[Pydotorg-redesign] I think that the current site structure needs to be scrapped altogether.

[Laura Creighton]

Right now, it looks as if somebody has said 'the most important thing
is to have all the links that are important up there on the first
page'. Pack 'em in. This is an efficient design for people who
already know what they want, more or less.

It is, however, an intimidating design for those who are coming here
for the first time, and I thought those were the ones that we were
designing for.

They want gateways to 'education' 'business use' and 'useful press
clippings you can use' and 'FIRST TIME HERE' and so on and so forth.
What they want to do is explore, and layouts for exploration look
very different from layouts for sitting down and getting a lot o
work done right now. The latter reads 'pure hacker', and there is
indeed a place for such a page -- but down someplace in the
'for Programmers' section.

If we don't produce some sort of consensus on how we are going to
limit the information shown on the first page, we are going to end
up with a very pretty site that still scares the managers, scientists,
educators, and journalists away.

Laura, who is still too busy to do anything until end of October.

[amk at amk.ca ()]

Post by Laura Creighton
Right now, it looks as if somebody has said 'the most important thing
is to have all the links that are important up there on the first
page'. Pack 'em in.

Correct; the links are jumbled together in no apparent order. I'd like to
at least put them into a sensible order, without making wider revisions such
as changing the graphic design or logical structure of the site.

Post by Laura Creighton
If we don't produce some sort of consensus on how we are going to
limit the information shown on the first page, we are going to end
up with a very pretty site that still scares the managers, scientists,
educators, and journalists away.

That's fine, but we shouldn't let future possibilities deter us from making
trivial improvements *right now*. Isn't this the primary lesson of XP? Not
to say, "Oh, we'll clean this up later", but to say "Let's clean it up
*now*".

The redesign effort is looking increasingly like the types-sig to me. The
types-sig expended lots of discussion time on complicated questions and
*never* *produced* *anything*. No add-on tools, no Python patches, no
specification, no plan, *nothing*. All that discussion time was therefore
wasted. If people deferred working on type-checking because "the types-sig
will have a solution soon", the community lost out because we never saw the
tools those people would have produced.

Similarly, redesign chatter circulates and recirculates endlessly around
marketing issues such as "What classes of users use the site?" and "Is Java
or VB Python's competition?" and "Pictures: threat or menace?". If we defer
making any improvements to the existing site because "there will be a
redesign in six months", what if the redesign discussion collapses? For six
months the site will have impeded users unnecessarily; what's the benefit of
that?

At this point I've had an improved sidebar arrangement for around two
months, and it's *still* not live. I'm becoming increasingly tempted to say
to hell with it and just make the revised version live. If people hate it,
that's fine; just go ahead and make fixes to CVS. At least there will be
*some* improvement to the site as a result, as I can't see anyone wanting to
go back to the existing sidebar arrangement.

--amk

[Aahz]

Post by amk at amk.ca ()
At this point I've had an improved sidebar arrangement for around two
months, and it's *still* not live. I'm becoming increasingly tempted
to say to hell with it and just make the revised version live. If
people hate it, that's fine; just go ahead and make fixes to CVS.
At least there will be *some* improvement to the site as a result,
as I can't see anyone wanting to go back to the existing sidebar
arrangement.

+1

I'm perfectly happy making the changes I requested myself. Thank you
for all the time and effort you've put into this.

[Dylan Reinhardt]

Post by amk at amk.ca ()

Post by Laura Creighton
Right now, it looks as if somebody has said 'the most important thing
is to have all the links that are important up there on the first
page'. Pack 'em in.

Correct; the links are jumbled together in no apparent order. I'd like to
at least put them into a sensible order, without making wider revisions such
as changing the graphic design or logical structure of the site.

You totally missed the point. In truth, most of those links shouldn't
be there _at all_. Order is a far less important concern.

If I were re-designing the front page, I would eliminate the top menu
(except for a search box) and pare down the links on the side bar to
about eight major topics:
- Introducing Python
- Download
- Documentation
- Community
- Projects
- Events
- Success Stories
- Links

Then I'd push the "news" currently in the content area to a right-hand
column and use the main content area to give an overview of Python.
Each of the main topic areas might "explode" out 6-8 main subtopics on
pages where they are being browsed. It's conventional, but there's a
reason conventional design got that way: it works.

Radical simplicity is what's needed here. There is no good reason to
confront a beginner with links to PEPs, PSF, PBF, Zope, bugs, cvs, PyPI,
etc, etc. They just aren't useful or relevant until you get more of the
big picture. Every one of the links on the front page can be
intuitively organized into one of the topic headings I listed.

Post by amk at amk.ca ()

Post by Laura Creighton
If we don't produce some sort of consensus on how we are going to
limit the information shown on the first page, we are going to end
up with a very pretty site that still scares the managers, scientists,
educators, and journalists away.

That's fine, but we shouldn't let future possibilities deter us from making
trivial improvements *right now*. Isn't this the primary lesson of XP? Not
to say, "Oh, we'll clean this up later", but to say "Let's clean it up
*now*".

XP is not a catch-all excuse for half-measures. The primary lesson of XP
is to deliver the product the *customer* wants, not what the programmer
wants to create.

We're currently ignoring 90% of our customers and it shows. The web
site is virtually useless as a tool for reaching or converting anyone
who isn't already bought in. We're holding potential converts at arm's
length. The site is unfriendly, cold, and unattractive. Nothing about
it is going to enhance our reputation or build our user base.

Post by amk at amk.ca ()
Similarly, redesign chatter circulates and recirculates endlessly around
marketing issues such as "What classes of users use the site?" and "Is Java
or VB Python's competition?" and "Pictures: threat or menace?". If we defer
making any improvements to the existing site because "there will be a
redesign in six months", what if the redesign discussion collapses? For six
months the site will have impeded users unnecessarily; what's the benefit of
that?

Well, here's the problem. The people who are willing to expend work on
a real redesign are not the same people as those who are currently in
charge of www.python.org. We're left spinning our wheels because the
people who have decision-making authority appear to think that the site
is actually pretty good. By almost anything other than open source
hacker standards, it is not.

That is why I am arguing in favor of making the decision to *do*
something instead of futzing around what what we already have. It will
be a lot easier to maintain momentum on a real redesign if there is a
solid commitment to working on it. Investing time in small cosmetic
changes directly undercuts that effort.

Dylan

[amk at amk.ca ()]

Post by Dylan Reinhardt
Well, here's the problem. The people who are willing to expend work on
a real redesign are not the same people as those who are currently in
charge of www.python.org. We're left spinning our wheels because the
people who have decision-making authority appear to think that the site
is actually pretty good. By almost anything other than open source
hacker standards, it is not.

To my mind, it's completely unclear who the people who have decision-making
authority *are*, which is a fundamental problem. This stalls small changes
as well as the larger redesign.

Anyway, there's nothing preventing anyone from grabbing portions of
python.org with wget, making an improved prototype and putting it on some
random IP address. Fred even made a (somewhat controversial) tarball a
little while ago; that could be repeated. So, make a prototype!

Post by Dylan Reinhardt
solid commitment to working on it. Investing time in small cosmetic
changes directly undercuts that effort.

This is not a zero-sum game. My investing time in the existing site doesn't
take time away from those people who are willing to work on a major
redesign.

--amk

[Aahz]

Post by amk at amk.ca ()
To my mind, it's completely unclear who the people who have
decision-making authority *are*, which is a fundamental problem. This
stalls small changes as well as the larger redesign.

Supposedly it's the PWC, but we haven't really gotten off the ground
yet. My priority has been PyCon more than the website recently (after
catching up from the metric tonne of e-mail from malware and its
byproducts).

Post by amk at amk.ca ()
This is not a zero-sum game. My investing time in the existing site
doesn't take time away from those people who are willing to work on a
major redesign.

+1

[Aahz]

Post by Dylan Reinhardt
We're currently ignoring 90% of our customers and it shows. The
web site is virtually useless as a tool for reaching or converting
anyone who isn't already bought in. We're holding potential converts
at arm's length. The site is unfriendly, cold, and unattractive.
Nothing about it is going to enhance our reputation or build our user
base.

