W3C home > Mailing lists > Public > public-bpwg@w3.org > June 2005

RE: The devices-universal websites "myth" and the semantic web

From: Antti Martikainen <Antti.Martikainen@sysopen.fi>
Date: Tue, 14 Jun 2005 13:20:11 +0300
Message-ID: <3D6EBE04E53BB54DBF66372BF200AAC25BD576@webmail.sysopen.fi>
To: "Nicolas Combelles" <nicolas.combelles@apocope.com>, <public-bpwg@w3.org>

Interesting topics are being touched in this topic, and
I'd like to join the discussion, so...

We have been working with the multi-device problems for a while now,
and yes, I would say that it is reality soon that applications can
be modelled to adapt to different kinds of device types. Although
standards are not likely to exist soon, different kinds of frameworks
will exist. For achieving an efficient development process, it is likely
that devices are categorised to map them with specific kinds of content. 
This is effectively achieved at least with WURFL, which really seems
to be on the right track with its patching abilities.

> What defines a website ? A URL ? Content ? Information design ?

>From the service provider's perspective, it's expensive and
laboriuos to maintain separate Web-sites, and as it is possible
to recognise the requesting device, it is a natural objective
for the service provider to provide a single URL for accessing
an adaptive application, which reuses most parts of the application's
specification across different kinds of devices.

>From the user's point of view, there should only be one URL, as it's
often not that nice to start looking for the mobile page's adress
on the road..

Another pointer that I would like to bring up, is that now that there
are Web browser implementations that try to adapt Web content for mobile
access (often with not that usable end results), it would be very important
to have a way, through which an already (server-side) adapted page could
let such a browser know that it should not be altered. This is crucial
for companies that in the future want to maintain their brand and level of
usability across device groups.

All in all: it is likely that not all Web sites in the future will adapt.
Then again, there are sites that will, and they require solid means for
identifying and categorising devices for appropriate content and layout
mappings. In addition, they require means to let adaptation-capable Web
browsers know if they they don't want the content to be altered by any
means.

> 1.1) Should I serve the same content to all devices ?
> -----------------------------------------------------

As was said in the conclusions, this is likely to be mostly
a site structural issue. In our experience, the overall
layout of a device-specific Web page (and the content that
is mapped to specific layout sections) is something that
must be designed. Adapt this automatically somewhere, e.g.
by the adaptive Web browser, and you are likely to go wrong
usability-wise. This does not mean that services should be
excluded from small devices, as they can be given additional
navigational points to the "hidden" content. Then again, image 
sizes are one thing that must change according to the device. 
But this can also be done on the server side. Therefore, making 
this happen somewhere else, e.g. by the operator gateway(!) can be
really bad considering the overall quality and appeal of
a commercial service.

> 1.2) Can I use a single structure ?
> -----------------------------------

No, not really, if you truly want to adapt content to the capabilities
of a device..

> Actually you will do best if your structure include several 
> content "length"
> (title, heading, short content, long content). You will then 
> have a heavy
> use of the CSS "display" property to show the more adapted 
> structure to the
> user depending on it's device (however "not displayed" 
> content is still
> downloaded, and heavy pages is not what we can call a "best 
> practice").

CSS apparently can be used for doing part of the job, but considering
the overall architecture of a large-scale commercial site, it is a
really bad solution. CSS is supposed to be about visualisation. Mapping
content to different kinds of devices is something else. And as was said,
it's bad to deliver additional content for small devices..

> 1.3) URL and communication
> -------------------------
> Conclusion : 
> ------------
> Having different URLS when you have differents versions of 
> your site is
> easier and sometime more suited to your marketing objectives, 
> and anyway you
> still can have a single URL even if you have different 
> versions. So unique
> URL is not an argument to justify a universal site.

Perhaps it's not as much about the single URL, but it's more
about adaptation, i.g. reusing same interaction and UI specifications
across device types as much as possible, for making it feasible
to build multi-device services.

> The "devices-universal websites concept" is irrealistic and 
> except for very
> specifics cases (students, geeks, and web-guru personal sites 
> and blogs :oD
> ....) is never suited to the webmaster nor the users needs or goals.

I would strongly disagree. Device-universal sites will not come
without platforms specifically designed for that purpose, which
doesn't mean it's impossible to build multi-device Web sites that
adapt very well to the devices detected.

I hope my points made any sense and I'm curious to see how the
conversation will proceed.

Best wishes,
Antti

Antti Martikainen
Senior Consultant
SysOpen Digia Plc
Hiomotie 19, FIN-00380 Helsinki, Finland
Email antti.martikainen@sysopen.fi
Tel. +358424 2020 653, GSM +358405507174
Received on Tuesday, 14 June 2005 10:23:26 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:42:55 UTC