python.org CMS

[Aahz]

The team for www.python.org is taking another crack at considering a
content management system for handling the website's content. This is
particularly important considering, for example, how far beind we are in
posting job ads -- getting the community to do our grunt work for us is
the primary motivator.

Right now, Plone/Zope2 is the primary competitor simply because it's
well-publicized, but none of us really has that much experience with
websites whose core maintainers are going to be programmers. ;-) We'd
like to know if there are other packages available that can do what we
want; we're more interested in getting information about each package
than in package comparisons. We're also interested in hearing about
experiences creating custom systems.

Here are the primary criteria that we're working with:

The data must be stored in a plain text format that would allow direct
access if the main system was not functioning. This doesn't necessarily
imply the ability to circumvent a web based layer by editing text files
directly, though there does need to at least be an easy way to export
and import such files for people who want to use a Real Editor [tm].

Workflow is presumed to be a hierarchial publish reject model (in other
words as simple as is possible to begin with). We'll need both
coarse-grained and fine-grained security (giving people a variety of
access levels to large and small sections of the site).

The system needs to work with browsers that have limited functionality
(i.e. Lynx).

 

[Ville Vainio]

aahz> The system needs to work with browsers that have limited
aahz> functionality (i.e. Lynx).

Isn't the functionality in Lynx a bit *too* limited? For text based
browsing there is w3m, which can render legibly a whole lot more web
pages...

 

[Max M]

The team for www.python.org is taking another crack at considering a
content management system for handling the website's content. This is
particularly important considering, for example, how far beind we are in
posting job ads -- getting the community to do our grunt work for us is
the primary motivator.

it *does* need a facelift

Right now, Plone/Zope2 is the primary competitor simply because it's
well-publicized

Good choice!

The data must be stored in a plain text format that would allow direct
access if the main system was not functioning. This doesn't necessarily
imply the ability to circumvent a web based layer by editing text files
directly

Your intention is unclear here. If you don't have access to eg. a plone
site, you cannot edit the "files" at all. But on the other hand Zope
allways work ;-)

though there does need to at least be an easy way to export
and import such files for people who want to use a Real Editor [tm].

You can easily do that in Plone by using ftp or webdav. Using either
some kind of structured tekst or html. The data is marshalled before and
after editing. This makes it really easy to us normal tools.

Workflow is presumed to be a hierarchial publish reject model (in other
words as simple as is possible to begin with). We'll need both
coarse-grained and fine-grained security (giving people a variety of
access levels to large and small sections of the site).

You can do that out of the box in Plone. Either setting permissions on
folders manually, or by using one of the workgroup tools that makes it a
bit simpler.

The system needs to work with browsers that have limited functionality
(i.e. Lynx).

hmm! Why on earth would anybody want to do that. Well I guess that some
techs can be stubborn. For editing it would be far easier to just ftp to
the documents you want to edit and then use your editor of choice.

Anyhoo it should work in Plone too.


regards Max M

 

[Duncan Booth]

The data must be stored in a plain text format that would allow direct
access if the main system was not functioning. This doesn't necessarily
imply the ability to circumvent a web based layer by editing text files
directly, though there does need to at least be an easy way to export
and import such files for people who want to use a Real Editor [tm].

If you use Zope/Plone then you can set up a ZEO server with multiple Plone
frontends. Then if for any reason your main point of access is down you can
access the data through an alternate frontend. You could for example have
the main site which gives read-only access to everything, and a secondary
site for those people with permission to contribute. These could be running
on completely separate Zope servers so if one is taken down the other is
not affected.

Giving access to edit via a real editor is a separate issue from allowing
access when the main system is down. That can be done by a variety of
methods (such as the ZopeExternalEditor product).

 

[Aahz]

aahz> The system needs to work with browsers that have limited
aahz> functionality (i.e. Lynx).