That's opinion, not fact. We still have not achieved consensus about
design goals, not even talking about the extent to which the current
design does or does not achieve those goals.

Post by Dylan Reinhardt
Well, here's the problem. The people who are willing to expend work on
a real redesign are not the same people as those who are currently in
charge of www.python.org. We're left spinning our wheels because the
people who have decision-making authority appear to think that the site
is actually pretty good. By almost anything other than open source
hacker standards, it is not.

Again opinion. If you look at successful sites such as Amazon.com or
cnn.com, I think it's hard to argue that python.org does significantly
worse at the organizational level.

[Dylan Reinhardt]

Post by Aahz

Post by Dylan Reinhardt
The site is unfriendly, cold, and unattractive.
Nothing about it is going to enhance our reputation or build our user
base.

That's opinion, not fact.

It is an informed opinion based on relevant experience. I spent several
years creating marketing and communications materials whose results were
measurable. What's your opinion based on?

Post by Aahz

Post by Dylan Reinhardt
We're left spinning our wheels because the
people who have decision-making authority appear to think that the site
is actually pretty good. By almost anything other than open source
hacker standards, it is not.

Again opinion.

Again, *informed* opinion.

Post by Aahz
If you look at successful sites such as Amazon.com or
cnn.com, I think it's hard to argue that python.org does significantly
worse at the organizational level.

First off, CNN's success has almost nothing to do with its web site and
Amazon's success has been propped by significant inflows from outside
their core business. Amazon's longer-term success is yet to be proven.

But even taken as given that they are good examples, I'm astonished you
can't tell how much better both of those sites are at presenting and
managing information. Both sites manage 1000's of times as much content
as www.python.org, yet each one is easier to navigate.

CNN's top bar consists of a search box, some branding and advertising
and *nothing* else. CNN enables browsing by sorting hundreds of pages
of new content per day into 14 key categories. The remaining 85% of the
screen is link-dense, but notice what an effort is made to put a *small*
number of links "above the fold" and how most of the really link-dense
content is stuffed below the fold and *sorted* into headers that mirror
the topic browsing buttons featured higher-up in the page.

Amazon makes a far more cluttered first impression the last couple
years, but it's still worlds better than the Python web site. Though
the first page is a bit scatter-shot, the site quickly adapts to your
usage and narrows its focus. If you allow cookies, that is. :-)

Also, I'll point out something *both* those sites make extensive use of:
graphics. Another thing: typography. Maybe they all look the same
when you're browsing in lynx, but for the other 99.99% of us using the
*graphical* web, there's a world of difference in terms of quality and
usability.

$.02,

Dylan

[Guido van Rossum]

I recommend that everyone who hasn't done so yet, read the book "Don't
make me Think" by Steve Krug.

http://www.amazon.com/exec/obidos/tg/detail/-/0789723107/ref=lib_dp_TFCV/103-8580480-8104610?v=glance&s=books&vi=reader#reader-link

I agree with almost everything Krug says, and wish I could put them in
practice for python.org -- alas, my time is needed in language design
and implementation, where there aren't quite so many people who are
better than me, so don't expect me to help with the redesign.

IMO the first and foremost recommendation from the book for python.org
is to put a search box on the front page. We shouldn't have to wait
to do that until the full redesign is ready. (BTW, the current
Ultraseek search engine also uses Python. :-)

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

[Aahz]

Post by Guido van Rossum
IMO the first and foremost recommendation from the book for python.org
is to put a search box on the front page.

I'll do that as soon as AMK installs his new design.

[Roy Smith]

Post by Dylan Reinhardt
We're currently ignoring 90% of our customers and it shows.

Putting on my product management hat for a moment, do we have any data
on what sells and what doesn't, i.e. which links on the current page
get used a lot and which don't?

For the ones that don't get used a lot, we need to figure out why.
Either they're not of interest to the majority of the people who walk
in the front door (in which case they should be banished to
lower-priced real estate), or they're not being presented well (in
which case they need more visibility).

[Tim Parkin]

After some c.l.python discussion, here's a revised version of some
tweaks to the current home page.

But why are we tweaking this design? Our web site needs far more than
any amount of incremental revision is likely to give it. Really, it's
long past time we took this old dog out behind the barn and shot it.

I disagree, the old site is awful and realistically it's going to be end
of the this year / start of next year before relaunching the new site.
In the meantime, lets make the current site more useful. It's against my
interests as the worse the current site looks, the better our redesign
appears but I'll be altruistic about it ;-)

The redesign effort is looking increasingly like the types-sig to me.

The

types-sig expended lots of discussion time on complicated questions and
*never* *produced* *anything*.

I'm sure you didn't mean nothing as on my public folder alone there's
9.1Mb of proposed designs/information architecture notes and logos. Also
on the wiki there's prototype html components and proposed navigation
concepts. Here's the latest design if people haven't seen it.
Loading Image...

Anyway, there's nothing preventing anyone from grabbing portions of
python.org with wget, making an improved prototype and putting it on

some

random IP address. Fred even made a (somewhat controversial) tarball a
little while ago; that could be repeated. So, make a prototype!

Well.. There is one thing that is stopping me in particular which is a
lack of official feedback on the design proposal. I don't mind spending
a lot of time building a prototype but would like some sort of
indication that :-

a) We (myself and people on the design/redesign list) are on the right
track
b) If not, why not (and why hasn't it been mentioned)
c) If so then what are the constraints in building html and for the
platform and do we have a provisional "yes we'll use this design as long
as it meets our requirements for accessibility/speed/etc"

I for one would love to get going but, in order to do what I have done
so far, I have had to make assumptions about what the PWC/Python Board
want. In particular (but in no means an indicative or substantial part)
the choice of how to support Netscape. I've had to presume Netscape 4.x
will degrade to a text only type design in order to make the most of the
massive advantages newer (post '98) version browsers can deliver.

We also have put some thought into platform / administration IA etc but
it seems risky carrying on without some buy-in from the PWC/Python
Board. We'd love to build something useful using Zope (and we have
addressed some of the concerns raised) but if it's a 100% no-no, then
there is little point.

To be honest I am just carrying on under the assumption that the scale
of feedback a project like this needs is unlikely and that, in addition,
the PWC/Board are not going to make any commitment apart from a yes/no
once they see a completed proposal. Moving the proposed web designs into
a fully standards compliant and browser compatible state is a
substantial (40+ hour) task to get perfect and I really can't afford to
'waste' that amount of time. And I do mean waste as, unlike python code,
web development builds are extremely difficult to refactor.

To be honest what worries me is carrying on and developing something
with many 100's of hours of mantime invested in it only to be told
"we're doing something else". That would not only upset me but would
severely test my sanity and trust. Understanding this, I shouldn't
really go any further as that isn't something I want to risk. However, I
love Python and see a good future for it.

What we really need is an official nod to say "great... we like... go
and make some html!"

What do you suggest?

Tim (not disillusioned, just wary of one sided relationships)

[amk at amk.ca ()]

Post by Tim Parkin
I'm sure you didn't mean nothing as on my public folder alone there's
9.1Mb of proposed designs/information architecture notes and logos. Also

Oh, the types-sig produced proposals and strawman implementations, too, but
that doesn't change the fact that, in this month of September 2003, there is
no standard type checking tool for Python. From the viewpoint of a
python.org user, the redesign process won't have produced anything
until the redesign is actually live.

Post by Tim Parkin
Well.. There is one thing that is stopping me in particular which is a
lack of official feedback on the design proposal. I don't mind spending

<sigh> Yes, the indecision problem is at the root of it all. I hope the PWC
group can come up with some way to make such decisions.

--amk

[Aahz]

Post by amk at amk.ca ()
<sigh> Yes, the indecision problem is at the root of it all. I hope
the PWC group can come up with some way to make such decisions.

