN, BRYAN L wrote:
> 
> 
> Jo,
>  
> AT&T would like to support the definition of Level 0 and step up as editor of the spec. We believe that this baseline is very important and needs to be supported in the overall CoreMob program including test assets. Our developer program team has the resources to support this effort, and is looking forward to getting more directly involved in W3C work in support of developers and mobile web user agent compliance.
>  
> Given that we provide the needed editing support, I would like to get confirmation that there is support in general for continuing this work in CoreMob.
>  
> Thanks,
> Bryan Sullivan
>  
> From: Jo Rabin [mailto:jo@linguafranca.org] 
> Sent: Wednesday, July 18, 2012 9:16 AM
> To: Tobie Langel
> Cc: Core Mobile
> Subject: Re: Are Level 0 features included in Level 1 Spec?
>  
> On 18 Jul 2012, at 14:15, Tobie Langel wrote:
> 
> 
> 
> Hi Sun,
> 
> On 7/10/12 5:16 PM, "Sun, Dan" <Dan.Sun@VerizonWireless.com> wrote:
> 
> 
> 
> Hi Tobie/Robin/Jo/All,
>  
> Level 0 was touched a little bit in the first F2F but no clear direction.
> 
> Think there's a clear direction. L0 is shelved. Unshelving it will require
> providing a new spec editor for it and getting group consensus to do so.
> So far, no editor has volunteered.
>  
> I think the F2F meeting determined that there is still interest in level 0 - at least that is the way I have written the meeting summary and the way that my response to my ACTION-2 [1]  reads. I completely hear your point that without an editor stepping forward it will be difficult to progress.
>  
> I'm sorry that these have not appeared sooner - they are coming to this list very shortly.
>  
> Thanks
> Jo
>  
> [1] http://www.w3.org/community/coremob/track/actions/2
>  

Received on Friday, 20 July 2012 19:46:02 UTC