Discussion:
[Pydotorg-redesign] CSS Layout Reservoir
Skip Montanaro
2003-10-07 12:15:59 UTC
Permalink
While looking for information on css @import stuff I came across this site:

http://www.bluerobot.com/web/layouts/

Folks interested in using css across a wide variety of web browsers might
want to take a look. The examples worked fine for me in Safari. In Lynx
and Netscape 4, the margin content appeared below the main body text. That
may be a problem if you expect your content to read roughly left-to-right,
top-to-bottom.

FYI.

Skip
Aahz
2003-10-07 20:15:35 UTC
Permalink
Post by Skip Montanaro
http://www.bluerobot.com/web/layouts/
Folks interested in using css across a wide variety of web browsers
might want to take a look. The examples worked fine for me in Safari.
In Lynx and Netscape 4, the margin content appeared below the main
body text. That may be a problem if you expect your content to read
roughly left-to-right, top-to-bottom.
That's easy enough to fix by putting the menu before the content in the
HTML itself.
--
Aahz (***@pythoncraft.com) <*> http://www.pythoncraft.com/

"It is easier to optimize correct code than to correct optimized code."
--Bill Harlan
Tim Parkin
2003-10-07 21:25:58 UTC
Permalink
Post by Aahz
Post by Skip Montanaro
The examples worked fine for me in Safari.
In Lynx and Netscape 4, the margin content appeared below the main
body text. That may be a problem if you expect your content to read
roughly left-to-right, top-to-bottom.
That's easy enough to fix by putting the menu before the content in the
HTML itself.
It would be of great benefit if possible, to have the main menu at the
bottom of the content and to provide a 'skip to navigation' link of some
sort at the top of the page. I have been told by the blind and partially
sighted users that I have had focus groups with, that the most annoying
thing is to have the menu read out on every page. The shortcut is
generally to offer a skip-menu feature which is ok but the best is to
have a 'skip to nav' with associated accesskey. I'll leave this one open
to debate as the html I've been building will work both ways. I've added
a reference regarding it from a very well thought of accessibility book
and also a link to Mark Pilgrim's Dive into Accessibility website.

Tim

http://diveintoaccessibility.org/day_10_presenting_your_main_content_fir
st.html

Joe Clark - Building Accessible (p150)

"Visitors who cannot readily skip all those links must either page
through them (as with Lynx, a text-only browser), or move the cursor to
select each link in turn (as with some mobility-impaired people using
any kind of browser), or attempt to skip the links using software
commands, if that is even possible. But in doing so, visitors may skip
links they actually want along with those they do not.

As a veteran Lynx user, let me tell you how it works. When presented
with dozens of navigation links, I press the spacebar over and over
again to bypass them a screenful at a time. If the first link of any
resulting screenful happens to be a text field (e.g., a search box), I
have to move the cursor off the field with the Downarrow or Tab keys,
then keep on paging through the links. Or I can guess how many
screenfuls of links are present and skip beyond that screenful: Typing
5P (p for page) brings me to the fifth screenful.

But I have it easy. While some mobility-impaired people do in fact
repeatedly press the Tab key to move around a page, some people with
more severe disabilities use adaptive technology that runs through a
sequence of possible actions ("switch access").You have to wait until
the keyboard action comes up, then home in on the Tab key (which itself
could take many steps), then actuate it. How would you like to go
through that 99 times just to skip nagavation?
Then what if the link you really want is actually near the bottom of the
page?
How long would you put up with that, if you don't already?"
Tim Parkin
2003-10-10 10:25:26 UTC
Permalink
Hi,

Heres the first stab at HTML for the python redesign.

http://pollenation.net/assets/public/draft-oct10.html

It's only a preliminary try but seems to work and also validates. It
still needs some accessibility additions (rel links and accesskeys and
also tabindex adjustments) but it's enough to start commenting on. I'm
very busy at work at the moment and will also be getting married at the
start of december. I'm going to have a crack at the home page HTML
sometime before then, if possible, and also try to get a draft release
of the supporting document for the proposal.

Tim
Richard Jones
2003-10-11 01:03:26 UTC
Permalink
Post by Tim Parkin
Heres the first stab at HTML for the python redesign.
http://pollenation.net/assets/public/draft-oct10.html
In general, I like this design - it's nice and clean. A few points:

1. Please remove the font sizing for the main body text
2. Not sure what the "dev" link in the top bar is...
3. "Download" and "documentation" are present twice in the sidebar. Surely
only one instance of each would be enough (I'd advocate the lower instance
of both)



Richard
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: signature
Url : http://mail.python.org/pipermail/pydotorg-redesign/attachments/20031011/5ae94914/attachment.bin
Tim Parkin
2003-10-11 11:55:55 UTC
Permalink
Post by Richard Jones
Post by Tim Parkin
Heres the first stab at HTML for the python redesign.
http://pollenation.net/assets/public/draft-oct10.html
1. Please remove the font sizing for the main body text
Hmm... not sure why you want this as the current text size is adjustable
via the browser menu *and* is also larger than the current default python
text size. I am going to be looking at adding an accessibility style
switcher that will allow you to change the default stylesheet, offering a
few options including a different size text. This will need javascript and
also cookies if you want it to remember the settings. A comparison of the
text sizes between redesign and existing site at default is here :-

Loading Image...

I may be adjusting body text colour to make it slightly darker (although
this is the recommended contrast for reading, I think some monitors will
show this at a lower contrast than others).
Post by Richard Jones
2. Not sure what the "dev" link in the top bar is...
Have a look at
Loading Image... for a
preliminary idea of what the developer link would be. It's intended as a
home for python developers (not python language developers, that confusion
needs to be addressed separately). Theres a lot of thinking behind doing
this that I won't go into here.
Post by Richard Jones
3. "Download" and "documentation" are present twice in the sidebar. Surely
only one instance of each would be enough (I'd advocate the lower
instance of both)
I wanted to have deep links that jump straight to popular pages. These can
be adjusted if we find out that people use other links. We don't want to
adjust the main menu all the time however. The quick links might change
depending on what section you are in (if you are in the downloads section
it may show the main downloads. In mailing lists, the most popular lists)
It's a way of getting quick links on the left hand bar, where people scan
first, without messing with the main menu.

Tim
Richard Jones
2003-10-12 23:38:12 UTC
Permalink
Skipped content of type multipart/mixed-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: signature
Url : http://mail.python.org/pipermail/pydotorg-redesign/attachments/20031012/2f68066a/attachment-0001.bin
Roy Smith
2003-10-11 23:08:53 UTC
Permalink
Post by Tim Parkin
Hi,
Heres the first stab at HTML for the python redesign.
http://pollenation.net/assets/public/draft-oct10.html
Overall, I like it a lot.

A minor technical glitch. When I viewed the page in Safari (Mac OSX),
if I made the window too narrow, the left-hand and right-hand text in
the nav bar run into each other and overlap. I've attached a small
screen shot to illustrate what I'm talking about.

It's unclear if this is really something worth worrying about -- if you
make anything too narrow, something's going to give somewhere.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: overlap.pdf
Type: application/pdf
Size: 8597 bytes
Desc: not available
Url : http://mail.python.org/pipermail/pydotorg-redesign/attachments/20031011/3de60ccf/overlap.pdf
-------------- next part --------------
Tim Parkin
2003-10-12 08:57:38 UTC
Permalink
Roy
Post by Roy Smith
Post by Tim Parkin
Heres the first stab at HTML for the python redesign.
http://pollenation.net/assets/public/draft-oct10.html
Overall, I like it a lot.
A minor technical glitch. When I viewed the page in Safari (Mac OSX),
if I made the window too narrow, the left-hand and right-hand text in
the nav bar run into each other and overlap. I've attached a small
screen shot to illustrate what I'm talking about.
It's unclear if this is really something worth worrying about -- if you
make anything too narrow, something's going to give somewhere.
It is inevitable that a website will become unusable at certain widths.
However I can make a small change to put a background on the utility nav
which will mean that it will slide over the breadcrumb trail as the window
gets smaller. As the utility nav is probably more important than the
breadcrumb, I think this is the correct choice, can't upload yet as I'm
away from civilisation but will do later. I've also played with a print
css sheet and a large text css sheet that should work quite niceley.

Tim

Tim
Laura Creighton
2003-10-12 09:19:20 UTC
Permalink
I like
Post by Tim Parkin
http://pollenation.net/assets/public/draft-oct10.html
a lot as well. Indeed, I think its wonderful.

However, you know me, always finding another nit .... :-)

I still would like the google text input window to be wider, all over
the top of the page, so that the search-widget-thing started, say
just over the I in DB-API.

I'm not so fond of the blue hook at the edge of the search-thing which
makes it look like a can-opener to me, but all I want to do is mention
that once -- I don't dislike it a lot, and not enough to fuss.

I will fuss about typing window size in search-things. I want more
room, or a good explanation why I shouldn't get it.

