Google web survey (webstats)

[Alan J. Flavell]

A chance remark on the German-language web authoring group (dciwam)
called my attention to http://code.google.com/webstats/index.html

An interesting survey of web pages. I haven't really drawn any
conclusions yet, but I thought readers here would find it of interest,
if they hadn't got there already.

(Needs an SVG viewer to see the plots. I took their advice and used
FF1.5).

--

[Chris Sharman]

Alan J. Flavell wrote:

A chance remark on the German-language web authoring group (dciwam)
called my attention to http://code.google.com/webstats/index.html

An interesting survey of web pages. I haven't really drawn any
conclusions yet, but I thought readers here would find it of interest,
if they hadn't got there already.

Yes - an unsurprising picture of widespread ignorance, for the most part
- and it tends to pose questions, rather than provide answers ("why do
people use table more than td ?").

I did see discussion of set-cookie2, however, which was new to me,
although I didn't discover why it was preferable to set-cookie, other
than a vague assertion that it was more secure.

Chris

[Eric Lindsay]

In article <dr*******************@news.demon.co.uk>,
Chris Sharman <ch***********@sorry.nospam> wrote:

Alan J. Flavell wrote:

A chance remark on the German-language web authoring group (dciwam)
called my attention to http://code.google.com/webstats/index.html

Yes - an unsurprising picture of widespread ignorance, for the most part
- and it tends to pose questions, rather than provide answers ("why do
people use table more than td ?").

I wonder what made the Google people decide to set the DTD as HTML 5,
and use the (almost unsupported) SVG.

It was a pretty depressing read however, when you see the ratio of
invalid markup on so many elements. I've started taking notes for
searches to at least check my pages for some of the worst of the faults.

--
http://www.ericlindsay.com

[Steve Pugh]

Eric Lindsay wrote:

Alan J. Flavell wrote:

A chance remark on the German-language web authoring group (dciwam)
called my attention to http://code.google.com/webstats/index.html

I wonder what made the Google people decide to set the DTD as HTML 5,
and use the (almost unsupported) SVG.

Because the research was carried out and written up by Ian Hickson...

Steve

[Andy Dingley]

On 30 Jan 2006 05:57:11 -0800, "Steve Pugh" <st**********@gmail.com>
wrote:

Because the research was carried out and written up by Ian Hickson...

"from misguided people using XHTML but sending it as text/html."

Yup, Hixie-spoor alright

[Alan J. Flavell]

On Mon, 30 Jan 2006, Steve Pugh wrote:

Eric Lindsay wrote:

Alan J. Flavell wrote:
> http://code.google.com/webstats/index.html
I wonder what made the Google people decide to set the DTD as HTML
5, and use the (almost unsupported) SVG.

In a way it's nice to see SVG getting some use, even if I had to
upgrade an old FF installation to 1.5 before I saw it (my usual
browser is Mozilla, to be honest). It makes a change from being told
that my up-to-the-minute Flash version is in need of upgrading[1]
Because the research was carried out and written up by Ian
Hickson...

Oh, I hadn't realised that, I must admit. It certainly seems to fit
the observations, anyway.

I still haven't drawn any firm conclusions from it, but it's kind-of
interesting, all the same.

cheers

[1] I digress, but it seems that some authors have found a way to
prevent their Flash content being seen at all by FlashBlock users -
they somehow pre-empt the FlashBlock menu, and instead they lecture
the reader about upgrading to the latest Flash version, even when the
reader already has the latest Flash version installed. Next time it
happens to me, maybe I'll bother to work out what they're doing.

[Pierre Goiffon]

Alan J. Flavell wrote:

A chance remark on the German-language web authoring group (dciwam)
called my attention to http://code.google.com/webstats/index.html

Oh, I just read your post today
Very nice document indeed ! Just a pity that no more information about
the source pages that were used for the study are present - "over a
billion documents" is a bit vague, but we can assume that such a big
number gives a good insurance that the study is quite representative of
the reality maybe :)

[Garmt de Vries]

Alan J. Flavell wrote:

A chance remark on the German-language web authoring group (dciwam)
called my attention to http://code.google.com/webstats/index.html

(Needs an SVG viewer to see the plots. I took their advice and used
FF1.5).

Strange. With my Opera 8.51, I can't see the SVG graphics, even though
this version of Opera supports SVG. The example at
http://www.opera.com/features/svg/ works fine for me.

What's the difference between the
<object data="/img/svg/circles.svg" width="220" height="150">
on Opera's example page, and the
<object data="charts/top19-elements.svg">
on the page about the Google survey?

Is it the lack of <?xml version="1.0" encoding="UTF-8"?> and DOCTYPE
declaration in the latter?

--
Garmt de Vries.

[Lauri Raittila]

in comp.infosystems.www.authoring.html, Garmt de Vries wrote:

Alan J. Flavell wrote:

A chance remark on the German-language web authoring group (dciwam)
called my attention to http://code.google.com/webstats/index.html

(Needs an SVG viewer to see the plots. I took their advice and used
FF1.5).
Strange. With my Opera 8.51, I can't see the SVG graphics, even though
this version of Opera supports SVG. The example at
http://www.opera.com/features/svg/ works fine for me.

Do you use Adobe plug-in? Didn't work with it for me in O8.51. "Works"
without in O9tp1
What's the difference between the
<object data="/img/svg/circles.svg" width="220" height="150">
on Opera's example page, and the
<object data="charts/top19-elements.svg">
on the page about the Google survey?
Width and height not specified.
Is it the lack of <?xml version="1.0" encoding="UTF-8"?> and DOCTYPE
declaration in the latter?

No, I don't think so.
--
Lauri Raittila <http://www.iki.fi/lr> <http://www.iki.fi/zwak/fonts>

[RobG]

Chris Sharman wrote:

Alan J. Flavell wrote: [...] Yes - an unsurprising picture of widespread ignorance, for the most part
- and it tends to pose questions, rather than provide answers ("why do
people use table more than td ?").

I think that is an incorrect conclusion - the horizontal axis isn't
labeled but I'm pretty sure it represents the percentage of pages that
have the element present, not a total count of their occurrence.

What is indicated is that about the same proportion of pages have TABLE
elements as have TD elements, which is pretty much to be expected.

Looking deeper, the svg width for the TABLE bar is 503.068887939895 and
for the TD bar is 502.729586816226. The display box is 600 wide, so
about 0.06% of pages surveyed (i.e. around half a million out of one
billion pages) have a TABLE element but no TDs. I guess those pages
have some other element such as TH in their tables and no TDs - or
invalid markup.

[...]
--
Rob