Isn't the functionality in Lynx a bit *too* limited? For text based
browsing there is w3m, which can render legibly a whole lot more web
pages...

w3m does a bit better at the visual presentation at the cost of much
functionality in navigation. If forced to, I usually use links because
there's a rudimentary JavaScript engine available for it.

 

[Shalabh Chaturvedi]

The team for www.python.org is taking another crack at considering a
content management system for handling the website's content.
[snip]

Right now, Plone/Zope2 is the primary competitor simply because it's
well-publicized, but none of us really has that much experience with
websites whose core maintainers are going to be programmers. ;-) We'd
like to know if there are other packages available that can do what we
want; we're more interested in getting information about each package
than in package comparisons. We're also interested in hearing about
experiences creating custom systems.

My experience creating a custom system with Quixote was very good. It is
simple, appears to be meant for developers and is easy to pick up. If
the current www.python.org is built using some kind of text to html
scripts it should be straightforward to port it to Quixote.

Quixote is not a CMS. It is a nice interface between Python classes and
the web. It does have session management. All CMS-like features would
have to implemented, but if you're dealing with reading and writing text
files, it should not be hard to do. Consider that Zope has a fairly
steep learning curve and customizing it by people not familiar with it
might take a substantial effort anyway. Though I must admit that I
haven't used Zope for a couple of years.

[snip]

 

[Tim Hoffman]

Hi Aahz

I have been actively devloping with Zope for 4 years now and I think it
is a good solution. Admittedly most of the applications and sites I have
developed with it have not utilised Plone but almost all have been based
on CMF, the work I have done with plone makes me feel it is a good
platform to build/base python.org on.

The only fly in the ointment I see (at least for an out of the box
Plone/Zope deployment) is the requirement for the content to be in the
filesystem.

Having said that, if you configure Zope with Directory Storage (
http://zope.org/Members/htrd/DirectoryStorage) then
the content would in fact be in the file system, so even that issue
is resolvable without much effort.

Zope/Plone workflow I think is more than adequate for what you want,
as well as its security capabilities. It just needs a little planning.

Lots of people say the learning curve is steep. Well I feel like any
sophisticated piece of software that is different from what you know,
has a steep learning curve, especially if it solves problems in
radically different ways. Certainly for the sort of customisations
I could envisage you would want to make to a basic plone/zope config
it would not be likely to tax anyone too much, and I am sure there are
more than a few Zope/Plone gurus around who can lend advice. Most
competant people I have worked with have been able to be quite effective
with Zope in a fairly short space of time.

my 2c worth.

Tim

P.S. I could probably volunteer some of my time (except whilst I am
sailing from Darwin to Broome in a 40' cat next month. ;-)

P.P.S Plone won't cause a problem for Lynx. Have a look at the plone.org
site with Lynx to see.

The team for www.python.org is taking another crack at considering a
content management system for handling the website's content. This is
particularly important considering, for example, how far beind we are in
posting job ads -- getting the community to do our grunt work for us is
the primary motivator.

Right now, Plone/Zope2 is the primary competitor simply because it's
well-publicized, but none of us really has that much experience with
websites whose core maintainers are going to be programmers. ;-) We'd
like to know if there are other packages available that can do what we
want; we're more interested in getting information about each package
than in package comparisons. We're also interested in hearing about
experiences creating custom systems.

Here are the primary criteria that we're working with:

The data must be stored in a plain text format that would allow direct
access if the main system was not functioning. This doesn't necessarily
imply the ability to circumvent a web based layer by editing text files
directly, though there does need to at least be an easy way to export
and import such files for people who want to use a Real Editor [tm].

Workflow is presumed to be a hierarchial publish reject model (in other
words as simple as is possible to begin with). We'll need both
coarse-grained and fine-grained security (giving people a variety of
access levels to large and small sections of the site).

The system needs to work with browsers that have limited functionality
(i.e. Lynx).

 

[Aahz]

hmm! Why on earth would anybody want to do that.

"I surf faster than you do, monkey boy."