Laura
Tim Parkin
2003-10-12 14:56:18 UTC
Permalink
Post by Richard Jones
I like
Post by Tim Parkin
http://pollenation.net/assets/public/draft-oct10.html
a lot as well. Indeed, I think its wonderful.
:-)
Post by Richard Jones
..I still would like the google text input window to be wider, all over
the top of the page, so that the search-widget-thing started, say
just over the I in DB-API.
I don't know where that would be as the search box is referenced to the
rhs and the db-api to the lhs. I'll make it as big as I think the design
can get away with.. Just as a matter of interest, what is the problem
with having smaller search box windows?

Just comparing sizes of other major sites :-

Google : 208 px
BBC Homepage : 180 px
*Proposed Search Width : 170 px
Amazon.co.uk : 168 px
Php : 162 px
*Current Search Width : 154 px
Developer Works : 123 px
Java.sun.com : 112 px
News.bbc.co.uk : 95 px

I'll up it to 170px as any more than that and the python logo and the
search box will collide at 640 wide (a browser width near to that which
many people with larger monitors use whilst arranging multiple windows).
As a side not, this size is also 15% bigger than Jakob Nielsen's
Recommendations for search box width.
Post by Richard Jones
I'm not so fond of the blue hook at the edge of the search->thing which
makes it look like a can-opener to me, but all I want to do is mention
that once -- I don't dislike it a lot, and not enough to fuss.
You'll have to enlighten me as I can't work out what that might be..
Heres a screenshot of what I get.

Loading Image...

I presume it's a browser bug of some sort. If you can send a screenshot
of what you have a your browsers type, browser version and operating
system that would be great.
Post by Richard Jones
I will fuss about typing window size in search-things. I want more
room, or a good explanation why I shouldn't get it.
I've not problem about making it bigger in this case but would love to
know why you see it as so important. Taking the java page as an example,
what problems do you feel the small search window causes or what
limitations does it place on you? I just want to learn as a convincing
argument in any situation is always useful.

Tim
Laura Creighton
2003-10-12 15:52:07 UTC
Permalink
Post by Tim Parkin
Post by Richard Jones
I like
Post by Tim Parkin
http://pollenation.net/assets/public/draft-oct10.html
a lot as well. Indeed, I think its wonderful.
:-)
Post by Richard Jones
..I still would like the google text input window to be wider, all over
the top of the page, so that the search-widget-thing started, say
just over the I in DB-API.
I don't know where that would be as the search box is referenced to the
rhs and the db-api to the lhs. I'll make it as big as I think the design
can get away with.. Just as a matter of interest, what is the problem
with having smaller search box windows?
I get really uncomfortable when what I want to paste into the search
box doesn't fit. Or when I type in long things. Here I am being
really precise to limit my search nicely, and I cannot see ...

It really gives me the willies. Like being blindfolded.
Post by Tim Parkin
Just comparing sizes of other major sites :-
Google : 208 px
BBC Homepage : 180 px
*Proposed Search Width : 170 px
Amazon.co.uk : 168 px
Php : 162 px
*Current Search Width : 154 px
Developer Works : 123 px
Java.sun.com : 112 px
News.bbc.co.uk : 95 px
I'll up it to 170px as any more than that and the python logo and the
search box will collide at 640 wide (a browser width near to that which
many people with larger monitors use whilst arranging multiple windows).
As a side not, this size is also 15% bigger than Jakob Nielsen's
Recommendations for search box width.
Post by Richard Jones
I'm not so fond of the blue hook at the edge of the search->thing which
makes it look like a can-opener to me, but all I want to do is mention
that once -- I don't dislike it a lot, and not enough to fuss.
You'll have to enlighten me as I can't work out what that might be..
Heres a screenshot of what I get.
http://pollenation.net/assets/public/screenshot-oct12.gif
I presume it's a browser bug of some sort. If you can send a screenshot
of what you have a your browsers type, browser version and operating
system that would be great.
Yes its a bug. Its not happening at work, but on myu laptop at home.
Bug me to send you that if I forget.
Post by Tim Parkin
Post by Richard Jones
I will fuss about typing window size in search-things. I want more
room, or a good explanation why I shouldn't get it.
I've not problem about making it bigger in this case but would love to
know why you see it as so important. Taking the java page as an example,
what problems do you feel the small search window causes or what
limitations does it place on you? I just want to learn as a convincing
argument in any situation is always useful.
Having what I type in vanish 'underneath' makes my skin crawl.

I fear we may need a psychologist to explain why ...
Post by Tim Parkin
Tim
Thanks Tim.

Laura
Tim Parkin
2003-10-12 22:38:39 UTC
Permalink
Post by Laura Creighton
I get really uncomfortable when what I want to paste into the search
box doesn't fit. Or when I type in long things. Here I am being
really precise to limit my search nicely, and I cannot see ...
Well I made it bigger now...

The new version is at

http://pollenation.net/assets/public/draft-oct13.html

And has an alternate large text style sheet and try print previewing for
a nicer output. And the utility menu rides over the breadcrumb now too.

Tim
Richard Jones
2003-10-12 23:24:51 UTC
Permalink
Post by Tim Parkin
Post by Laura Creighton
I get really uncomfortable when what I want to paste into the search
box doesn't fit. Or when I type in long things. Here I am being
really precise to limit my search nicely, and I cannot see ...
Well I made it bigger now...
The new version is at
http://pollenation.net/assets/public/draft-oct13.html
And has an alternate large text style sheet and try print previewing for
a nicer output. And the utility menu rides over the breadcrumb now too.
Looking great!


Richard
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: signature
Url : http://mail.python.org/pipermail/pydotorg-redesign/attachments/20031013/22cb8b2d/attachment.bin
Laura Creighton
2003-10-13 00:15:46 UTC
Permalink
Post by Tim Parkin
Post by Laura Creighton
I get really uncomfortable when what I want to paste into the search
box doesn't fit. Or when I type in long things. Here I am being
really precise to limit my search nicely, and I cannot see ...
Well I made it bigger now...
The new version is at
http://pollenation.net/assets/public/draft-oct13.html
you sure Tim? On Mozilla on my laptop it looks the same. And I get
the can opener effect as well. Do you know any good way for me to
just save all the bits on my screen as a png or something so I can
mail it to you? I don't.

Is this how I ask my mozilla to tell me what version it is so I can tell
you? about: under help menu says ...

Mozilla 1.0
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020623 Debian/1.0.0-0.woody.1


Right now the S in Search is aligned just about between 'in' and 'relational'
and I would like to yank it all the way over to above the I in API.

Were I doing this for me, I would skip the 'search' if google allows that.
For me, everything that is not a button and a text window is wasted space.


I just tried it under konquerer, and it looks really bad. I've got to
learn a way to send you the bits of my screen so you can see it, but it
is nothing like what you are asking for ...

But now I need bed more than anything else.

Laura
Post by Tim Parkin
Tim
Matt Goodall
2003-10-13 01:31:54 UTC
Permalink
Post by Laura Creighton
Post by Tim Parkin
Post by Laura Creighton
I get really uncomfortable when what I want to paste into the search
box doesn't fit. Or when I type in long things. Here I am being
really precise to limit my search nicely, and I cannot see ...
Well I made it bigger now...
The new version is at
http://pollenation.net/assets/public/draft-oct13.html
you sure Tim? On Mozilla on my laptop it looks the same. And I get
the can opener effect as well. Do you know any good way for me to
just save all the bits on my screen as a png or something so I can
mail it to you? I don't.
It depends what you're using. GNOME 2.2 has a built in screenshot
utility (available as a panel button), KDE has something called
ksnapshot, Imagemagick includes a command line app called import.
Post by Laura Creighton
Is this how I ask my mozilla to tell me what version it is so I can tell
you? about: under help menu says ...
Mozilla 1.0
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020623 Debian/1.0.0-0.woody.1
Wow, are you really using Debian Stable? I thought everyone immediately
switched to testing ;-).

This must be a Mozilla 1.0 problem as the page works ok in 1.4 and seems
to work ok in 1.1 and 1.3. I may be able to install 1.0 to help Tim but
is it honestly worth it? Anyone know how many Moz 1.0 users the site gets?

[snip]
Post by Laura Creighton
But now I need bed more than anything else.
Me too!

Cheers, Matt
--
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenation.net
e: ***@pollenation.net
Laura Creighton
2003-10-13 08:10:49 UTC
Permalink
Post by Matt Goodall
It depends what you're using. GNOME 2.2 has a built in screenshot
utility (available as a panel button), KDE has something called
ksnapshot, Imagemagick includes a command line app called import.
Ksnapshot it is, then
Post by Matt Goodall
Post by Laura Creighton
Is this how I ask my mozilla to tell me what version it is so I can tell
you? about: under help menu says ...
Mozilla 1.0
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020623 Debian/
1.0.0-0.woody.1
Wow, are you really using Debian Stable? I thought everyone immediately
switched to testing ;-).
Well, at work, I use unstable. But this is my laptop, which has to
work on airplanes, trains, and other places where I cannot just
get the odd file when something breaks ....
Post by Matt Goodall
This must be a Mozilla 1.0 problem as the page works ok in 1.4 and seems
to work ok in 1.1 and 1.3. I may be able to install 1.0 to help Tim but
is it honestly worth it? Anyone know how many Moz 1.0 users the site gets
?
Don't do it for me -- or for the can opener.
Post by Matt Goodall
[snip]
Post by Laura Creighton
But now I need bed more than anything else.
Me too!
Cheers, Matt
Ok, awake now.