My take is that it's not the job of the PWC to drive things. So far, I
haven't seen consensus on even the base goals: some people seem to think
we need a glitzy web site to attract people to Python where others want
an upgraded website that focuses on usability first.

I don't think these goals are *necessarily* in conflict, but I also
cringe every time I see someone arguing for ridding the home page of
links to e.g. the Dev Guide. I think it would be easier for me to push
the PWC to make a decision if there were at least a clear description of
the options for goals.

What I suggest is that people write up short (100-200 words) descriptions
of their web site vision (maybe on a Wiki?). People who see a lot of
commonality can then group together to come up with a single vision to
sign jointly. (Maybe sending the visions to marketing-python to see if
anyone else wants to sign on.) Then each group submits its vision and
signatories to the PWC.

How does this sound?

[Steve Holden]

-----Original Message-----
Sent: Monday, September 22, 2003 4:41 AM
Subject: RE: [Pydotorg-redesign] Next iteration (was "I think that the
current site ....")

After some c.l.python discussion, here's a revised

version of some

tweaks to the current home page.

But why are we tweaking this design? Our web site needs far

more than

any amount of incremental revision is likely to give it. Really, it's
long past time we took this old dog out behind the barn and shot it.

I disagree, the old site is awful and realistically it's
going to be end
of the this year / start of next year before relaunching the new site.
In the meantime, lets make the current site more useful. It's
against my
interests as the worse the current site looks, the better our redesign
appears but I'll be altruistic about it ;-)

I agree that at the moment the web site redesign appears to be in limbo,
and applaud Andrew Kuchling's efforts to get going with an incremental
change rather than trying to do a complete job in the absence of any
guidelines. I'm glad to see you agree, Tim, as I think that could at
least get us moving in the right direction.

The redesign effort is looking increasingly like the types-sig to me.

The

types-sig expended lots of discussion time on complicated

questions and

*never* *produced* *anything*.

I'm sure you didn't mean nothing as on my public folder alone there's
9.1Mb of proposed designs/information architecture notes and
logos. Also
on the wiki there's prototype html components and proposed navigation
concepts. Here's the latest design if people haven't seen it.
http://pollenation.net/assets/public/python-latest.png

Because we have a committee "in charge" of this activity I've kind of
backed off this, but I think there's an amazing collection of good ideas
been proposed in the last months, and I'm somewhat surprised that there
does seem to be a residual feeling in certain quarters that the existing
web site is "good enough" - it isn't. Notwithstanding the need for
textual navigation, we would be foolish to omit the appeal that good
graphical design can add.

I'm absolutely positive there has been no fundamental change to the
design in the last five years, and probably it goes back rather further
than that, and it looks like it. The content has grown well beyond the
ability of the site's architecture to direct the reader's attention to
the appropriate areas. In short, like my own current web site (which is
if anything even worse) it looks like a "geek" site, and it seems clear
that it isn't being organized with maximal reader convenience in mind.

Anyway, there's nothing preventing anyone from grabbing portions of
python.org with wget, making an improved prototype and putting it on

some

random IP address. Fred even made a (somewhat controversial)

tarball a

little while ago; that could be repeated. So, make a prototype!

Well.. There is one thing that is stopping me in particular which is a
lack of official feedback on the design proposal. I don't
mind spending
a lot of time building a prototype but would like some sort of
indication that :-
a) We (myself and people on the design/redesign list) are on the right
track
b) If not, why not (and why hasn't it been mentioned)
c) If so then what are the constraints in building html and for the
platform and do we have a provisional "yes we'll use this
design as long
as it meets our requirements for accessibility/speed/etc"

Hear, hear! I've been loath to propose further ideas because they just
get sucked into the cauldron of indecision that is this mailing list.

I for one would love to get going but, in order to do what I have done
so far, I have had to make assumptions about what the PWC/Python Board
want. In particular (but in no means an indicative or
substantial part)
the choice of how to support Netscape. I've had to presume
Netscape 4.x
will degrade to a text only type design in order to make the
most of the
massive advantages newer (post '98) version browsers can deliver.
We also have put some thought into platform / administration
IA etc but
it seems risky carrying on without some buy-in from the PWC/Python
Board. We'd love to build something useful using Zope (and we have
addressed some of the concerns raised) but if it's a 100% no-no, then
there is little point.

Yes, it would be nice to see some of the other talent that's on offer
being utilized to improve the situation.

To be honest I am just carrying on under the assumption that the scale
of feedback a project like this needs is unlikely and that,
in addition,
the PWC/Board are not going to make any commitment apart from a yes/no
once they see a completed proposal. Moving the proposed web
designs into
a fully standards compliant and browser compatible state is a
substantial (40+ hour) task to get perfect and I really can't
afford to
'waste' that amount of time. And I do mean waste as, unlike
python code,
web development builds are extremely difficult to refactor.

Agreed.

To be honest what worries me is carrying on and developing something
with many 100's of hours of mantime invested in it only to be told
"we're doing something else". That would not only upset me but would
severely test my sanity and trust. Understanding this, I shouldn't
really go any further as that isn't something I want to risk.
However, I
love Python and see a good future for it.
What we really need is an official nod to say "great... we like... go
and make some html!"
What do you suggest?
Tim (not disillusioned, just wary of one sided relationships)

Personally I think we need some goals, and I would think the committee
should be the group to provide them, but I know that they are all busy
people and I'm sure that their silence is attributable to overload
rather than lack of good intentions. The confusions and cross-posting
between pydotorg, pydotorg-redesign and marketing-python haven't helped,
either.

Guido's clearly too busy with other matters to put a hand on the tiller
here, but I'm sensing a vacuum that isn't helpful for people like Tim,
whose goodwill we need to nurture. Proposals are made, then commented
upon, but no decision are resulting, and no goals are available by which
to measure specific designs against published criteria.

Goals have certain characteristics: specifically, they should be SMART:

S(imple)
M(easurable)
A(chievable)
R(ealistic), and with a
T(imescale)

So here's a sample goal as an "Aunt Sally". Would it help?

The new web site design should allow reader to reach all existing
content and provide a framework for the inclusion of new content. Any
page should be reachable from the home page with no more than four
mouse-clicks, and the more following sections should be more
accessible:

1 click [name the most important areas]
2 clicks [the next most important areas]
3 clicks [stuff more important than the 1995 conference]

Other than "leaf" content, navigational pages should not exceed 64 Kb
in
content-length [we can argue about this - personally I'd go for 128
Kb].

The redesign should be completed by October 31, and the content should
be publicly available in the revised format by November 31.

There, now tell me to shut up. I'll get back to pycon-organizers for
another month :-)

regards
--
Steve Holden http://www.holdenweb.com/
Python Web Programming http://pydish.holdenweb.com/pwp/
Interview with GvR August 14, 2003 http://www.onlamp.com/python/

[Steve Holden]

-----Original Message-----
Sent: Tuesday, September 23, 2003 6:04 PM
Subject: [Pydotorg-redesign] What are the goals?

Post by amk at amk.ca ()
<sigh> Yes, the indecision problem is at the root of it all. I hope
the PWC group can come up with some way to make such decisions.

My take is that it's not the job of the PWC to drive things.
So far, I
haven't seen consensus on even the base goals: some people
seem to think
we need a glitzy web site to attract people to Python where
others want
an upgraded website that focuses on usability first.
I don't think these goals are *necessarily* in conflict, but I also
cringe every time I see someone arguing for ridding the home page of
links to e.g. the Dev Guide. I think it would be easier for
me to push
the PWC to make a decision if there were at least a clear
description of
the options for goals.
What I suggest is that people write up short (100-200 words)
descriptions
of their web site vision (maybe on a Wiki?). People who see a lot of
commonality can then group together to come up with a single vision to
sign jointly. (Maybe sending the visions to marketing-python
to see if
anyone else wants to sign on.) Then each group submits its vision and
signatories to the PWC.
How does this sound?

Like nothing will continue to happen. We don't need visions, we need
direction, hence my (attempted) focus on goals.

