Python Popularity: Questions and Comments

[Paul Everitt]

I think there's a lot of merit to what you (Bill) say below.  Zope is a 
big project in scope, and though there is a lot of documentation, it is 
scattered and not organized as it should be.

The current Zope (2.x) also suffers from more of a "one size fits all" 
approach, where content authors, web designers, and site developers are 
all wedged into the same basic box.  Thus, documentation and GUI suffer 
from having different audiences.  All in all, Zope has a higher learning 
curve than it should.

Tackling this basic issue is the goal of Zope3:

   http://dev.zope.org/Wikis/DevSite/Projects/ComponentArchitecture/

In particular, read "VisionStatement" on that page for a pretty blunt 
statement showing we're in tune with your points.  Zope3 actually has 
documentation, interface files, and unit tests being written *before* 
code.  All in all, Zope3 is proving to be much more Pythonic, more 
modular, more explicit than implicit, etc.

This doesn't preclude the need to do this for Zope2.  I'm interested in 
hearing more about the Pet Shop example you mentioned.  Is it a 
canonical example that is written up somewhere else for a different system?

All in all, Zope is doing quite well.  I'd say, with all the books out 
and coming out, we're ahead of where PHP was at a similar age.  Also, 
it's important to note that around half of the people using Zope had 
never used Python before.  It's a significant achievement.  Like Python, 
Zope has a large and active community with bright and pleasant people 
all around the world.

I'd also like to point out that Zope isn't a straight-up competitor to 
J2EE.  In many ways, Zope competes more with Vignette, Interwoven, and 
other content management systems as much as it competes with application 
servers.  This explains why appealing to content authors and site 
designers has been on our radar.

It also explains why I think WebWare, Quixote, and other Python web 
systems are complementary to Zope.  It's a big market with lots of 
audiences.  And we all need Python to succeed.

In fact, with Zope3 I'd like to see more ways for us to collaborate with 
other Python web systems, as there is a lot of good work going on.  With 
Zope3 being more Pythonic and more modular, there's a better basis for 
two-way sharing.  But only if other projects are also interested.

2002 is going to be a critical year for Zope.  On the one hand, we have 
the validation from multiple books and some nice customer success 
stories (hooray, we can finally talk about CBS!).  Also, Zope3 combined 
with Python 2.2 will make Zope more friendly to Python developers.  On 
the other hand, we either win back people like you this year, or we face 
the thought that we've peaked.

Bill, can you send me an offline note and explain the Pet Shop example?

--Paul

Bill Tate wrote:

> Ron Stephens <rdsteph at earthlink.net> wrote in message news:<3C2B9E04.9957801F at earthlink.net>...
> 
>>I recently read an interview with a Digital Creations top manager where he said
>>that he hopes someone starts a successful business based solely on providing
>>ISP services using Zope, with full technical support. Is anyone looking into
>>this? Does anyone think this could be a successful business model? I am
>>interested in opinions on this.
>>
> 
> This past spring, I spent a full month pouring over Zope in detail.  I
> pulled every example product I could find that I thought would be
> relative to the purpose of setting up a commercial-based application. 
> I downloaded the tutorial and the draft zope book.  I pulled down
> every bit of documentation I could find for developers.  In so many
> cases, I found whatever documentation was available that key aspects
> of the implementation took the "assumed facts-already-in-evidence"
> approach; omitting key details that would were critical to improving
> my understanding of using Zope.  I consider myself a masochist in the
> sense that I'm willing to go through various contortions to figure
> something out, but like many others, I reached a point of saying that
> I still don't get it.  I won't even touch the issue of ZOPE's CMF or
> Zope Templates - the documentation on that couldn't be more confusing
> if it tried.
> 
> My point is, the product may be perfect for my needs, but please do
> away with tutorial examples referring to using zoo animals to
> illustrate object navigation or elvis sightings to solidfy
> understanding of the product.  There were simply too many disjointed
> pieces in the documentation to allow me to pull all the pieces
> together and believe me when I say I'm a person who will go to great
> lengths to try to "get something."  If I'm having trouble, than I
> suspect more than a few others are as well.
> 
> Any commercial enterprise is likely to be supported by either some
> data model or object-class model that almost invariably has a RDBMS on
> the back-end for persistent storage (I've also worked with ODBMS so
> ZODB isn't an issue for me).  If you want to gain converts to zope,
> build a tutorial that follows some resemblance of a what a normal web
> site has to provide by way of example - perhaps something akin to the
> Pet Store example in Zope would be useful???  How about building an
> example auction site from scratch?  I would be more than happy to
> devote time and energy contributing to Zope but I can't until I figure
> out how Zope works.  To some degree, I think my confusion with Zope
> increased AFTER I went through the documentation.  Truthfully, I don't
> even know whether Zope would be even a reasonable candidate for doing
> something akin to the Pet Store example or even an auction site for
> that matter.
> 
> PHP has virtually exploded on the scene by providing a everything
> including the kitchen sink - it appears to have taken little
> convincing to gain converts to PHP and accordingly to the latest web
> surveys, its implementation is huge.  I much rather see tools like
> Zope achieve that kind of status. But I don't imagine that too many
> folks are going to be willing to work that hard to achieve Zope Zen
> when other alternatives exists and that's a damn shame because I
> suspect Zope is an excellent product.
> 
> I truly do commend and appreciate DC's contribution's to Python, but I
> hope somebody can point to something (either examples or
> documentation) that gives me a reason to try again.
> 
> Truthfully, I just don't think its me.
> 
> Bill Tate
>