Here are 2 pngs, the first from my laptop mozilla, the second from my
laptop konqueror.

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/png
Size: 90951 bytes
Desc: not available
Url : Loading Image...
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: image/png
Size: 85073 bytes
Desc: not available
Url : Loading Image...
-------------- next part --------------
And I think I understand what is happening. Tim has made something I
am going to call a Google-widget. It is that pretty grey box with the
Google advertising and Search and the Go button, and of course the
all important text window. Now, as far as I am concerned, anything
but the Go button and the text window is 'chart junk'. (for those
of you who are reading this, and haven't read Tufte, I am being
constructively critical here. And reading Tufte is a joy). However,
'Google' is a legal requirement, and, I don't know, maybe there are some
people on the planet who don't recognise a search window when they see it.

My problem is that Tim's Google-widget is of fixed size. You resize
the window, and the Google widget moves back and forth, stuck to the
right hand border of the window, and eating up the space between the
'Python' and its left most border as needed. I want a different
Google-widget, one that automatically sizes its text window to take up
all the space that is available. I want the left most border of the
Google-widget glued to some space just to the right of 'Python' the logo,
the right most border glued to the rhs, wherever that is, and the whole
thing to float in width as I resize it.

That way I can paste things like:

'Haskell monadic modular interpreters'

into the window, and see whether we have had a discussion about them
someplace, before I start another round of 'making Python more
Haskell-like' or whatever is my favorite poison of the moment. (I am
assuming that I will get to access the mailing lists via this
Google-widget as well, and I think that a 'advanced search' click link
for people who only want to search the mailing lists, say, would be
useful.)

Am I making any sense?

Laura
Tim Parkin
2003-10-13 07:57:48 UTC
Permalink
No attachment!! :-) I'm sure they'll arrive soon...

----------------------------------------------
Tim Parkin
Managing Director
Pollenation Internet Ltd
www.pollenation.net
m : 07980 59 47 68
t : 01132 25 25 00

-----Original Message-----
From: Laura Creighton [mailto:***@strakt.com]
Sent: 13 October 2003 07:21
To: Matt Goodall
Cc: Laura Creighton; ***@pollenation.net; pydotorg-***@python.org;
***@strakt.com
Subject: Re: [Pydotorg-redesign] Draft HTML for redesign proposal
Post by Matt Goodall
It depends what you're using. GNOME 2.2 has a built in screenshot
utility (available as a panel button), KDE has something called
ksnapshot, Imagemagick includes a command line app called import.
Ksnapshot it is, then
Post by Matt Goodall
Post by Laura Creighton
Is this how I ask my mozilla to tell me what version it is so I can
tell
Post by Matt Goodall
Post by Laura Creighton
you? about: under help menu says ...
Mozilla 1.0
Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.0) Gecko/20020623
Debian/
Post by Matt Goodall
1.0.0-0.woody.1
Wow, are you really using Debian Stable? I thought everyone immediately
switched to testing ;-).
Well, at work, I use unstable. But this is my laptop, which has to
work on airplanes, trains, and other places where I cannot just
get the odd file when something breaks ....
Post by Matt Goodall
This must be a Mozilla 1.0 problem as the page works ok in 1.4 and
seems
Post by Matt Goodall
to work ok in 1.1 and 1.3. I may be able to install 1.0 to help Tim but
is it honestly worth it? Anyone know how many Moz 1.0 users the site
gets
Post by Matt Goodall
?
Don't do it for me -- or for the can opener.
Post by Matt Goodall
[snip]
Post by Laura Creighton
But now I need bed more than anything else.
Me too!
Cheers, Matt
Ok, awake now.

Here are 2 pngs, the first from my laptop mozilla, the second from my
laptop konqueror.
Laura Creighton
2003-10-13 08:31:45 UTC
Permalink
Post by Tim Parkin
No attachment!! :-) I'm sure they'll arrive soon...
----------------------------------------------
Tim Parkin
Managing Director
Pollenation Internet Ltd
www.pollenation.net
m : 07980 59 47 68
t : 01132 25 25 00
They are delayed awaiting moderator approval because they are big.
I have to go to the dentist. Back later.

Laura
Tim Parkin
2003-10-13 07:58:48 UTC
Permalink
It may be better to mail the attachments directly to me for now. That
goes for everyone else too.. I'll put them up on a site for
demonstrations if they're not instantly fixable.

Tim
Tim Parkin
2003-10-13 09:41:32 UTC
Permalink
Ruchard
Attached is the comparison on my system. You can see that on my system,
the
specifying pf 100% (as in the current pydotorg site) and 86% (as
specified in
the new stylesheet) actually results in text sizes reflecting those
specifications, unlike on your system, which somehow renders the 86%
font
larger than your default font size...
Could you send your operating system type and version and also your
browser type and version. Also if you've made any changes to the
properties for your browser those would be good too as it looks very
strange. It would also be great if you could take a screenshot of this
page for me and send to my email address (we don't want to flood the
list with screenshots). What makes me think something strange is
happening is that 99% of browsers use serif fonts for default text
(times-roman usually) and yours looks like a sans serif.

http://pollenation.net/assets/public/font-test.html

Cheers

Tim
Tim Parkin
2003-10-13 22:23:59 UTC
Permalink
Post by Laura Creighton
Post by Tim Parkin
http://pollenation.net/assets/public/draft-oct13.html
I just tried it under konquerer, and it looks really bad. I've got to
learn a way to send you the bits of my screen so you can see it, but it
is nothing like what you are asking for ...
I can tweak things a little when we get some feedback as theres no point
making many small tweaks only to find out that something needs changing
and have to go through the whole process again.

The konquerer version you are using in debian stable is 2.2.2 which is 2
years old and did have many css rendering problems. I'll still have a
try to get it usable if possible but can't promise anything. How would
it seem if I used a javascript method for checking for konqueror and
used an alternative stylesheet. This stylesheet could be selected
manualy for people with javascript disabled. This would offer a custom
stylesheet in the same vein as netscape 4. I think this method could be
used to extend the usabilty of the site to close to 100% of users,
albeit with a manual intervention needed for a very small percentage
(automatically applied for those with javascript enabled).

I've also been trying out something else which is not polished yet but
wanted to know what people think.

http://pollenation.net/assets/public/collapsetest

The left hand menu has a collapse text. I don't currently like this
positioning. As it would only really be used at the "document" level, I
thought it could be added to the rhs document level navigation box. It
would also have an associated 'hot key' so power users wouldn't need a
menu item at all. Oh and theres not 'back to default size' on the make
text larger link.

Tim
Tim Parkin
2003-10-14 08:29:04 UTC
Permalink
Post by Laura Creighton
Post by Tim Parkin
http://pollenation.net/assets/public/draft-oct13.html
I just tried it under konquerer, and it looks really bad. I've got to
learn a way to send you the bits of my screen so you can see it, but it
is nothing like what you are asking for ...
I can tweak things a little when we get some feedback as theres no point
making too many small tweaks only to find out that something needs
changing
and have to go through the whole process again.

The konquerer version you are using in debian stable is 2.2.2 which is 2
years old and did have many css rendering problems. I'll still have a
try to get it usable if possible but can't promise anything. How would
it seem if I used a javascript method for checking for konqueror 2 and
used an alternative stylesheet. This stylesheet could be selected
manualy for people with javascript disabled. This would offer a custom
stylesheet in the same vein as netscape 4. I think this method could be
used to extend the usabilty of the site to close to 100% of users,
albeit with a manual intervention needed for a very small percentage
(automatically applied for those with javascript enabled).

I've also been trying out something else which is not polished yet but
wanted to know what people think.

http://pollenation.net/assets/public/collapsetest

The left hand menu has a collapse text. I don't currently like this
positioning. As it would only really be used at the "document" level, I
thought it could be added to the rhs document level navigation box. It
would also have an associated 'hot key' so power users wouldn't need a
menu item at all. Oh and theres not 'back to default size' on the make
text larger link.

Tim
Laura Creighton
2003-10-14 13:45:30 UTC
Permalink
My laptop died yesterday. The screen is shot. I'm sorry, but I cannot
help you any more .... at least until I can see if I can get it repaired.
SInce I hate the thing, I am sort of hoping _not_ ...