Since my suggestions that we set goals was pretty much ignored, I'll
assume that those with more energy than I wish to continue the current
fruitless debate while AMK hacks away at the overgrowth.

regards
--
Steve Holden http://www.holdenweb.com/
Python Web Programming http://pydish.holdenweb.com/pwp/
Interview with GvR August 14, 2003 http://www.onlamp.com/python/

[Aahz]

Post by Steve Holden
Like nothing will continue to happen. We don't need visions, we need
direction, hence my (attempted) focus on goals.

How do you suggest we set goals, then?

[Dylan Reinhardt]

Post by Aahz

Post by Steve Holden
Like nothing will continue to happen. We don't need visions, we need
direction, hence my (attempted) focus on goals.

How do you suggest we set goals, then?

Jumping in here, our essential problem is we *lack* the ability to set
goals. As a consequence, we discuss goals in an decision-making vacuum,
which remains amusing for only a few weeks at a time.

The way things stand, interested parties are encouraged to gather what
information they can and put together a proposal that is evaluated in a
PEP-like process. That sounds nice, but doesn't show very much
awareness of how much effort goes into a non-trivial design proposal.

Precious few pieces of effective communications are developed by
committee in an open, collaborative process. A small group (1-4 people)
is going to need to go off and create something. This is only likely to
happen if the powers that be are *involved* and demonstrate a high
degree of predisposition toward taking this group's output seriously.

To move forward, we will have to identify and/or delegate
decision-making authority. That authority will need to be *involved* in
the redesign process. I don't see a lot of other ways to succeed.

FWIW,

Dylan

[Aahz]

Post by Dylan Reinhardt

Post by Aahz

Post by Steve Holden
Like nothing will continue to happen. We don't need visions, we need
direction, hence my (attempted) focus on goals.

How do you suggest we set goals, then?

Jumping in here, our essential problem is we *lack* the ability to set
goals. As a consequence, we discuss goals in an decision-making vacuum,
which remains amusing for only a few weeks at a time.

There's a difference IMO between discussing goals versus clearly
articulating sets of suggested goals then presenting them to a
decision-making body. The latter is what I'm suggesting that subgroups
of this list do. From my POV, a completely articulated -- but
summarized -- set of goals equates to "vision statement".

Nevertheless, I've gone ahead and forwarded AMK's message to the PWC,
asking whether anyone there thinks that the PWC should be a driver.

Post by Dylan Reinhardt
The way things stand, interested parties are encouraged to gather what
information they can and put together a proposal that is evaluated in a
PEP-like process. That sounds nice, but doesn't show very much
awareness of how much effort goes into a non-trivial design proposal.

That's exactly what I'm trying to get around. Once we have a set of
approved design goals, it should be easier both to create designs and to
evaluate proposed designs.

Post by Dylan Reinhardt
Precious few pieces of effective communications are developed by
committee in an open, collaborative process. A small group (1-4 people)
is going to need to go off and create something. This is only likely to
happen if the powers that be are *involved* and demonstrate a high
degree of predisposition toward taking this group's output seriously.
To move forward, we will have to identify and/or delegate
decision-making authority. That authority will need to be *involved* in
the redesign process. I don't see a lot of other ways to succeed.

While I don't disagree with you, I think that there are a lot of
different ways to be involved.

[amk at amk.ca ()]

Post by Aahz
How do you suggest we set goals, then?

I think the PWC, as the representatives chosen by the PSF, has to do it.
The five-person committee can converge more quickly on decisions
than a random group of people, some vocal and some not, on a mailing list.

Now, I'm not sure what decisions the PWC needs to make. Off the top of my
head:

1) There seems to be general agreement a more marketing-savvy site is needed.
Is this site going to be www.python.org?

If yes, is there a separate developer-oriented site at dev.python.org
or python-dev.org? Do we want a family of subsites (apps.python.org,
database., etc.)

If no, where is this marketing site going to be? go-python.org,
use.python.org, etc.?

2) What tack should this marketing site take? The marketing-python list has
discussed a zillion possible approaches: "faster programming", "happier
programming staff", "solve your problems", "an open source Java {or VB}
replacement". Choose an angle or combination of angles. Should it
emphasize success stories, the Python community, the open source aspect,
the readable pseudocode aspect.

3) What are the constraints on a new design: appearance, technologies (can
it use CSS, DHTML, Flash?).

3.5) Is a CMS wanted? Do we want the community to be able to post things,
or should it stay centralized?

4) Designers such as Tim Parkin need a way to get specific feedback. They
can request feedback from the entire list, but that wind blows every
which way. The committee needs to be able to decide "Here's what's good
and what's not", and do so reasonably quickly.

Those are some possible starting points, anyway.

--amk

[Steve Holden]

[Aahz]

Post by Steve Holden

Post by Dylan Reinhardt

Post by Aahz

Post by Steve Holden
Like nothing will continue to happen. We don't need

visions, we need

Post by Dylan Reinhardt

Post by Aahz

Post by Steve Holden
direction, hence my (attempted) focus on goals.

How do you suggest we set goals, then?

Jumping in here, our essential problem is we *lack* the

ability to set

Post by Dylan Reinhardt
goals. As a consequence, we discuss goals in an

decision-making vacuum,

Post by Dylan Reinhardt
which remains amusing for only a few weeks at a time.

There's a difference IMO between discussing goals versus clearly
articulating sets of suggested goals then presenting them to a
decision-making body. The latter is what I'm suggesting that
subgroups
of this list do. From my POV, a completely articulated -- but
summarized -- set of goals equates to "vision statement".

Unfortunately, as I suspect my "goal setting" email proved, if there's
no clear feedback from the PWC the discussions on goals will be as
fruitless as the discussions on redesign. I think we have a different
idea of a "vision statement", though.

A vision statement is something like "to be the primary source of
information about the use and development of the Python language", which
is neither simple nor measurable (without a research project), may or
may not be achievable and realistic, and definitely doesn't include a
timescale. SO it has few of the characteristics of a goal, or a set of
goals.

Post by Steve Holden
Nevertheless, I've gone ahead and forwarded AMK's message to the PWC,
asking whether anyone there thinks that the PWC should be a driver.

If the PWC won't set goals the redesign "project" is probably just going
to carry on spinning its wheels.

Post by Steve Holden

Post by Dylan Reinhardt
The way things stand, interested parties are encouraged to

gather what

Post by Dylan Reinhardt
information they can and put together a proposal that is

evaluated in a

Post by Dylan Reinhardt
PEP-like process. That sounds nice, but doesn't show very much
awareness of how much effort goes into a non-trivial design

proposal.
That's exactly what I'm trying to get around. Once we have a set of
approved design goals, it should be easier both to create
designs and to
evaluate proposed designs.

Yes! Yes! Goals! Measurability!

Post by Steve Holden

Post by Dylan Reinhardt
Precious few pieces of effective communications are developed by
committee in an open, collaborative process. A small group

(1-4 people)

Post by Dylan Reinhardt
is going to need to go off and create something. This is

only likely to

Post by Dylan Reinhardt
happen if the powers that be are *involved* and demonstrate a high
degree of predisposition toward taking this group's output

seriously.

Post by Dylan Reinhardt
To move forward, we will have to identify and/or delegate
decision-making authority. That authority will need to be

*involved* in

Post by Dylan Reinhardt
the redesign process. I don't see a lot of other ways to succeed.

While I don't disagree with you, I think that there are a lot of
different ways to be involved.

Indeed there are, and I respect all the individuals involved in this
effort. Together they could be an incredibly powerful team. I just want
to see a clear sense of direction and some goals so that I can decide
whether I have the skills and knowledge to help the group achieve what
it wants to.

So far we don't appear to have got beyond the "motherhood and apple pie"
stage, and while a lot of effort has gone into certain aspects of the
design there hasn't been any clear statement of either the vision for
the site or the goals we have to meet to achieve that vision.

regards
--
Steve Holden http://www.holdenweb.com/
Python Web Programming http://pydish.holdenweb.com/pwp/
Interview with GvR August 14, 2003 http://www.onlamp.com/python/

[Aahz]

Post by Steve Holden
Unfortunately, as I suspect my "goal setting" email proved, if there's
no clear feedback from the PWC the discussions on goals will be as
fruitless as the discussions on redesign. I think we have a different
idea of a "vision statement", though.

