Re: Colloquial Tidbits

Hi Sean,  

On Friday, 16 September 2011 at 16:07, Sean B. Palmer wrote:
> On Fri, Sep 16, 2011 at 5:32 AM, Marcos Caceres <w3c@marcosc.com (mailto:w3c@marcosc.com)> wrote:
>  
> > > * Quality assurance scenarios. How do people actually check their
> > > websites? The scope here is people who write the HTML / CSS /
> > > JavaScript themselves, so people who are pumping material through a
> > > CMS don't count.
> >  
> > Those still seem like still valid cases, they just make those templates
> > available as a platform to scale.
>  
> In these cases, though, who is most commonly doing the QA and what
> methods are they using? If I'm using a default Wordpress or Drupal
> theme, as a user I would probably hope to be doing as little QA as
> possible in technological terms.  
Right, but this is significant because it presupposes an even greater question: do people do "QA" at all on the markup. Look at the google home page, for instance: it uses the HTM5 doctype yet it doesn't validate, but I imagine it's had the hell QA'd out of it.  

So, this means:  
0. What is QA (to people out there, not us in the WG… in the colloquial sense)?  
1. How many QA VS how many don't?  
3. Of those that do, if statistically significant, what do they QA? and why?  
4. Of those that don't, why not?  
> Obviously there may be other forms of
> QA. If it's a company weblog, for example, I may have to keep to
> company guidelines for the content I product.
Right.  
> On the other hand, the people producing the templates will have to be
> doing extensive QA.
This can't be assumed. But we can test this! We can pull all the templates from wordpress, and validate them. It's at least small indicator. Or we could just pull wordpress sites that are not using the default template. That should give us at least a small indication (so long as the templates match a template in the Wordpress database).  
>  So CMS and CMS template designers would count as
> subjects for the point suggested above.
Absolutely :)  
>  I wonder, too, about certain
> kinds of complex CMS. MediaWiki for example allows a range of
> structured output, go so far as to allow the embedding of raw HTML. Of
> course if people are actually needing to alter the results of their
> pages in terms of layout and interface even in a CMS, then those cases
> would count.
Agreed.  
>  
> > I guess what would also be interesting here is Q) what percentage
> > of the Web is using a custom CMS vs … um, static pages?…
>  
> I'd like to know that too, indeed. Your point about defining a CMS is
> a very good one: obviously here one of the most prominent
> characteristics that I have in mind is that a CMS is anything that
> abstracts the layout and interface design requirements away from the
> user. But there may be other characteristics of a CMS.
Seems like a good start for our wiki :)  
> Since more people are subscribed to the list than I had realised—and
> welcome to all those who have signed up—I may give a few more formal
> posts now outlining some of the background to the philosophy of the
> CG. The principles along whose lines we're investigating are mainly
> borrowed from the cognitive sciences.
Just wondering, do we have any stats people in the group? My better-half has a lot of stats training, and I often run to her for algorithms, established practices, etc. I can alway go and annoy her if we need proper stats models, etc.  

> > > Will probably involve spending lots of time in the DOM inspector. I'm
> > > constantly having to draw in boxes to check why the box model isn't
> > > quite working as I think it should be.
> >  
> > Would we be talking one on one to developers? Do you see this being
> > a survey?
>  
> We could do, pursuing them ourselves, and of course developers would
> be free to contribute of their own volition by mailing the CG at
> public-colloquial. I'm hoping that some level of curated wiki access
> would be possible too, but I'm not sure that will be possible under
> the W3C's contribution framework.
Lets think about that one. But it would be good to have a "mixed method" approach to gather data (i.e., quantitative and qualitative).  
> > There is research in the general standardisation literature around this.
> > That is, the explosion of consortia in the last century and towards the
> > movement of towards informal consortia like the WHATWG
>  
> If you have references to this literature, please feel free to share
> them with the list and add them to the wiki!
>  
> http://www.w3.org/community/colloquial/wiki/
Will do.  

> Such references are precisely the kind of things that will enrich our
> discourse. I've been concentrating mainly on the QA aspect, and have a
> few things along those lines to share.
Excellent. I've looked at the economics and dynamics of standardisation when I was a student three years ago. Having said that, we should be careful to really limit our scope… like having a very little project for us to collaborate on, just to see if the group can work together (as none of us is here getting paid!… and if you are, good for you :) ).  

> > standardisation is increasing, but not in traditional settings (e.g., the
> > W3C). Obviously, the W3C has recognised this, which is why these
> > community groups were created
>  
> Which makes it quite meta that we're investigating the very process
> that gave rise to the community groups mechanism in the first place,
> indeed.
:)  
>  The W3C community groups are exciting, and I wonder how much
> further the W3C could step in the same direction.
Well, lets hope this translates into some money for them somehow… lets just say this is a freemium model.  
>  These early groups
> are a testbed for this, and I'm hoping that Colloquial CG will serve
> as a kind of primary meta-analysis group.
Sounds good. But again, lets think about the first problem we want to tackle. There are a lot of great things for us to look at… and I'm really excited about building tools that will enable this to happen. I know in the social sciences, they could really benefit from the tools we could end up building to conduct our own research. Looking at the member's list of names, I see some kickass coders.  
>  
> > (to fold groups back into the W3C, while providing the tools that
> > enables standardisation easily… such as IRC, this mailing list, the
> > Website, etc.).
>  
> Would you apply the term standardisation to cross-browser polyfills?
Absolutely.  
> Or what about an effort like pdf.js? (I don't have a well formed
> opinion on this, so I'm not challenging your understanding; I'm really
> asking!)

And here comes term that needs definition number 2: "standardisation". I will provide the text from my PhD around this.  

To answer your question above, probably would not to fall under standardisation because it's an implementation of some specification. Standardisation generally only produces specifications or guidelines that are then implemented to create products, and pdf.js didn't create the PDF spec (is there one?), even if lots of people collaborated on it. Contrast that to Common.js, or micro-formats: where an agreed set of conventions is standardised.  

> > > * Fashions. The web ten years ago was very different to the web now
> > > not just because it had fewer users, but also because the technologies
> > > have changed dramatically.
> >  
> > Can you give some examples of what you are thinking here? is it hardware
> > or software? I guess both?
>  
> Hardware has mainly had an effect in that sites such as YouTube
> wouldn't have been possible ten years ago, as far as I know. Software
> has probably had a bigger effect, but I don't know. I'm sure that
> things have changed, but it's an open research question for us as to
> why these changes have come about.
There are also great books written about this. I'm writing this from an airport offline, which is why I don't have any references. I'll provide them as soon I'm back online.  

> > > Could we have predicted CMSes ten years ago?
> > We might need a definition of CMS… like I implied above, I see PHP, for
> > instance, as a CMS on which other CMSs can be built.
>  
> I agree, this is a very good point.

Thanks!

Kind regards,
Marcos

Received on Saturday, 17 September 2011 07:44:20 UTC