YOu may have fixed the rendering problem by removing the one user
who had it. Boy, do I need to learn that trick for some customers of
mine. :-)

Laura
Post by Tim Parkin
Post by Laura Creighton
Post by Tim Parkin
http://pollenation.net/assets/public/draft-oct13.html
I just tried it under konquerer, and it looks really bad. I've got to
learn a way to send you the bits of my screen so you can see it, but it
is nothing like what you are asking for ...
I can tweak things a little when we get some feedback as theres no point
making too many small tweaks only to find out that something needs
changing
and have to go through the whole process again.
The konquerer version you are using in debian stable is 2.2.2 which is 2
years old and did have many css rendering problems. I'll still have a
try to get it usable if possible but can't promise anything. How would
it seem if I used a javascript method for checking for konqueror 2 and
used an alternative stylesheet. This stylesheet could be selected
manualy for people with javascript disabled. This would offer a custom
stylesheet in the same vein as netscape 4. I think this method could be
used to extend the usabilty of the site to close to 100% of users,
albeit with a manual intervention needed for a very small percentage
(automatically applied for those with javascript enabled).
I use mozilla and firebird for me, so I don't care about konqueror myself.
But I think that it is well used, so I wanted to see how it looked, and
it wasn't good ...
Post by Tim Parkin
I've also been trying out something else which is not polished yet but
wanted to know what people think.
http://pollenation.net/assets/public/collapsetest
Interesting, now I get the canopener effect at work too. You not
getting it?
Post by Tim Parkin
The left hand menu has a collapse text. I don't currently like this
positioning. As it would only really be used at the "document" level, I
thought it could be added to the rhs document level navigation box. It
would also have an associated 'hot key' so power users wouldn't need a
menu item at all. Oh and theres not 'back to default size' on the make
text larger link.
Tim
Really making this window very narrow is --- odd. But I'm not saying that
it is wrong. more 'interesting' to me .... i might get perverse and like it.

Laura
Michael Geary
2003-10-14 09:50:41 UTC
Permalink
Hi Tim!

You know what I'm going to say, don't you? :-)

Yeah, it's the text.

It's a lot better than it was before--the font size is completely reasonable
now.

(As a reminder, I use a Thinkpad A30p with a 15" 1600x1200 display, i.e. 133
pixels per inch--a typical pixel density for high-end notebooks. I'm running
Windows XP with ClearType enabled, using Mozilla Firebird with 18 pixel
Georgia as my default font and a minimum font size of 14 pixels.)

Here's what I don't like, though:

* Not black. I know that you consider less contrast to be easier on the
eyes, but I have to tell you that with my eyes, on my computer, it just
ain't so. I spend all day, every day, looking at text, and I've tried lots
of variations of font and color. Black on white is by far the best for me.
This is especially true now that I use ClearType--black on white gives
ClearType the most room to work with, and colored text suffers especially
because it severely limits how much subpixel antialiasing can be done. But
even before I used ClearType, black on white was the best on my LCD
displays.

* Arial font. I saw your discussion on your site about Arial vs. Verdana,
but for me Verdana is really quite a bit more readable. Also, I think that
the fact that Verdana distinguishes between upper case "I" and lower case
"l" is a big point in its favor on a site where that distinction may be
significant.

* But why force a sans-serif font at all? I know that traditionally, serif
fonts have suffered from poor readability on a computer display, and the
traditional default browser font has been a serif font, so web designers
have forced sans-serif fonts in an attempt to gain more readability. But for
me, the combination of the Georgia font, ClearType, and the high-resolution
LCD panel have changed that completely. For the first time, I can really
make use of the benefits of a serif font (which as you know is more readable
than sans-serif in print).

Do me a favor and beg, borrow, or steal a notebook with a high-resolution
LCD like mine--a 15" 1600x1200 or a 14.1" 1400x1050 will do the trick--and
take a look at this:

Loading Image...

The text on the left is from your sample, and the text on the right is from
the original python.org page using my default browser font settings (black
Georgia), both rendered in Mozilla Firebird. I don't know how they will
compare on a lower resolution LCD or on a CRT (ClearType might mess both of
them up on a CRT), but on my high-resolution LCD there's no contest.

But what the heck, I could live with a sans-serif font if it were black
Verdana. :-)

Sorry if I sound like a one-trick pony, but when I'm reading text, text is
what I care about.

Thanks for listening to my input!

-Mike
Laura Creighton
2003-10-14 13:53:20 UTC
Permalink
I really like serifs on my fonts as well.

Laura
Skip Montanaro
2003-10-14 16:27:46 UTC
Permalink
Michael> Do me a favor and beg, borrow, or steal a notebook with a
Michael> high-resolution LCD like mine--a 15" 1600x1200 or a 14.1"
Michael> 1400x1050 will do the trick--and take a look at this:

Michael> http://www.geary.com/python/pyfonts.png

Can you add the other two combinations to that png (low-contrast serif font
and high-contrast sans serif font) just for completeness?

Thx,

Skip
Anna Ravenscroft
2003-10-14 17:42:17 UTC
Permalink
Post by Michael Geary
* Not black. I know that you consider less contrast to be easier on the
eyes, but I have to tell you that with my eyes, on my computer, it just
ain't so. I spend all day, every day, looking at text, and I've tried lots
of variations of font and color. Black on white is by far the best for me.
My feeling exactly. I consider the font to be eyestraining very quickly. The
size is fine (most websites I tend to zoom to 110% just to read the font; I
don't need to do that here), but the impact is too low. Please, can you put
in a new ribbon and retype it so it's dark enough to read?

Oh, sorry - it just brought me flashbacks of old days... ;-) Seriously though,
it reminds me of a copy done on a copier running out of toner.

I can read it but I feel like I need to strain a lot... and it's easy for my
eyes to get distracted to other parts of the page. I keep finding my eyes
skipping off the text to other parts of the page...

The font and size are fine... I just really dislike the low contrast.

Thanks for listening

Anna
running Opera 7.11
--
There is a type 3 error in which your mind goes totally blank whenever you try
to remember which is which of type 1 and type 2.
-Richard Dawkins
Michael Geary
2003-10-14 10:25:16 UTC
Permalink
(Read this after my other message for it to make the most sense...)

p.s. Realizing that you may prefer the sans-serif font, I did another
side-by-side example, this time with 16-pixel Verdana on the right:

Loading Image...

I do like the bit of extra line spacing the text on the left has, but I like
the font and color better on the right.

And just so it's all in one place, here's the one with 18-pixel Georgia
again:

http://www.geary.com/python/pyfonts.png

Again, these use ClearType and are meant for a high-resolution LCD.

Also, it goes without saying--make sure your browser doesn't shrink the
images when you look at these! :-)

Thanks,

Mike
Tim Parkin
2003-10-14 12:09:12 UTC
Permalink
Post by Michael Geary
* Not black. I know that you consider less contrast to be easier on the
eyes, but I have to tell you that with my eyes, on my computer, it just
ain't so. I spend all day, every day, looking at text, and I've tried
lots
Post by Michael Geary
of variations of font and color. Black on white is by far the best for
me.
Post by Michael Geary
This is especially true now that I use ClearType--black on white gives
ClearType the most room to work with, and colored text suffers
especially
Post by Michael Geary
because it severely limits how much subpixel antialiasing can be done.
But
Post by Michael Geary
even before I used ClearType, black on white was the best on my LCD
displays.
I think perhaps you're right when you're using clear type on an LCD
display 15" monitor at 1600x1200. What I propose doing is adding an
alternate sytlesheet which removes all styling from the body content.
This means that for 99% of people who don't set their default fonts (and
probably don't know how to) and who would see times roman if they didn't
(one of the most unreadable fonts when non-antialiased, which is how ie
shows it) get to see an optimum view. For those power users who have
exceptional circusmstances, there will be an alternate stylesheet
available, which is how the w3c proposes handling situations similar to
this . I agree that the contrast is a little low and am now implementing
a stylesheet specifically for people who have dyslexic style reading
difficulties. I don't want to use full black on white as a lot of people
have complained that this is too stark for them so I'm upping the
contrast to 80%. Then, people who see this as too dark can set their own
contrast by using the default view stylesheet. People who see this as
too light can also do the same.
Post by Michael Geary
* Arial font. I saw your discussion on your site about Arial vs.
Verdana,
Post by Michael Geary
but for me Verdana is really quite a bit more readable. Also, I think
that
Post by Michael Geary
the fact that Verdana distinguishes between upper case "I" and lower
case
Post by Michael Geary
"l" is a big point in its favor on a site where that distinction may be
significant.
I don't think the l versus I is a big issue, context provides incredible
clues and we don't read letter by letter any way. As an example, does
this look like the plural or virus? VIRll... Or did it read as Virll? I
can imagine a situation in an olde world drama about roman gods "So lo,
daughter of Inachus" but maybe this is most unlikely. I personally think
that verdana is too extended (opp. Condensed) and also I chose arial to
match the opus/officana font used in the Python logo, not the main
consideration but it works niceley nevertheless. If a majority of people
complain about the use of arial and all state a preference for verdana I
shall obviously change it. However I will make sure that people who wish
to set their own custom styles will have the opportunity.
Post by Michael Geary
* But why force a sans-serif font at all? I know that traditionally,
serif
Post by Michael Geary
fonts have suffered from poor readability on a computer display, and
the
Post by Michael Geary
traditional default browser font has been a serif font, so web
designers
Post by Michael Geary
have forced sans-serif fonts in an attempt to gain more readability.
But for
Post by Michael Geary
me, the combination of the Georgia font, ClearType, and the
high-resolution
Post by Michael Geary
LCD panel have changed that completely. For the first time, I can
really
Post by Michael Geary
make use of the benefits of a serif font (which as you know is more
readable
Post by Michael Geary
than sans-serif in print).
Because the vast majority of people use browser/os combinations that
don't support clear type, that don't anti alias fonts and that default
to times-roman, it would be remiss to configure for this as the default.
But I do agree, good size monitors with high dpi's and a system like
cleartype is more like print and hence serif fonts work very well for
narrative.
Post by Michael Geary
Do me a favor and beg, borrow, or steal a notebook with a
high-resolution
Post by Michael Geary
LCD like mine--a 15" 1600x1200 or a 14.1" 1400x1050 will do the
trick--and
I have a sony vaio PCG-C1MHP which is an 8.9" UW-SXGA which equates to
160dpi (your monitors are 133dpi and 124dpi respectively). Like I said,
if I activate cleartype and boost font size it looks great. If I don't
do this then sans-serif looks better.
Post by Michael Geary
Sorry if I sound like a one-trick pony, but when I'm reading text, text
is
Post by Michael Geary
what I care about.
No problem, I appreciate the detailed reply and have made some changes
that go toward what you are after. I hope these are enough to balance
the satisfaction of the individual power user and the majority passive
browser. (I'll upload the changes this evening)

Tim
Tim Parkin
2003-10-14 12:29:05 UTC
Permalink
Michael
Post by Michael Geary
p.s. Realizing that you may prefer the sans-serif font, I did another
Just my comparison of serif vs sans-serif as would be viewed in majority
of browsers compared with viewed with cleartype.

