RE: You've been framed :-)

["Nikola Kasic" <nikola@xxxxxxx>]

Mark, do you have any book on HCI that you would recommend?
Especially for web design purpose.
Nik


-----Original Message-----
From: Mark Hetherington [mailto:mark.egroups@xxxxxxx]
Sent: 09 April 2003 23:02
To: ukha_d@xxxxxxx
Subject: RE: [ukha_d] You've been framed :-)

mark_harrison_uk2 wrote:
> It has become a seminal text among the designers of BIG websites.
> Some particularly alarming things for all of us to bear in mind,
> including the fact that NONE of the world's top 10 websites uses
> frames any more, since they have all discovered that, in practice,
> they don't give good usability.

The "Frames vs. No Frames" argument has been ongoing for years
and there
is likely a "seminal text" refuting the very reasons stated for
not
using frames. This text would likely quote the world's top 10 top
websites as well. Statistics are as volatile as reasons for bombing the
hell out Iraq... they change by the hour and the person you ask.

The main concern with website design is the target audience and this is
the only point the many writers of books about web design ever actually
have in common but they largely tend to put this aside for certain
elements if the author is biased a certain way. The site should always
be designed around the target audience and if they are likely to like
frames, then no book in the world should stop you using them.

> Observed evidence in the field is that 90% of clicks made from a page
> will come from stuff visible on the first screenful. Hence, the
> overall recommendation is to layer your site... On entry level pages,
> it is vital that all the possible choices are provided on a single
> screenful. Once someone has drilled down to a particularly detailed
> page, then scrolling is acceptable, since by then, the user is likely
> to be engrossed in READING the page, rather than simply scanning it
> to find which link is appropriate.

One simple rule of web design which seems to have been lost in the move
to "flashy" sites and the increase in "experts" writing
books, is that
actual content should be accessible within 3 clicks. This fits in well
with your quotation(?)/comment on drill down since that is the long
established point at which a viewer will give up turn to another site.
So for say a shopping site, this means specific product information
within 3 clicks of the home page (e.g. select category -> product
page->
specific product). Further links from that are of course acceptable, but
aim for something meaty within 3 clicks.

> Rendering of pictures fast is harder, but the trick here is to
> include explicit picture geometry directives in the HTML, so that
> rather than putting in a little box for a picture placeholder, then
> redrawing the whole page once the picture is loaded, the client can
> draw a box of the correct size for the placeholder, and fill in the
> picture only.

This is basic stuff. All image tags should be complete rather than
partial but this applies to HTML as a whole as much as to images
specifically. Where possible, tables should also be completely defined
with size values since the few bytes extra it takes will usually result
in a final layout rendered prior to page download completion.

Text is obviously faster to load than graphics but not always
appropriate.

> Now - what is my reference "great" website????
>
> It's Google....
>
> ... everything you want to do, 100% of the time, on the first screen
> of the homepage.

Actually, no. There are google features which are only available through
drill down. The feature set of google is in essence limited to a pure
database search with 5 main categories so despite the huge size of the
database, it is trivial to provide a simple front end website for it.
Google does not make a good template for all web sites.

For newcomers to the net, Google actually loses out to Yahoo, MSN and a
number of portals since there is nothing to click on to get to something
immediately. While a more experienced net user might find google's plain
interface friendly, a newcomer has no way to just click around and must
type something.

Whenever I introduce someone to the net and start them at Google, they
lose interest there and then. Start them at say Yahoo, or a portal site,
and they are happy to click about and later actually start typing
something in a search box.

Although personally I like the no frills google interface for web
searches, I miss the missing features from it's predecessor on which it
is based - Deja. Google have recently begun resurrecting the full Deja
feature set back but are IMO not quite there yet.

> ... Once you've done a search, and are therefore interested in the
> contents OF THAT PAGE rather than using that page solely to navigate
> to the next, then scrolling is fine.

Scrolling on target pages is a quite a widespread web design comment.
Where long pages lose out to frame based navigation is the need to
scroll to quickly navigate since a non-framed site will usually provide
navigation at the top or bottom or both. A well designed framed site
offers main navigation features at all points while reading any given
page. Of course you can fix this with the likes of DHTML, but then you
also restrict your audience.

There are many arguments over how long a page ought to be before being
split into a sequential series of pages. This then brings in other
navigational issues so I doubt anyone has been prepared to actually give
figures for it.

> ... No frames, but explicit "where you are in the site" text
on every
> page, means that you can't get lost by being incorrectly divered to
> an inneer frame

But you will have to scroll to find it. "Where you are" text is a
relatively recent invention but is often helpful on any large site
whether framed or not. Google is not a large site when you consider what
it actually offers and this is how it gets away with very simple
"tricks".

> ... Clear branding, so you never lose site of the fact that you're on
> Google, but that takes up precious little space (38 by 51 pixels),
> and is explicitly sized in the HTML so that the page loads around it,
> rather than having to refresh once the image is downloaded.

As are images on all well designed sites.

Google gets it pure speed from a lack of graphics. Since they have found
alternative revenue streams to traditional banner advertising, they can
present a purely text based page with merely their logo as graphical
content. Not all sites can dismiss graphics completely from their design
whether or not they rely on sponsor banners.

Unfortunately, the expectation of the newcomer to the web is something
flashy and often graphical, so removing all the graphics for speed will
not make you website a winner if your main viewers expect graphics to be
there.

To be honest, if someone tells you the frames are the devils work,
ignore them. If someone tells you they are the best thing since sliced
bread, ignore them. A site should be designed around the end user.
Sometimes frames are the best option, sometimes not.

The art of web design is creating a site that works well and is easy to
use for any visitor. There are basic rules to help you achieve it, but
"Frames vs. No Frames" is merely "Linux vs. Windows" in
another guise
and has comparable numbers of fanatics in either camp.

If you are serious about wanting to create the perfect web experience,
first study HCI (Human Computer Interfacing). There are a number of
simple rules that have often been ignored in web design features.
Examples are why animation will distract the viewer from the content of
a page due to primal instincts within us and why pictures and text are
better positioned to one side or another.

Only then should you read web design articles. If you do want to buy a
book for web design, buy a couple of conflicting ones. The important
points are where they agree. The arguments are where you look to your
target audience to decide. If you read up on HCI first, you can merely
ignore anything they claim which contradicts that since HCI has been
around much longer than the web.

HCI can be quite an interesting subject if you get the right texts since
a lot of the discussion is about the human brain and it's reactions to
certain things. For our particular group's interests, there are a lot of
things which can be applied to HA implementation.

Mark.