Rupert
>I want to list an image with google, is it possible just to submit images to
>this search engine.
>
Have a look at google.com and click on the pictures bit and search for
a phrase.
I did a search for cherry tree and it appears that cherry tree is the
in the graphic name.
I would assume google crawls web pages and looks at the graphics
names. It might look at alt text too - who knows! If it doesn't
already then it may in the future.
--
http://www.voucherfreebies.co.uk
UK Shopping Discounts
> ....
> I would assume google crawls web pages and looks at the graphics
> names. It might look at alt text too - who knows! If it doesn't
> already then it may in the future.
I would hope that it's the *alt text* that it indexes on rather than the
filename! I'm normally quit careful about giving an alt description,
but I've never thought about going through all my images and renaming
the filenames so they appear on google!
--
_. _|._.o _.._ |_ _ |o __-+- _ ._
(_](_][ |(_][ ) [_)(_)||_) | (_)[ )
http://www.boliston.com/
I just did a quick search as I knew certain pictures of mine appeared in
the index and it seems to be the filename and URL that lists images -
obviously the filename being part of the URL - when I tried different
combinations, it was directory or URL names without spaces, that yielded
greatest results, as opposed to the name properly written.
E.g. I did a search on 'petercox' as one of my sites has this in the URL
and just about all of the site's image files were listed, several pages
of them, but a search for 'Peter Cox' yielded only a handful of my
photos, indicating to me that the ALT text didn't figure at all - as all
pictures of him would obviously feature his full name in the ALT text.
I was rather alarmed however to find some of my image files weren't
linked to my own site, but were being shown on a gay man's adoration
page! Time for a word about copyright, methinks!
--
Julia
Spread the World Web Site Design
http://www.spreadtheworld.com/
http://www.lancs-net.co.uk/
> I would hope that it's the *alt text* that it indexes on rather than the
> filename!
Some search engines index alt-text, and some don't. Last time I
looked, AltaVista did index it (try
http://www.altavista.com/sites/search/web?q=small.red.bullet
to see some horrible examples), whereas Google didn't, for example.
Small red bullet Response to Terrorism
Former Yugoslav Army Chief Dragoljub Ojdanic Sent
to The Hague Animated red arrow
Horizontal line
Google now seems to find a few matches - but I suspect that's a result
of mis-parsing defective HTML, rather than intentional indexing of ALT
text. Mostly, the matches that Google finds for "small red bullet"
are pages discussing HTML authoring techniques, rather than dumb sites
that used it as ALT text.
> I'm normally quit careful about giving an alt description,
Most ALT texts that consist of a "description" of the image are wrong;
there are some examples above.
As someone (not me) perceptively remarked on a WAI group: "think of
the text and the image as alternative representations of content".
The text should actually be doing the _job_ that the image was doing
on the page. Sometimes a short description of the significant
feature(s) of the image might be OK, but in my experience that seems
to be wrong far more often than it's right.
> Most ALT texts that consist of a "description" of the image are wrong;
> there are some examples above.
As most of my site consists of photo galleries created with the gPhoto
gallery creator, then I guess that leaving the default alt tags of
"001", "002", "003" etc would be better than trying to duplicate the
short description under each photo thumbnail into the alt text fields.
> "Alan J. Flavell" wrote:
>
> > Most ALT texts that consist of a "description" of the image are wrong;
> > there are some examples above.
>
> As most of my site consists of photo galleries created with the gPhoto
> gallery creator,
I said "most".
> then I guess that leaving the default alt tags of
> "001", "002", "003" etc would be better than trying to duplicate the
> short description under each photo thumbnail into the alt text fields.
You're obviously trying to be silly.
If the information is already there on the page, then duplicating it
as an ALT attribute is pointless. Either do something more
creative/constructive/useful, or nothing.
If the clear-text descriptions already link to the fullsize pics and
you've no additional information to offer, then what the heck is wrong
with alt="" ?
I simply don't see the point of this:
St Mary's Church
St Marys Church
001
Hammett Street
Hammett Street
002
St Mary's Church
St Mary's
003
The Parade
The Parade
004
Castle Bow
Castle Bow
005
and so on, and so on.
- other than the fact that without the ALT text you'd have nothing to
click on.