Loading Image...

Loading Image...

I've also shown the contrast that I'd prefer to use.

Loading Image...

Tim
Tim Parkin
2003-10-14 12:44:51 UTC
Permalink
Final note...

Verdana is actually a bigger font than arial, to this effect, I've
modified the font sizing to reflect this.

Loading Image...

Tim
Roy Smith
2003-10-14 13:53:03 UTC
Permalink
Post by Tim Parkin
I've also shown the contrast that I'd prefer to use.
http://pollenation.net/assets/public/python-fonts-contrast.gif
Could just be my 40-something eyes, but I find the black text
significantly easier to read than the 80% text. I'm looking at this on
a 17" CRT set to 24-bit color, 1280 x 960, 75 Hz, which works out to
about 90 dpi.
Skip Montanaro
2003-10-14 16:38:29 UTC
Permalink
Post by Tim Parkin
http://pollenation.net/assets/public/python-fonts-contrast.gif
Roy> Could just be my 40-something eyes, but I find the black text
Roy> significantly easier to read than the 80% text.

Ditto here (Mac 15" PowerBook, 1280x854, nearly 50-something).

Skip
Tim Parkin
2003-10-14 13:59:01 UTC
Permalink
Post by Laura Creighton
My laptop died yesterday. The screen is shot.
YOu may have fixed the rendering problem by removing the one user
who had it. Boy, do I need to learn that trick for some customers of
mine. :-)
Took me a lot of hacking to make that happen ;-)
Post by Laura Creighton
I use mozilla and firebird for me, so I don't care about
konqueror myself. But I think that it is well used, so I
wanted to see how it looked, and it wasn't good ...
I think It's 2.2.2 as I said, so it won't look v good
Post by Laura Creighton
Interesting, now I get the canopener effect at work too. You not
getting it?
Not unless I use konquerer 2.2.2
Post by Laura Creighton
Really making this window very narrow is --- odd. But I'm not saying
that
Post by Laura Creighton
it is wrong. more 'interesting' to me .... i might get perverse and
like it.
I'm sorry I probably don't understand the context. You mean making the
browser narrow makes it odd or the screen goes narrow when you collapse
the menu?

Tim

Ps cheers for feedback
Tim Parkin
2003-10-14 14:32:09 UTC
Permalink
So far, my favorite is the antialias Verdana. It shows up clear and
clean on
Opera 7.11. The nonantialias versions are worst - with Arial slightly
worse
than Verdana...
Unfortunately most users (most windows and some linux) won't have a
choice of using anti-alias. The argument about legibility of fonts can
be confused as there are two types of reading. One if word reading,
where the individual letter clarity is important, and for this verdana
wins hands down. The other is sentence reading, where word shapes are
more important as the brain doesn't process letter by letter. For this,
in my opinion, arial wins as it's condensed face creates word shapes
that are more in line with the typical narrative word shapes used in
print etc. Like Michael says, if we were to have very highresolution
with excellent anti-aliasing, I would choose a serif font for body
content. Unfortunately we have low resolution with no anti-alias being
the default, majority view and for this, I feel arial wins due to it's
familiar rendering of word shapes. This is also why you don't tend to
see body content in 'Avant Garde' but you do see a lot in 'helvetica'
where the screen versions of these are, loosely, 'verdana' and 'arial'
respectivley (arial was ms's rendering of helvetica for screen and
verdana was the same but optimised for ultra low resolution,
non-antialiased.

At the end of the day this probably says it all...

http://usability.gov/guidelines/fonts.html

http://www.acm.org/sigchi/chi95/proceedings/intpost/tst_bdy.htm

In summary..

"Research shows no reliable differences in reading speed or user
preferences between 10-point Times Roman, Georgia serif fonts,
Helvetica, or Verdana sans serif font"

Although contrast did produce some differences :-

http://hubel.sfasu.edu/research/AHNCUR.html

In this case it was noted that in no survey did black on white produce
the best response times. This was typically produced at lower contrasts,
although in these case the background was adjusted rather than the
foreground. On my monitor here (22" Iiyama CRT) black on white makes my
eyes water and I prefer a low contrast foreground. Perhaps I suffer from
a dyslexia type disorder, but I find 80% contrast so much more readable
than 100%. If it turns out I am the exception then I'll have to change
it but if what I suffer at 100% contrast is repeated for a large
percentage of the population then I'd be very reluctant to give in on
this point.

Tim
Michael Geary
2003-10-15 07:07:11 UTC
Permalink
Post by Skip Montanaro
Michael> Do me a favor and beg, borrow, or steal a notebook with a
Michael> high-resolution LCD like mine--a 15" 1600x1200 or a 14.1"
Michael> http://www.geary.com/python/pyfonts.png
Can you add the other two combinations to that png (low-contrast
serif font and high-contrast sans serif font) just for completeness?
It looks like Tim's additional samples covered those--or are there still
some cases missing?

-Mike
Michael Geary
2003-10-15 07:26:23 UTC
Permalink
From: Tim Parkin
Unfortunately most users (most windows and some linux) won't have
a choice of using anti-alias.
Hmm... Just about all Windows users have the option of using anti-aliasing.
Of course, only Windows XP users will have ClearType available, but nearly
all new Windows machines come with XP.
The argument about legibility of fonts can
be confused as there are two types of reading. One if word reading,
where the individual letter clarity is important, and for this verdana
wins hands down. The other is sentence reading, where word shapes are
more important as the brain doesn't process letter by letter. For this,
in my opinion, arial wins as it's condensed face creates word shapes
that are more in line with the typical narrative word shapes used in
print etc.
Interesting. You really find Arial to be more readable than Verdana for body
text? For my eyes, Verdana wins at small type sizes whether it's
anti-aliased or not.

At larger type sizes--once the stems get to be two pixels thick--I don't
mind either one.

Like Michael says, if we were to have very highresolution
with excellent anti-aliasing, I would choose a serif font for body
content. Unfortunately we have low resolution with no anti-alias being
the default, majority view and for this, I feel arial wins due to it's
familiar rendering of word shapes. This is also why you don't tend to
see body content in 'Avant Garde' but you do see a lot in 'helvetica'
where the screen versions of these are, loosely, 'verdana' and 'arial'
respectivley (arial was ms's rendering of helvetica for screen and
verdana was the same but optimised for ultra low resolution,
non-antialiased.
At the end of the day this probably says it all...
http://usability.gov/guidelines/fonts.html
Ha! They *say* it doesn't matter, but they *use* Verdana!!! ;-)

(Well, they use Verdana for *most* of their body copy, but they mix in a bit
of Arial too--really a bad idea to switch back and forth like that.)
http://www.acm.org/sigchi/chi95/proceedings/intpost/tst_bdy.htm
That's a very old paper that predates Verdana, anti-aliased text, and LCD
displays. I'd tend to take anything it says with a grain of salt.
In summary..
"Research shows no reliable differences in reading speed or user
preferences between 10-point Times Roman, Georgia serif fonts,
Helvetica, or Verdana sans serif font"
Although contrast did produce some differences :-
http://hubel.sfasu.edu/research/AHNCUR.html
OUCH! That paper really is painful to look at. Green on yellow? What were
they thinking?
In this case it was noted that in no survey did black on white produce
the best response times. This was typically produced at lower contrasts,
although in these case the background was adjusted rather than the
foreground. On my monitor here (22" Iiyama CRT) black on white makes my
eyes water and I prefer a low contrast foreground. Perhaps I suffer from
a dyslexia type disorder, but I find 80% contrast so much more readable
than 100%. If it turns out I am the exception then I'll have to change
it but if what I suffer at 100% contrast is repeated for a large
percentage of the population then I'd be very reluctant to give in on
this point.
Forgive a dumb question, but what kind of refresh rate are you running your
monitor at? I remember back in the day when all I had was a 60 Hz monitor, I
went to great lengths to turn down the contrast. I even recall patching the
Windows 1.x/2.x display drivers to force white backgrounds to be light gray.
But when I finally got up to a decent refresh rate, black on white started
looking fine. And it's always looked great on LCDs. (One observation
here--we just got in some new computers and monitors at Adobe, and I noticed
that just about everyone who got a new monitor chose an LCD.)