Probably. But I think that while it's fair to expect the PWC to give
feedback on a packaged set of goals, I'd also suggest that it's realistic
not to expect the PWC to drive the creation of goals. I'm suggesting
that instead of trying to get consensus on this list, each person who
wants to create a set of goals will get a hearing from the PWC.

Post by Steve Holden
So far we don't appear to have got beyond the "motherhood and apple
pie" stage, and while a lot of effort has gone into certain aspects of
the design there hasn't been any clear statement of either the vision
for the site or the goals we have to meet to achieve that vision.

And so far, I haven't seen even a single person step up to suggest a
coherent set of goals. I'm not sure why you think the PWC would have
more time/energy than the group here. It's difficult to create good
goals, but at this point it's more useful than anything else we can do.

Here. I'll go ahead and throw something out that people can use to
agree or argue with -- but do it *ONLY* as a packaged set. Don't bother
arguing with me about individual points:

* www.python.org is the only central web site for Python; there are no
separate domains for different pieces.

* The front page should cater to all the expected primary users: Python
newbies, management, and existing developers.

* All important information should be available within one or two clicks
off the main page.

* The web site must be accessible and readable through Lynx and other
low-end browsers, as well as with JavaScript disabled. The site may have
optional advanced features that make it more attractive in high-end
browsers.

* No page other than explicitly labeled download or marketing pages
should take more than thirty seconds to load over a 28.8 modem: images,
stylesheets, and all.

This isn't particularly good, but it took me less than ten minutes. It
also focuses on the user experience; it says nothing about how the site
is maintained. Let's keep that as a separate discussion.

[Tim Parkin]

Post by Aahz
This isn't particularly good, but it took me less than ten minutes. It
also focuses on the user experience; it says nothing about how the site
is maintained. Let's keep that as a separate discussion.

I think we have to take best practices for granted unless there is a
situation where a best practice is in conflict with another best
practice. It is at these pointes we *NEED* feedback. Case in point: full
support for Netscape vs. best practices in w3c wai accessibility and
standards/semantic compliance. In this I raised it as a conflict and
looked for feedback but received very little and had to assume that
support for Netscape was less important than good accessibility and
standards/semantic compliance.

Other points that have already been stated as part of the initial pycon
brief, further discussion through irc or on the mailing list itself are
shown here :

. make it look more "professional".
. make the site more obvious to use, particularly for beginners.
. prioritise links to download and documentation
. introduce an obvious site search facility
. reduce the clutter in commercial exits
. head towards valid markup
. promoting Python, leverage evangelism
. use existing metrics to focus improvements
. simplify the homepage
. documentation types confusing
. simple search on home page
. Should bear in mind that mirroring is important
. No 3d rotating head of Guido with glowing lights around him (Kevin
Altiss)
. All sections should adhere to new design templates (Kevin Altiss)
. a google site search on python.org is very ineffective right now
(Kevin Altiss)
. make it work in lynx (Aahz)
. we want blue (Aahz)
. reduce the maintenance hassles for the webmasters (Kevin Altiss)
. top global nav and left hand secondary nav (Aahz)
. front page under 60k (Aahz)
. breadcrumb trail across the site (Aahz)
. Steve Krug - Don't Make Me Think (Guido Van Rossum)
. No flash, No dark backgrounds (Guido Van Rossum)
. Don't design by committee (Guido Van Rossum)
. Browsing then searching then sitemap (Guido Van Rossum)
. home page should have a lot more "sell" with convenient links to the
developer resources (Kevin Altiss)
. there should not be two sites. Python.org is *the* homepage (Aahz)

Generally these break down into four overall core requirements

. Usability / Clarity
. Professional Design
. Both Marketing / Informational Purposes
. Best Practices in Web Development

On top of this, there is the marketing document at :

http://wingide.com/pub/misc/pymarket/py-market-plan-0.0.6.rst

Most of these have been available for at least a month. General best
practices have been taken for granted.

Other than the previously stated PWC/PSF guidelines, here's a list of
general best working practices that I typically work to and the books
that go into some more detail about them.

* Page should have a time to first click of less than 8 second on an
average connection.
(This actually means the time until you can usefully use the navigation
or content. This affects sites that use graphics as navigational
elements. In these site, all of the headers graphics are loading before
it gets around to loading the menu item graphics. This typically means
the time to first click is similar to the time to download the whole
page. If a page uses html text as the menu items, the page will be
usable before even the whole html has loaded. In this way the page may
be usable before even half of all page elements are downloaded. This
increases the 'perceived' speed of a site.)

* All pages should answer : Where am I? Where do I go Next? How do I get
back?

* A preferred vocabulary should be created that is widely understood and
cannot be confused

* The site should sit comfortably in the perceived market space.

* Pictures of people are widely recognized as engaging more than any
other method, use if there is a choice.

* Don't hijack the browser. Leave it alone. Make it do what people
expect.

* Recognise that most people will be deep linking into the site

* Scrolling should only be needed when a user finds the page they want.

* Pages should print clearly

* The site content should be optimised to benefit accurate search
indexing

These are only some of the best practices that are not widely
recognised. If we were to make a list of all of the practices we would
be writing a book and there are already enough books out there. I have
copied and pasted most of this content directly from the web proposal
that's currently being written that will complement the marketing
document mentioned previously. A lot of extra information comes from
developing best practise guidelines for British Airways' Internet
strategy, a document I was co-author of.

Books for Best Practice
-----------------------
* "Building Accessible Website" - Clark
* "Don't Make Me Think" - Krug
* "Homepage Usability" - Nielsen & Tahir
* "Designing Web Usability" - Nielsen
* "Practical Information Architecture" - Reis
* "Information Architecture" - Wadtke
* "Top 100 Internet Mistakes" - Peter Burns
* "CSS - Separating Content From Presentation" - Briggs, Champeon,
Costello, Paterson
* "Collaborative Web Development" - Brudman
* "WWW Layout" - Glenwright
* "WWW Type" - Pring
* "Reality Chec"k - Wieners & Prescovitz
* "Hotwired Style" - Veen
* "Secrets of Successful Websites" - Siegel
* "Click Here" - Pirouz
* "Accessible Websites" - Thatcher, Bohman, et al
* "Web Metrics" - Sterne
* "Submit Now" - Chak
* "Designing with Web Standards" - Zeldman
* "Designing CSS Pages" - Schmidt
* "Usability" - Braun, Gadney, et al
* "Usability for the Web" - Brinck, Gergle, Wood
* "Eric Meyer on CSS" - Meyer
* "Customer Centred Design" - Hyatt

Loading Image...

What I, personally, believe the PWC should be doing is dictating a
framework within which people who are willing to contribute can work, if
this can't be achieved then at least a date by which a decision will be
made should be proposed and the process by which a choice will be made.

As it is, it shouldn't take a great deal of time or coordination to get
to 'yes we like the look of that' or 'no, perhaps it should be more...'.
As it is there has been no feedback. We can only assume from this that,
because all the work has been available for quite some time, there is
approval and a desire for us to carry on. If there was a negative
opinion or no wish to use what has been created then why would there be
no communication of that?

As the situation lies, the design/marketing/redesign groups have had
numerous discussions and reached a consensus about what is felt to be
the right direction. This is now being documented. I'm not sure what the
advantage of having multiple groups doing the same thing is unless there
is dissent as to the direction that should be taken. I haven't noticed
any constructive dissent so assume we have a certain amount of
consensus.

I am continuing to develop the proposal, of which the material presented
is only a fraction.