My preference would be to have the image and its accompanying text
both within the scope of the <a href="..."> , and use alt="" on the
img. Of course that wouldn't be legal if the image and text were in
separate table cells, as you have them right now: but what you're
doing with table cells is structurally unsound anyway, and what's the
odd <br> between friends?
best regards
> My preference would be to have the image and its accompanying text
> both within the scope of the <a href="..."> , and use alt="" on the
> img. Of course that wouldn't be legal if the image and text were in
> separate table cells, as you have them right now: but what you're
> doing with table cells is structurally unsound anyway, and what's the
> odd <br> between friends?
What do you mean by structurally unsound?
The page passes WC3 HTML & CSS validation.
I'm not sure if there are any other HTML photo gallery generators that
produce better code than the gPhoto generator?
> What do you mean by structurally unsound?
My apologies - due to initially mis-reading the source. It's no more
unsound than _any_ page which uses tables for layout.
> The page passes WC3 HTML & CSS validation.
But with respect, that wasn't the point. A page could still be
semantic nonsense and pass syntax validation.
> I'm not sure if there are any other HTML photo gallery generators that
> produce better code than the gPhoto generator?
The specific authoring tool isn't an issue that I'd want to get
dragged into discussing. I was looking at the generated markup, and
how I'd be inclined to simplify it. (There's no markup there that
couldn't be generated by e.g a simple Perl script - but as I say, that
isn't the issue I'm raising here.)
See this markup detail (outer table cell contents) like:
<td>
<table>
<tr><td><a href="picture-006.html"><img alt="The Castle"
src="thumbnail-006.jpg" border="1"></a></td></tr>
<tr><td>Taunton Castle</td></tr>
<tr><td>006</td></tr>
</table></td>
Even if one retains the original concept of the outer table for
layout, it seems to me that these cell contents can be simplified, and
IMHO* a number of accessibility issues addressed, if one changes it to
something like
<td>
<a href="picture-006.html"><img alt=" "
title="short description" longdesc="if.you.want.html"
src="thumbnail-006.jpg" border="1"><br>
Taunton Castle<br>
006
</a></td>
[*I'd obviously be interested to see reactions from others here who
take an interest in accessibility issues. In the fullness of time I'd
want to migrate to a more logical markup, using CSS positioning to get
the intended visual effects in the mainstream presentation situation.]
Naturally you'd want to supply appropriate CSS styles to get your
link texts presentation to suit your tastes. You could close the
anchor after the caption text, before the serial number, if you
prefer.
The text views on w3m would then be something like
St Mary's Hammett Street St Mary's The Parade Castle Bow
Church 002 003 004 005
001
as compared with something like
River Tone River Tone 2 BT exchange Tone view Sewer crossing
River Tone River Tone 2 BT exchange Tone View Tone
011 012 013 014 Sewer Crossing
015
as you're getting now; and this is much the same as what you get on
graphical browsers (Mozilla, Opera, etc.) if auto image loading is
disabled. As someone on a slow connection might be inclined to do.
Sure, the result isn't ideal on Lynx (it would be better without the
<br>, and suggesting the layout by other means), but I sense it would
be a lost cause trying to persuade you to do anything better for Lynx.
At least the simplified version looks like this kind of thing:
St Mary's Church
001
Hammett Street
002
St Mary's
003
The Parade
004
Castle Bow
005
as opposed to this kind of thing that you're getting now:
River Tone
River Tone
011
River Tone 2
River Tone 2
012
BT exchange
BT exchange
013
Tone view
Tone View
014
Nuff said about that.
By the way, it's illogical to specify the text size in pt units for
screen display. If you're determined to make the text fit the images,
then it might be defensible to size the text in px units. (More
logical, in theory, would be to size the thumbnails in em units, but
browsers aren't particularly good at supporting that!).
You know, your rubric "Please click on a thumbnail to view the
fullsize image" pre-supposes that the reader is using a specific user
interface, and is actually loading the thumbnails; it also says to the
user that you don't consider them competent to use what seems to be a
perfectly straightforward web page - which I think they're entitled to
find rather patronising. I would be inclined to trust them to
recognise links when they see them (providing you don't work _too_
hard to camouflage them).
Why not offer them a link to "Start the tour", instead of that
patronising rubric? The individual pics already have the relevant
Previous/Index/Next links in the content, so you're all set for that.
(I'm missing the LINK REL= entries for those references, though...).
cheers