Naw, you're a designer (and a darn good one, despite our differences on type
style <g>), you wouldn't have a monitor with a bad refresh rate--would you?
:-)

Also, just out of curiosity, what resolution do you run the 22" monitor at?

Thanks!

-Mike
Michael Geary
2003-10-15 07:37:24 UTC
Permalink
From: Tim Parkin
Verdana is actually a bigger font than arial, to this effect, I've
modified the font sizing to reflect this.
http://pollenation.net/assets/public/python-fonts-arialverdana.gif
Am I seeing things, or is the Verdana in that sample less tall than the
Arial? It looks like the widths are about the same, and Verdana is a wider
font than Arial, so it would make sense that it's less tall.

Both are too small for easy reading on my display--but of course they would
end up bigger when rendered at my 120 DPI setting.

-Mike
Tim Parkin
2003-10-15 09:10:28 UTC
Permalink
Post by Michael Geary
From: Tim Parkin
Unfortunately most users (most windows and some linux) won't have
a choice of using anti-alias.
Hmm... Just about all Windows users have the option of using
anti-aliasing.
Post by Michael Geary
Of course, only Windows XP users will have ClearType available, but
nearly
Post by Michael Geary
all new Windows machines come with XP.
This still leaves the majority of users having no idea or not having the
capability of antialiased fonts. I can't find the option of switching
antialiasing on in windows 2000 or NT, and I asked a few friends to
switch it on for their windows XP boxes and most couldn't find how.
Those that did find it rightly pointed out that it's meant for LCD
monitors.
Post by Michael Geary
Naw, you're a designer (and a darn good one, despite our differences on
type
Post by Michael Geary
.style <g>), you wouldn't have a monitor with a bad refresh rate-
-would you? :-)
22" Colour Balanced CRT with 100Hz or 150Hz Refresh depending on res.
Post by Michael Geary
(One observation
here--we just got in some new computers and monitors at Adobe, and I
noticed
Post by Michael Geary
that just about everyone who got a new monitor chose an LCD.)
I guarantee any designers worth there wages will choose crt's and more
than likely colour balanced Lacie monitors (typically with a calibratpr
to ensure no colour drift over time) eg
http://www.lacie.com/products/product.htm?id=10036 I couldn't afford one
:-(

I do think LCD monitors are great however, they're in the majority as
far as screen sales go but remember that the majority of monitors out
there are only 1024x768 or 1280x1024 so boosting font size to get multi
pixel thick lines isn't an option even if most users could work out how
to do it or even realise it was possible.
Post by Michael Geary
Also, just out of curiosity, what resolution do you run the 22" monitor
at?
1600x1200 or 2048x1536
Post by Michael Geary
and a darn good one, despite our differences on type
style <g>
Actually I'm an electrical engineer, I user to work in R&D modelling
magnetic and electric fields for general electric. The design side
started when my fellow mathmeticians didn't know what to do to market
the software we'd written. I still don't think of myself as a 'designer'
more of a 'web page synthesist' or 'average plagiarist'. Our designer
(who originated the idea) is a different story. He's lectured in
branding and design and top German universities, created websites for
NatWest, Yahoo and HM the Queen and now is a design director for Ornge
Communications. I won't turn the compliment down though (and we're
getting along niceley this time :-)

Tim
Tim Parkin
2003-10-15 09:24:43 UTC
Permalink
Post by Michael Geary
From: Tim Parkin
Verdana is actually a bigger font than arial, to this effect, I've
modified the font sizing to reflect this.
http://pollenation.net/assets/public/python-fonts-arialverdana.gif
Am I seeing things, or is the Verdana in that sample less tall than the
Arial? It looks like the widths are about the same, and Verdana is a
wider
Post by Michael Geary
font than Arial, so it would make sense that it's less tall.
No you're right.. I tried to get the fonts the same size (the size I
would use if I used either one) and adjusted the line spacing to be the
same so we could have an equal test of readability. I used the
quantifier that the same amount of content should fit in the same area.
This meant that verdana worked out smaller height. Obviously if I made
the verdana lower case characters the same height as the arial lower
cases, the verdana full character height would be substantially bigger
than arial (by about 13%).
Post by Michael Geary
Both are too small for easy reading on my display--but of course they
would
Post by Michael Geary
end up bigger when rendered at my 120 DPI setting.
I understand that and promise I won't say "IT's AN IMAGE!!" ;-)

Tim
Roy Smith
2003-10-15 12:56:41 UTC
Permalink
Post by Michael Geary
Post by Tim Parkin
http://hubel.sfasu.edu/research/AHNCUR.html
OUCH! That paper really is painful to look at. Green on yellow? What
were
they thinking?
They were thinking about their own data. They measured "Reaction Time"
for various fonts and color combintations and found Green on Yellow to
be the one that gave the fastest reaction times.

I wonder about their experimental population, though. In the first
paragraph of the paper, they point out that age and color-blindness are
important factors, but they specifically excluded color-blind people
from the cohort, as well as anybody with less than 20/20 corrected
vision. They make no mention of the age distribution of the subjects
they used, but if it's done in a university setting, I wouldn't be
surprised if it was mostly 20-somethings (grad students hungry to make
$50 for an hour of their time).

Color blindness affects a significant portion of the population.
According to http://www.hhmi.org/senses/b130.html, 7% of American males
are red/green deficient.

Their experimental technique was to present short (150 or so word)
passages and ask the subject to find a given "shape" word (i.e.
"square", "circle", "triangle", etc). I'm not convinced that scanning
for a specific word is anything like reading for comprehension. On the
other hand, for something like a reference manual, scanning for a given
word may indeed be what you want to optimize.

There is also another thing to consider. Even assuming that some case
could be made that Green on Yellow increases reading comprehension, it
is (IMHO) ugly. One of the goals of the web site is to be visually
pleasing, and Green on Yellow doesn't meet that goal.
Tim Parkin
2003-10-15 16:31:33 UTC
Permalink
Post by Roy Smith
Color blindness affects a significant portion of the population.
According to http://www.hhmi.org/senses/b130.html, 7% of American males
are red/green deficient.
Unfortunately when you have no other research, you try to re-evaluate
the statistical evidence given in the research you do have. In this
case, there was obviously a reason why they concluded that lower
contrasts gave better readability results, what we have to do is say.
Well there was either very little difference or there was a slight
readability improvement for lower contrasts.
Post by Roy Smith
Their experimental technique was to present short (150 or so word)
passages and ask the subject to find a given "shape" word (i.e.
"square", "circle", "triangle", etc). I'm not convinced that scanning
for a specific word is anything like reading for comprehension. On the
other hand, for something like a reference manual, scanning for a given
word may indeed be what you want to optimize.
I won't go into detail about how people read as there is an amazing
amount of research out there but it's got a lot more to do with scanning
word shapes than recognising letters.
Post by Roy Smith
There is also another thing to consider. Even assuming that some case
could be made that Green on Yellow increases reading comprehension, it
is (IMHO) ugly. One of the goals of the web site is to be visually
pleasing, and Green on Yellow doesn't meet that goal.
Green and yellow is horrible as we know which is why I've not used it.
However having done a little more research about scotopic sensitivity or
Meares Irlen Syndrome, I've found that this supposedly affects a
significant proportion of the population (between 4% and 12% depending
on where you look) This isn't a subset of dyslexia and can appear on
it's own (although between 4% and 10% for dyslexia). There is a also a
strong link between Aspergers/Autism and with autism and aspergers
rising rapidly and aspergers itself being a very common condition among
programmers (geeks) I think maybe it's worth considering as long as it
doesn't affect any other large group of people. I've yet to have anyone
say they have difficulty reading the 80% contrast text. Maybe a large
survey would be good? I'll look into it some more and maybe do some user
testing when possible.

Tim

Tim
Roy Smith
2003-10-15 16:49:05 UTC
Permalink
I've yet to have anyone say they have difficulty reading the 80%
contrast text.
Did you not notice these two messages?

http://mail.python.org/pipermail/pydotorg-redesign/2003-October/
000595.html
http://mail.python.org/pipermail/pydotorg-redesign/2003-October/
000600.html
Tim Parkin
2003-10-15 17:00:52 UTC
Permalink
Post by Roy Smith
I've yet to have anyone say they have difficulty reading the 80%
contrast text.
Did you not notice these two messages?
Yes and they didn't say that had difficulty reading the 80% contrast,
they said they thought it was easier to read the 100% contrast. I think
th other person ditto'd. I find cooking beans on toast easier than
cooking a fry up. Doesn't mean I find fry ups difficult. I find it
easier to walk than it is to run, again running isn't difficult. My
point is exactly that people aren't finding it difficult to read at 80%
wheras some people DO find it difficult to read at 100%.

Tim
Roy Smith
2003-10-15 17:13:42 UTC
Permalink
OK, let me be more clear. I find it difficult to read the 80% text.