Constructively your,

Tim Parkin

[Todd Grimason]

Post by Tim Parkin
practice. It is at these pointes we *NEED* feedback. Case in point: full
support for Netscape vs. best practices in w3c wai accessibility and
standards/semantic compliance. In this I raised it as a conflict and
looked for feedback but received very little and had to assume that
support for Netscape was less important than good accessibility and

As a +1 for valid CSS/XHTML, Redhat's (admittedly lowkey/simple) new
site for "Fedora" uses CSS and no tables. Other points in favor of
this approach are meeting reqs such as:

- low bandwidth: this type of markup *greatly* reduces byte counts

- accessible: to nearly every "browser"-type device in existence -
even Netscape 4 - it's just not purty in NS4 or earlier. Works great
in mobile phones even! (for looking up functions on the site from a
pub?)

- IE gets most of it right, so the all-important "manager-types" can
use the site with no problems. Geekier users will likely be using
something besides NS4, so we're covered there (Konqueror, Mozilla,
Opera, Lynx(!), Safari, etc.)

- minimizes work: clean use of CSS (separation of content/style) helps
reduce maintenance work

- visibility: well-known sites are starting to convert, such as
previously mentioned Redhat site, ESPN, Wired News, etc. Each time
they do, they get a fair bit of press in the "design/standards world";
this is a good group to attract new users from as well.

- future maintainability: won't have to do this in a year when full
CSS/XHTML is even more widespread and beneficial

- 100s more reasons all documented over the web and the excellent
booklist Tim posted...

[Aahz]

[This is far too long a message for me to respond to in any substantive
way, so I'm just going to take a few potshots at points that I think are
important.]

Post by Tim Parkin
I think we have to take best practices for granted unless there is a
situation where a best practice is in conflict with another best
practice. It is at these pointes we *NEED* feedback. Case in point: full
support for Netscape vs. best practices in w3c wai accessibility and
standards/semantic compliance. In this I raised it as a conflict and
looked for feedback but received very little and had to assume that
support for Netscape was less important than good accessibility and
standards/semantic compliance.

So far, everyone has said that it's possible to have something readable
in Netscape, even if it doesn't look particularly good, so I don't see
what the issue is.

Post by Tim Parkin
Generally these break down into four overall core requirements
. Usability / Clarity
. Professional Design
. Both Marketing / Informational Purposes
. Best Practices in Web Development

Right. But I don't know how much agreement there is in achieving those
requirements.

Post by Tim Parkin
http://wingide.com/pub/misc/pymarket/py-market-plan-0.0.6.rst

That's a long document that's difficult to wade through, not to mention
a lack of clarity about how relevant it is to the website redesign.

Post by Tim Parkin
As it is, it shouldn't take a great deal of time or coordination to get
to 'yes we like the look of that' or 'no, perhaps it should be more...'.
As it is there has been no feedback. We can only assume from this that,
because all the work has been available for quite some time, there is
approval and a desire for us to carry on. If there was a negative
opinion or no wish to use what has been created then why would there be
no communication of that?

Speaking strictly for myself, one problem is that a key design proposal
(yours) has not been available as an HTML page. I'm not prepared to
spend time looking at it until either most other people have approved it
or it's available in HTML. I'm just not going to download pictures of
an evolving design.

Post by Tim Parkin
As the situation lies, the design/marketing/redesign groups have had
numerous discussions and reached a consensus about what is felt to be
the right direction. This is now being documented. I'm not sure what the
advantage of having multiple groups doing the same thing is unless there
is dissent as to the direction that should be taken. I haven't noticed
any constructive dissent so assume we have a certain amount of
consensus.

Some, yes. But it's not clear how much consensus. And the three groups
(marketing, website redesign, and logo) have different emphases and
different groups of people.

[A.M. Kuchling]

Post by Aahz
So far, everyone has said that it's possible to have something readable
in Netscape, even if it doesn't look particularly good, so I don't see
what the issue is.

Tim is doubtless worried that the mandate will be for the site to look
identical in both NS4 and current Mozilla/IE/Safari/whatever. Netscape
4's CSS implementation is very buggy, with bugs ranging from rendering
glitches to browser segfaults.

* If it doesn't matter if NS4 is unreadable or crashes, the designer
can just use CSS
and let the chips fall. Easy.
* If the site has to be readable in NS4 but getting a different, less
attractive, presentation
doesn't matter, the designer can just hide or simplify the CSS
stylesheet served
to Netscape4 using Apache directives or some CSS tricks.
* If the site has to look identical in both NS4 and Mozilla, it's very
difficult; you end up
doubling the work required, or you implement for NS4 and avoid using
modern CSS.

You seem to be OK with the middle option.

--amk

[Aahz]

Post by A.M. Kuchling
Tim is doubtless worried that the mandate will be for the site to look
identical in both NS4 and current Mozilla/IE/Safari/whatever.

That's ridiculous. From my POV, NS4 is simply one of a long line of
non-mainstream browsers that must be supported (Lynx, Links, Internet
appliances, various cell phone and PDA browsers). There's no way we can
mandate an idential look across all these platforms, particularly when
you start considering that 800x600 vs 1280x1024 might as well be
different platforms.

Post by A.M. Kuchling
* If it doesn't matter if NS4 is unreadable or crashes, the designer
can just use CSS
and let the chips fall. Easy.

That is not an option.

Post by A.M. Kuchling
* If the site has to be readable in NS4 but getting a different, less
attractive, presentation
doesn't matter, the designer can just hide or simplify the CSS
stylesheet served
to Netscape4 using Apache directives or some CSS tricks.

That's the way to go.

Post by A.M. Kuchling
* If the site has to look identical in both NS4 and Mozilla, it's very
difficult; you end up
doubling the work required, or you implement for NS4 and avoid using
modern CSS.

Well, I'm not opposed to the idea of ditching CSS. ;-) (You figure out
how much winking that is.)

Post by A.M. Kuchling
You seem to be OK with the middle option.

Absolutely. No, I'm not just okay, that's a requirement.

[Todd Grimason]

Post by Aahz

Post by A.M. Kuchling
Tim is doubtless worried that the mandate will be for the site to look
identical in both NS4 and current Mozilla/IE/Safari/whatever.

That's ridiculous. From my POV, NS4 is simply one of a long line of
non-mainstream browsers that must be supported (Lynx, Links, Internet
appliances, various cell phone and PDA browsers). There's no way we can
mandate an idential look across all these platforms, particularly when
you start considering that 800x600 vs 1280x1024 might as well be
different platforms.

It is ridiculous, but you'd be (shocked|amazed|disappointed|&c) by how
many clients still demand it. Thankfully that number is slowly
declining now, but it has been (a very painful) standard requirement
from clients for years now.

Compliant CSS/XHTML will guarantee legibility on all clients, and
nice presentation on the majority of clients. Also reduced bandwidth
and easier maintenance.

It appears an agreement has been reached!(?)

[Fred L. Drake, Jr.]

Post by Todd Grimason
It is ridiculous, but you'd be (shocked|amazed|disappointed|&c) by how
many clients still demand it. Thankfully that number is slowly
declining now, but it has been (a very painful) standard requirement
from clients for years now.

Clients have to pay for what they get. When someone pays for NS4
support on python.org, it'll get a higher priority (heck, it might
even happen in that case!).


-Fred

[Skip Montanaro]

Post by A.M. Kuchling
* If it doesn't matter if NS4 is unreadable or crashes, the designer
can just use CSS and let the chips fall. Easy.

aahz> That is not an option.

Seems like a perfectly valid option to me. A thread here some time ago
suggested NS4's share of the python.org hits was just over 1%. At some
point, you have to just cut it loose. Lynx is a different matter, not
because it necessarily has a larger share of the browser market (it might at
this point for all I know), but because it serves a unique segment of the
user population, those without access to graphical user interfaces. I
suspect most people who use NS4 *could* use something else. It doesn't hurt
that it has a vocal supporter either. I've never heard a peep on webmaster
from anyone complaining that the Python site doesn't work in NS4. Maybe the
people coding the site have been diligent or just lucky. In any case,
Netscape 4's warts mean that at some point you have to decide it's not worth
the effort.