If you would like to conclude that a small enough number of people said
they find the 80% text difficult to read that you're going to use it
anyway, OK, I can live with that. But please don't dismiss my
statement as having not been said because it doesn't fit with your
conclusion.
Post by Tim Parkin
Post by Roy Smith
I've yet to have anyone say they have difficulty reading the 80%
contrast text.
Did you not notice these two messages?
Yes and they didn't say that had difficulty reading the 80% contrast,
they said they thought it was easier to read the 100% contrast. I think
th other person ditto'd. I find cooking beans on toast easier than
cooking a fry up. Doesn't mean I find fry ups difficult. I find it
easier to walk than it is to run, again running isn't difficult. My
point is exactly that people aren't finding it difficult to read at 80%
wheras some people DO find it difficult to read at 100%.
Tim
Tim Parkin
2003-10-15 18:01:34 UTC
Permalink
Post by Roy Smith
OK, let me be more clear. I find it difficult to read the 80% text.
If you would like to conclude that a small enough number of people said
they find the 80% text difficult to read that you're going to use it
anyway, OK, I can live with that. But please don't dismiss my
statement as having not been said because it doesn't fit with your
conclusion.
I didn't dismiss your statement as not having been said and don't know
how you think I did. I said that nobody had mentioned they had
difficulty with the 80% contrast text and they hadn't.

The reason I find it difficult to believe that a significant portion of
the python population would have difficulty reading the 80% text is that
a huge amount of the current python site is rendered at 63% contrast and
some very important parts at 43% contrast. In fact I would have to
assume that, if you feel it is difficult to read 80% contrast text then
the whole of the web would be a pain for you to use.

This is getting a little silly. I'm trying to build a website that works
for the majority of people, I can honestly say that nearly every website
in the world used a significant amount of content at less than 70%
contrast (including the rnib, the oldie, uk government pension guides).
I'll be making sure that there will be an unstyled css option that you
can choose if you have serious problems reading 80% contrast text. I
don't feel that this is the general problem you feel it to be.

Please show me some empirical evidence that a significant proportion of
the population will have problems with 80% contrast and then I can add
it to the large body of evidence I've used to work out what the optimum
contrast for screen readability is.

I think we've stated enough opinion for to benefit the mailing list so
if we need to carry on, can we do so off list. (I'll even restate now
that you're welcome to your opinion about the difficulties in reading
80% text and I respect that you have this affects you personally and so,
if it becomes apparent that this afflicts a significant number of python
users, current or future, and the adjustments nescessary would not be
detrimental to a greater number of people in the similar way or a
smaller number of people in a greater way, then I'd agree that the
changes would be worth making)

Tim
Tim Parkin
2003-10-15 18:38:35 UTC
Permalink
Here are the last test cases for people to work out what they like the
most, please post me off list with your opinions and I'll compile them
:-

Loading Image...
Loading Image...
Loading Image...
Loading Image...

Tim
Tim Parkin
2003-10-17 10:54:44 UTC
Permalink
Hi all,

After a *huge* discussion about contrasts and fonts, we've come to some
conclusions and some queries.

Contrast wise the voting has been 50/50 between people liking 80% and
100%. A surprising number of people responded with different contrasts
depending on the font. A sample demographic split

1) designers preferred 80%
2) oldies slightly preferred 100%
3) youngsters preferred 80%
4) the balance was split between people professing difficulty/tiredness
reading 80% and 100%

Overall I think providing an 80% (or maybe 85% as this seemed to be a
demarcation point for some people) and allowing people to choose an
unstyled version would be appropriate at this point. This decision is by
no means set in stone as it would be trivial to change at any point in
the future.

As for fonts, although a lot of people liked verdana, almost as many
preferred arial. We extended slightly to include various other common
fonts as follows

Tahoma
Trebuchet MS
Lucida Sans
Bitstream Vera Sans
Helvetica
MS Sans Serif

Tahoma, Trebuchet, MS Sans Serif and Helvetica were awful. This left
Lucida Sans and Bitstream Vera Sans. Out of these Bitstream Vera Sans
won hands down. We had a surprising discovery. On closer inspection we
discovered that Bitstream Vera Sana was antialiasing at smaller sizes
(which no other windows fonts we looked at did). This in combination
with the quality of the font (and it's license) means we would recommend
using the following order of fonts for the website

"Bitstream Vera Sans", Verdana, "Lucida Sans", Geneva, Helvetica,
sans-serif

And I'd even go further to suggest that users of the Python site
download the bitstream vera font if they want to get high quality body
text.

A sample of the results are here :-

Loading Image...
Loading Image...
Loading Image...
Loading Image...

Cheers

Tim
Anna Ravenscroft
2003-10-17 14:34:06 UTC
Permalink
Post by Tim Parkin
Hi all,
After a *huge* discussion about contrasts and fonts, we've come to some
conclusions and some queries.
<snip>
Post by Tim Parkin
Tahoma, Trebuchet, MS Sans Serif and Helvetica were awful. This left
Lucida Sans and Bitstream Vera Sans. Out of these Bitstream Vera Sans
won hands down. We had a surprising discovery. On closer inspection we
discovered that Bitstream Vera Sana was antialiasing at smaller sizes
(which no other windows fonts we looked at did). This in combination
with the quality of the font (and it's license) means we would recommend
using the following order of fonts for the website
"Bitstream Vera Sans", Verdana, "Lucida Sans", Geneva, Helvetica,
sans-serif
And I'd even go further to suggest that users of the Python site
download the bitstream vera font if they want to get high quality body
text.
You've done a great job Tim and I'm really impressed by your responsiveness to
the community. Thank you!

Anna Ravenscroft
Guido van Rossum
2003-10-17 15:44:32 UTC
Permalink
Post by Tim Parkin
As for fonts, although a lot of people liked verdana, almost as many
preferred arial.
I do most of my browsing on a Linux box (Red Hat 7.3), and so do many
other existing and potential Python users. Those fonts don't exist
there. Therefore I'm not sure that picking fonts should be too much
of an issue at this point.

--Guido van Rossum (home page: http://www.python.org/~guido/)
Matt Goodall
2003-10-17 16:26:40 UTC
Permalink
Post by Guido van Rossum
Post by Tim Parkin
As for fonts, although a lot of people liked verdana, almost as many
preferred arial.
I do most of my browsing on a Linux box (Red Hat 7.3), and so do many
other existing and potential Python users. Those fonts don't exist
there. Therefore I'm not sure that picking fonts should be too much
of an issue at this point.
Actually, both Verdana and Arial are available for X aslong as you can
use TrueType fonts. I don't know about RH7.3 specifically.

http://www.gnome.org/fonts/
http://corefonts.sourceforge.net/

I work on a Debian testing/unstable box and I have many of the fonts Tim
mentioned installed. They're much nicer to work with that the standard X
bitmap fonts too.

In any case, a bit further down that message, Tim suggested the
following font order: "Bitstream Vera Sans", Verdana, "Lucida Sans",
Geneva, Helvetica,
sans-serif.

Lucida (not Lucida Sans) is a standard X11 font, Bitstream and Verdana
can be installed; Verdana and Lucida Sans are standard Windows fonts,
Bitstream can be installed; Geneva is a standard Mac font; Helvetica is
available on everything but Windows although I'm sure it's available.
There's the sans-serif fall back at then end anyway.

That list of fonts covers a lot of OSs although fonts and font sizes are
a right pain on the web and it's quite difficult to cover all bases
effectively.

Cheers, Matt
--
Matt Goodall, Pollenation Internet Ltd
w: http://www.pollenationinternet.com
e: ***@pollenation.net
Tim Parkin
2003-10-17 17:03:33 UTC
Permalink
Post by Guido van Rossum
Post by Tim Parkin
As for fonts, although a lot of people liked verdana, almost as many
preferred arial.
I do most of my browsing on a Linux box (Red Hat 7.3), and so do many
other existing and potential Python users. Those fonts don't exist
there. Therefore I'm not sure that picking fonts should be too much
of an issue at this point.
The linux platform core fonts don't really provide a choice that people
can have strong opinions about. Lucida on X11 is awful leaving only
Helvetica. Other than this, many people install extra fonts on their
system and, in doing so, install the windows core fonts or the bitstream
vera fonts.

The reason for the conversation was that these fonts present quite
diverse ranges of condensed/expanded, x-heights and densities. Arial and
verdana represent the extremes of both of these and are the most common
fonts available to Python users (over 70% of users on Windows systems,
18% on Linux and 1% on Mac). It may seem anal but a little feedback can
have a siginificant affect, if only for those users who have issues with
reading.

Tim
Guido van Rossum
2003-10-17 17:14:00 UTC
Permalink
Post by Tim Parkin
The linux platform core fonts don't really provide a choice that people
can have strong opinions about. Lucida on X11 is awful leaving only
Helvetica. Other than this, many people install extra fonts on their
system and, in doing so, install the windows core fonts or the bitstream
vera fonts.
I don't believe that "many people install extra fonts." The reality
is that if it's not on the dominant vendor's CD, it's not going to be
on most boxes.
Post by Tim Parkin
The reason for the conversation was that these fonts present quite
diverse ranges of condensed/expanded, x-heights and densities. Arial and
verdana represent the extremes of both of these and are the most common
fonts available to Python users (over 70% of users on Windows systems,
18% on Linux and 1% on Mac).
Are these Python stats or font availability stats? Where did you get
them?
Post by Tim Parkin
It may seem anal but a little feedback can have a siginificant
affect, if only for those users who have issues with reading.
Yes, I'm glad you care about those people too.

--Guido van Rossum (home page: http://www.python.org/~guido/)
Tim Parkin
2003-10-17 17:27:13 UTC
Permalink
Post by Guido van Rossum
Post by Tim Parkin
The linux platform core fonts don't really provide a choice that
people
Post by Guido van Rossum
Post by Tim Parkin
can have strong opinions about. Lucida on X11 is awful leaving only
Helvetica. Other than this, many people install extra fonts on their
system and, in doing so, install the windows core fonts or the
bitstream vera fonts.
I don't believe that "many people install extra fonts." The reality
is that if it's not on the dominant vendor's CD, it's not going to be
on most boxes.
Just from a survey on

http://www.codestyle.org/css/font-family/sampler-UnixResults.shtml

all the people I know who use unix have the windows fonts installed,
most also have the bitstream fonts. This cross section is mostly
programmers but a few designers and a couple of desktop office users
although like many 'friend surveys' will have it's own demographic tilt.

As an aside, it is intended that KDE would use Bitstream Vera range as
part of it's core standard fonts. Other distributions would include
Freetype, XFT2 and X Render extensions of the XFree86 project, Pango,
KDE and Trolltechs QT
Post by Guido van Rossum
Post by Tim Parkin
The reason for the conversation was that these fonts present quite
diverse ranges of condensed/expanded, x-heights and densities. Arial
and verdana represent the extremes of both of these and are the most
common fonts available to Python users (over 70% of users on Windows
systems, 18% on Linux and 1% on Mac).
Are these Python stats or font availability stats? Where did you get
them?
These are stats for september 2003 from www.python.org/wwwstats

Tim
Guido van Rossum
2003-10-17 17:37:26 UTC
Permalink
Post by Tim Parkin
Post by Guido van Rossum
I don't believe that "many people install extra fonts." The reality
is that if it's not on the dominant vendor's CD, it's not going to be
on most boxes.
Just from a survey on
http://www.codestyle.org/css/font-family/sampler-UnixResults.shtml
all the people I know who use unix have the windows fonts installed,
most also have the bitstream fonts. This cross section is mostly
programmers but a few designers and a couple of desktop office users
although like many 'friend surveys' will have it's own demographic tilt.
Trust me on this one. We're geeks. Inevitably, our friends are
mostly geeks. This is *not* a representative survey. :-)
Post by Tim Parkin
As an aside, it is intended that KDE would use Bitstream Vera range as
part of it's core standard fonts. Other distributions would include
Freetype, XFT2 and X Render extensions of the XFree86 project, Pango,
KDE and Trolltechs QT
For most of my colleagues here at Elemental Security, it won't exist
until it comes from a Red Hat CD. The same is true for our enterprise
customers.
Post by Tim Parkin
Post by Guido van Rossum
Post by Tim Parkin
The reason for the conversation was that these fonts present quite
diverse ranges of condensed/expanded, x-heights and densities. Arial
and verdana represent the extremes of both of these and are the most
common fonts available to Python users (over 70% of users on Windows
systems, 18% on Linux and 1% on Mac).
Are these Python stats or font availability stats? Where did you get
them?
These are stats for september 2003 from www.python.org/wwwstats
I was afraid so. That's 70% Windows *downloads*, not *users*. Since
Python is pre-installed on *all* Linux distros (not just Red Hat :-),
Linux users don't have to download anythign in order to use Python.
(The same will be true for Mac OS X Panther, but that's not out yet.)

--Guido van Rossum (home page: http://www.python.org/~guido/)
Skip Montanaro
2003-10-17 17:53:06 UTC
Permalink
Guido> (The same will be true for Mac OS X Panther, but that's not out
Guido> yet.)

Actually, it's also true of Mac OS X Jaguar (10.2). /usr/bin/python is 2.2.

Skip
Guido van Rossum
2003-10-17 17:59:25 UTC
Permalink
Post by Skip Montanaro
Guido> (The same will be true for Mac OS X Panther, but that's not out
Guido> yet.)
Actually, it's also true of Mac OS X Jaguar (10.2). /usr/bin/python is 2.2.
Yes, but that's not a framework build (or something -- it's not quite
as complete as what Panther will have; Jack and Just know the
details).

--Guido van Rossum (home page: http://www.python.org/~guido/)
Tim Parkin
2003-10-17 18:09:15 UTC
Permalink
Post by Tim Parkin
Just from a survey on
http://www.codestyle.org/css/font-family/sampler-UnixResults.shtml
trust me on this one. We're geeks. Inevitably, our friends are
mostly geeks. This is *not* a representative survey. :-)
Almost all my Python using friends are geeks too, proudly so. My old
boss wasn't a geek but he kept complaining that his linux pc wouldn't
show websites like his laptop did so we installed ttf windows fonts. I
would imagine a lot of Python users would be geeks too, whether or not
they'd like to be called that is a different matter :-)

The survey I used, though flawed, does seem to be good enough to show
general trends and does reflect the experience of myself and that of a
few of my colleagues. Some environments have strict guidelines on
platform setup (we weren't allowed to install new fonts at a dot com I
worked for, explaining I was designing for a client who wouldn't like us
to change there logo finally got round that one).
Post by Tim Parkin
Post by Guido van Rossum
Are these Python stats or font availability stats? Where did you
get
Post by Tim Parkin
Post by Guido van Rossum
them?
These are stats for september 2003 from www.python.org/wwwstats
I was afraid so. That's 70% Windows *downloads*, not *users*. Since
Python is pre-installed on *all* Linux distros (not just Red Hat :-),
Linux users don't have to download anythign in order to use Python.
(The same will be true for Mac OS X Panther, but that's not out yet.)
The stats I picked of the site weren't for downloads, they were for
general browsing. I was presuming that this would be the best source of
information about what the balance of browser/operating system support
for the new site would be. I'd love to know if there are any other stats
available, actually that raises a question I was going to ask a while
back. Can I get access to the raw 'combined' logs for the python server.
I know they'd be huge but I've got some excellent data mining software
that I could use to get some good information on usage.

I think the conversation about fonts was pretty much over anyway, any
further feedback was coming directly back to me, the last post was just
a summary of findings. We could always move back to logos and straplines
;-)

Tim
Guido van Rossum
2003-10-17 18:19:21 UTC
Permalink
Post by Tim Parkin
Almost all my Python using friends are geeks too, proudly so. My old
boss wasn't a geek but he kept complaining that his linux pc wouldn't
show websites like his laptop did so we installed ttf windows fonts. I
would imagine a lot of Python users would be geeks too, whether or not
they'd like to be called that is a different matter :-)
The survey I used, though flawed, does seem to be good enough to show
general trends and does reflect the experience of myself and that of a
few of my colleagues. Some environments have strict guidelines on
platform setup (we weren't allowed to install new fonts at a dot com I
worked for, explaining I was designing for a client who wouldn't like us
to change there logo finally got round that one).
Well that argument definitely won't fly for the casual visitors of
python.org whome we want to please (if I recall the "goals" discussion
on the marketing list correctly).
Post by Tim Parkin
Post by Guido van Rossum
Post by Tim Parkin
Post by Guido van Rossum
Are these Python stats or font availability stats? Where did you
get
Post by Guido van Rossum
Post by Tim Parkin
Post by Guido van Rossum
them?
These are stats for september 2003 from www.python.org/wwwstats
I was afraid so. That's 70% Windows *downloads*, not *users*. Since
Python is pre-installed on *all* Linux distros (not just Red Hat :-),
Linux users don't have to download anythign in order to use Python.
(The same will be true for Mac OS X Panther, but that's not out yet.)
The stats I picked of the site weren't for downloads, they were for
general browsing. I was presuming that this would be the best source of
information about what the balance of browser/operating system support
for the new site would be.
Yes, probably. (Interesting that the stats match the 70% download
stats for Windows.)
Post by Tim Parkin
I'd love to know if there are any other stats
available, actually that raises a question I was going to ask a while
back. Can I get access to the raw 'combined' logs for the python server.
I know they'd be huge but I've got some excellent data mining software
that I could use to get some good information on usage.
I suggest you write to Thomas Wouters (thomas at xs4all.net).

--Guido van Rossum (home page: http://www.python.org/~guido/)

Roy Smith
2003-10-17 18:17:03 UTC
Permalink
Post by Tim Parkin
all the people I know who use unix have the windows fonts installed,
most also have the bitstream fonts.
Just another data point...

At home I've got Mac/OSX and RedHat 8.0. At work I've got a variety of
Linux and Solaris systems. Heck, I've even got a windows laptop.
None of the above have any special fonts installed (unless they got
installed by default along with an application).

I install stuff I need to do my work. Databases, new versions of
python or gcc, etc. But, fonts are just something I expect to be
there. At this point in my life (remember, I'm the guy with the
40-something eyeballs :-)), when it comes to computers, I want a tool,
not a hobby.

For what it's worth, I checked my two home boxes. OSX comes with
Verdana and Ariel, RedHat 8.0 doesn't.
Loading...