Post by A.M. Kuchling
* If the site has to be readable in NS4 but getting a different, less
attractive, presentation doesn't matter, the designer can just hide
or simplify the CSS stylesheet served to Netscape4 using Apache
directives or some CSS tricks.

aahz> That's the way to go.

I'd still argue that NS4 support is not a primary design criterion. If,
after all is said and done, you can (easily) tweak the CSS or add a
RewriteRule to the Apache config to deliver a different CSS file to Netscape
4 users, that's fine. Just like we don't take pains to support Unicos,
Cygwin or OS/2 (someone from the Python community has to pony up some
patches) we shouldn't take great pains to support NS4 either. In fact, I
like that analogy quite well. Just like Python itself, this is an open
source programming project (maybe the CVS repository should be available via
anoncvs) being done with limited contributed resources. Like any other open
source project, the programmers get to scratch their itches. They aren't
obligated to scratch someone else's.

Post by A.M. Kuchling
* If the site has to look identical in both NS4 and Mozilla, it's
very difficult; you end up doubling the work required, or you
implement for NS4 and avoid using modern CSS.

aahz> Well, I'm not opposed to the idea of ditching CSS. ;-) (You
aahz> figure out how much winking that is.)

Like it or not, CSS is the future.

Skip

[Aahz]

Post by Skip Montanaro

Post by A.M. Kuchling
* If it doesn't matter if NS4 is unreadable or crashes, the designer
can just use CSS and let the chips fall. Easy.

aahz> That is not an option.
Seems like a perfectly valid option to me. A thread here some time
ago suggested NS4's share of the python.org hits was just over 1%.
At some point, you have to just cut it loose. Lynx is a different
matter, not because it necessarily has a larger share of the browser
market (it might at this point for all I know), but because it serves
a unique segment of the user population, those without access to
graphical user interfaces. I suspect most people who use NS4 *could*
use something else. It doesn't hurt that it has a vocal supporter
either. I've never heard a peep on webmaster from anyone complaining
that the Python site doesn't work in NS4. Maybe the people coding
the site have been diligent or just lucky. In any case, Netscape 4's
warts mean that at some point you have to decide it's not worth the
effort.

Not being familiar with NS4's problems, I have no idea why we don't
currently get complaints, but given how much traffic python.org gets, I'd
have to say that 1% is a substantial amount of traffic. I will agree
that it's not worth going out of the way to ensure that things are
readable with NS4, but as soon as a single complaint comes in that our
site is unreadable or crashes NS4, I think it should be fixed. That's
what I mean by "not an option".

Post by Skip Montanaro

Post by A.M. Kuchling
* If the site has to be readable in NS4 but getting a different, less
attractive, presentation doesn't matter, the designer can just hide
or simplify the CSS stylesheet served to Netscape4 using Apache
directives or some CSS tricks.

aahz> That's the way to go.
I'd still argue that NS4 support is not a primary design criterion.
If, after all is said and done, you can (easily) tweak the CSS or add
a RewriteRule to the Apache config to deliver a different CSS file
to Netscape 4 users, that's fine. Just like we don't take pains to
support Unicos, Cygwin or OS/2 (someone from the Python community has
to pony up some patches) we shouldn't take great pains to support
NS4 either. In fact, I like that analogy quite well. Just like
Python itself, this is an open source programming project (maybe
the CVS repository should be available via anoncvs) being done with
limited contributed resources. Like any other open source project,
the programmers get to scratch their itches. They aren't obligated to
scratch someone else's.

Thing is, given how many different browsers there are out there, I think
it's fair to say that a bug that shows up in one browser is pretty
likely to show up in another. The point isn't whether we're catering to
NS4 in specific, the point is the extent to which we code the site to
fit the lowest-common denominator. I'd say that extent is pretty high.

[Tim Parkin]

Post by Aahz
So far, everyone has said that it's possible to have something readable
in Netscape, even if it doesn't look particularly good, so I don't see
what the issue is.

Only Simon Willison and Myself have said this, the only other
consensus was that the number of netscape 4 visitors was 1.17%.
Unless the issue was resolved elsewhere.

The issue is that a web design CAN be built
in XHTML/CSS for netscape 4.x but is substantially
more work and sacrifices a lot of the advantages of CSS.
Some people see this as important just as
Todd says. I'm glad that we agree it's a
waste of time.

Post by Aahz
(re marketing document) That's a long document
that's difficult to wade through, not to mention
a lack of clarity about how relevant it is to the
website redesign.

I presume all of it is relevant if the website is to be
any sort of marketing tool. I've used the whole document
in the process of developing a site architecture/visual design.

Post by Aahz
Speaking strictly for myself, one problem is that a key design proposal
(yours) has not been available as an HTML page. I'm not prepared to
spend time looking at it until either most other people have approved

it

Post by Aahz
or it's available in HTML. I'm just not going to download pictures of
an evolving design.

Why not? This is how most design processes go. I hope the PWC can
agree it would be an incredible waste of time if I were to
develop the site in HTML only for them to say "I don't like
the design, can you change it" which would probably mean
recreating a whole load of HTML. I do have a full time job aswell
(actually running a company which is probably more like 5 full time
jobs).

If the visual design is irrelevant to you then
I'm happy to create a text only version that
reflects the information architecture and the
way the document design would degrade in lynx.
I really appreciate that you are the one of the
only PWC representatives supplying feedback and
I hope you don't mind me speaking frankly. If the
visual design is relevant and yet the effort can't
be found to assess a single image then I'm sorry
to say I'm a ittle dissapointed, especially
considering the amount of work that has been put
in by myself and other parties.

I wasn't going to build the HTML without any
further mandate from the PWC/PSF but as it
looks like this is unlikely to happen,
I will carry on regardless under the presumption
That; the mailing lists are high enough profile
that people would have objected had they disliked
what was available.

It should be remembered that the current design
proposal was originally prototyped in HTML to ensure
the design could be achieved and the results
acceptable and this has been available on the
redesign wiki for quite a while. I'm not sure how
obvious this had been made so. Please note that little
effort has been made regarding cross browser
compatibility or that the results are optimised for lynx
however as the design is semantic HTML it should render
fairly well.

http://www.pollenation.net/assets/public/index_about.html

This demonstrates how the menu, the breadcrumb
trail and the semantic content would appear. Although
this doesn't indicate the content or the
information architecture. If a mock up of the
lynx version of the site would be beneficial,
this would be realistic to achieve in a short
period of time. If this is the case then I'll
proceed with this in order to demonstrate the
proposal. If not then I'll continue under the
previous assumption.

Post by Aahz

Post by Tim Parkin
. Usability / Clarity
. Professional Design
. Both Marketing / Informational Purposes
. Best Practices in Web Development

Right. But I don't know how much agreement
there is in achieving those requirements.

Does this mean the PSF, having proposed these requirements,
can't agree whether it's a) worth achieving. Or
b) possible to achieve. I'm not sure I understand.

Is it that all of the requirements aren't in one place?
That's one of the reasons I had compiled all of the
statements by PSF/PWC members togather. If any one has
any objections to the design requirements stated in the
previous post then raise them so they can be discussed.
Otherwise, tacit approval would seems to have been reached.

I'm not trying to be difficult :-) , all I want is that
there is a continuing movement forward and that effort
is not going to waste when a little feedback could stop it.

Verbosely... Tim

[Laura Creighton]

I have a shorter idea.

I have no idea who it is that has approval of this -- PWC, PSF -- who
the people are.

I suggest that somebody who knows posts the list of people here.
Then, we give them each by the end of the week to look at the current
design, and say, 'I like the current design and think further work
would not be wasted' -- or if they cannot stand by that statement,
say something else.

Laura

[Fred L. Drake, Jr.]

Post by Laura Creighton
I have no idea who it is that has approval of this -- PWC, PSF -- who
the people are.

The PWC decides what happens with the site, in the end.

Post by Laura Creighton
I suggest that somebody who knows posts the list of people here.

I don't have the list handy (sorry), but I'm on it.

Post by Laura Creighton
Then, we give them each by the end of the week to look at the current
design, and say, 'I like the current design and think further work
would not be wasted' -- or if they cannot stand by that statement,
say something else.

If by current design, you mean what's on www.python.org now:

The current design doesn't sicken me, but it could stand a *lot*
of improvement.

The stuff I've seen from Tim Parkin looks quite promising, and I
regret that I've not been able to be more responsive to his
inquiries. I definately intend to take some time to look further into
it, and I appreciate the efforts everyone has made to work this out.

Now, on the related issue of site goals, I'm more ambivalent:

The site should be easier to use for newbies (and old-timers), but
making this a "marketing" or "advocacy" site sounds scary.

It's not at all clear to me what the win is in trying to convert the
unconverted; they'll convert on their own if they determine it's right
for them.

As far as the information architecture goes:

What's on the site now is pretty messy, and doesn't effectively
serve either the goal of making it easy for programmers to find
information or for the audiences of marketing hype to get the hype
that's fit for them.

So it realy needs work. I applaud the efforts Andrew Kuchling has led
to make the site easier to navigate (structurally).

How's that? ;-)


-Fred

[Fred L. Drake, Jr.]

Post by Tim Parkin
The issue is that a web design CAN be built
in XHTML/CSS for netscape 4.x but is substantially
more work and sacrifices a lot of the advantages of CSS.
Some people see this as important just as
Todd says. I'm glad that we agree it's a
waste of time.

I stated my position on the NS4 issue earlier today in response to a
different email, but I'll summarize here in case anyone missed it:

If we need to do any more than the @import trick to support NS4,
we're doing too much. If the @import trick is sufficient, we
should do it.

NS4 is dying off, and we shouldn't stop it. (If we can hurry it
along, that would be a Good Thing, even if NS4 users don't know that
yet.)


-Fred

[Roy Smith]

Post by Skip Montanaro

Post by A.M. Kuchling
* If it doesn't matter if NS4 is unreadable or crashes, the designer
can just use CSS and let the chips fall. Easy.

aahz> That is not an option.
Seems like a perfectly valid option to me. A thread here some time ago
suggested NS4's share of the python.org hits was just over 1%. At some
point, you have to just cut it loose.

I've stayed out of this for the most part for a while, but I'm with
Skip 100% on this. It just doesn't make sense to cater to 1% of the
audience, especially when as Skip points out, there's nothing keeping
NS4 users from upgrading to something newer. We're also not talking
about 1% new adopters of a technology, so there's an expectation that
the audience share will grow. We're talking about the hangers-on, with
the expectation that it will continue to decline.

Post by Skip Montanaro
Lynx is a different matter, not
because it necessarily has a larger share of the browser market (it
might at
this point for all I know), but because it serves a unique segment of
the
user population, those without access to graphical user interfaces.

Again, I'm with Skip. I can't remember the last time I actually used a
computer that didn't support some kind of graphical environment, but
it's fairly common that I've got a slow network between me and a remote
Unix box I'm working on. If I've got to download something from a web
site to that remote Unix box, firing up Lynx on the remote box is often
the simpliest way to get the job done. If my local box is a Mac or Win
box with no X server installed, then Lynx is my only choice if I want
to do the download directly to the remote box.

There are certainly enough other non-graphical applications (PDA's, web
crawlers, browsers for the visually impaired) that the ability to
render reasonably on a text-only display has value out of proportion to
the raw usage stats. But NS4 is a complete non-starter as far as I'm
concerned.

Post by Skip Montanaro
Like it or not, CSS is the future.

Once again, Skip I agree with Skip. I'm not particularly thrilled with
CSS, but the fact is it's the standard, it's here to stay, and it's
supported in all mainstream browsers. To talk of putting effort into
NS4 while ignoring CSS is lunacy (IMHO).

On a slightly different thread, I agree with the idea of not doing web
page mockups in HTML. In the design stage, you really want to
concentrate on the design, not the implementation. The design should
drive the coding. If you do your design in HTML, it's too easy to let
the coding drive the design. It's the same reason people mock-up GUI's
with paint/draw programs instead of coding them. First answer the
question, "what do we want?", then worry about "How do we get what we
want?" later.

[Fred L. Drake, Jr.]

Post by Roy Smith
Again, I'm with Skip. I can't remember the last time I actually used a
computer that didn't support some kind of graphical environment, but
it's fairly common that I've got a slow network between me and a remote
Unix box I'm working on. If I've got to download something from a web

I just poked at the current python.org with "links" (different from
"lynx", at least in that it happened to be installed on this box), and
the site held up fairly well; the organization was still sloppy, but
the presentation degraded cleanly and was about as usable it is on the
recent versions of Mozilla I'm accustomed to using.

That's the kind of behavior I'd like to see on the new design. I'm
not concerned with seeing the exact same thing on every browser. I'd
like the site to be navigable on NS4, but degraded presentation is
quite tolerable. (Heck, it's almost desirable to get NS4 users out of
their slump!) If we can do it using a simple stylesheet with an
@import to load portions of the style that would break NS4, that seems
like the right level of effort. Making a dynamic selection on the
server seems like the wrong approach, and doesn't lend itself to

Post by Roy Smith

Post by Skip Montanaro
Like it or not, CSS is the future.

...

Post by Roy Smith
Once again, Skip I agree with Skip. I'm not particularly thrilled with
CSS, but the fact is it's the standard, it's here to stay, and it's
supported in all mainstream browsers. To talk of putting effort into
NS4 while ignoring CSS is lunacy (IMHO).

Whatever the faults of CSS, it's much better than having the style
information inline. That's good enough for me. We're not getting
beyond it for quite a while.


-Fred

[Roy Smith]

Post by Aahz
as soon as a single complaint comes in that our
site is unreadable or crashes NS4, I think it should be fixed.

Why? Did we sell somebody a support contract while I wasn't looking?

[Fred L. Drake, Jr.]

Post by Aahz
as soon as a single complaint comes in that our
site is unreadable or crashes NS4, I think it should be fixed.
Why? Did we sell somebody a support contract while I wasn't looking?

If there's a complaint that the site "doesn't work" under NS4, then it
would be nice if we could make it work for them. That doesn't mean we
should break it or degrade it for people using current browsers. It
also doesn't mean we need to exert unlimited levels of effort to make
it work for NS4.

The idea that we should support all once-mainstream browsers doesn't
make sense to me; we're not supporting Mosaic now either.

The text browsers are more interesting because they offer different
functionality than older graphical browsers. The older graphical
browsers aren't offering functionality that's different, so present no
cause to continue support.

After all this discussion about NS4 compatibility, I'm inclined to
make a site that can't possibly work with it. ;-)


-Fred

[Tim Parkin]

Post by Aahz
I'm not prepared to
spend time looking at it until either most other people have approved

it

Post by Aahz
or it's available in HTML. I'm just not going to download pictures of
an evolving design.

http://pollenation.net/assets/public/python-lynx.html

http://pollenation.net/assets/public/python-lynx-interior.html

These pages are built with the intention of showing how the proposed
design will render in Lynx. As a by product, they also show how the
'worst case scenario' netscape 4.x look will render. They are also the
starting point for developing the CSS for the site. Some changes will be
made in implementing the CSS but these are unlikely to affect the
semantic part of the HTML.

Some of the order of items may not seem totally straightforward but this
is to do with ensuring the accessibility for speech readers. Some
changes to the order can be made, however the exact order of these items
will affect how the site may be built using html and css and hence some
orderings may generate more than a justifiable level of extra work.

This is a preliminary design, and like all designs there may be small
variances between this and the finished product. The (X) HTML may not
validate at the moment.

Tim

[Aahz]

Thanks for the Lynx page; I'm going off for a long weekend today, so I
don't know when I'll get to it.

[Tim Parkin]

Post by Tim Parkin
http://pollenation.net/assets/public/python-lynx.html
http://pollenation.net/assets/public/python-lynx-interior.html

For people without easy access to Lynx,

Loading Image...
Loading Image...

Tim