[Zope] - Suggestions for Zope 1.9.0

[Jose.Lacal@icn.siemens.com]

---------------------- Forwarded by Jose Lacal/Service/ICN on 01/28/99
15:06 ---------------------------


Jose Lacal
01/28/99 12:36

To:   Paul Everitt <Paul@digicool.com>
cc:
Subject:  RE: [Zope] - Suggestions for Zope 1.9.0  (Document link not
      converted)

<!-- Lotus Notes is not very good at handling SMTP mail, so I had to
re-format the message. ->

>Jose wrote:

>> b.) The ocumentation (installation and initial configuration)
>> *needs* to be
>> improved from just a set of dysfunctional README.txt and
>> WEBSERVER.txt files scattered all over the place.
>
> Like Python, this is an Open Source project, so we'd love
> specific suggestions and fixes.

It is understood that Digicool is not expected to provide all the goodies.
Rather the community must contribute their effort to improve the product.

>> c.) For those interested, I have compiled a step-by-step set
>> of instructions for installing Zope on a Linux box, along with
>> the PyGreSQLDA. The file is available in PDF format. The
>> document is still in draft form, but it's usable.
>
> Great, as soon as you're ready and have a URL, let us know
> and we'll link to it.

Here it is: http://www.volks-pc.org/document/zope.pdf This is my
contribution to the community.

The document is still a *draft* so be gentle with it. The purpose of the
doc is not just to explain *how* to do things but also *why* you are doing
what the document instructs you to do. I do Customer Support for a living,
so the documents I write tend to be geared towards a non-techie audience.
Some of the Zope mailing list's members may take offense to the overt
Windows bias of the document. Sorry for that.

>> - - - - -
>>
>> 2.- Positioning.
>>
>> a.) Zope's website states that: "Zope is a free, open source
>> web application platform used for building high-performance,
>> dynamic web sites." That definition, IMHO, is not even close
>> to what Zope
>
> Well it's _one_ of the things Zope is :^)

?? "The Lexus 400 takes you from your house to the office in exactly the
same time as a Volkswagen Beetle does, traffic gridlock permitting." Yes,
that is one of the things the Lexus 400 does, but that doesn't mean Lexus
is going to position the 400 as just doing that.

>> is. Plus, it is a very dry, unattractive Marketing one-liner.
>> No offense intended, but it certainly looks like a site by
>> nerds and for nerds.
>
> Hmm, if I did a nielsen rating of people visiting the site,
> I'm sure very few people would come out as "Joe Sixpack the
> TV Clicker".

It's a circular question and answer issue: if all you have in your website
is a techie-focused message, that's all the audience you'll ever get. If
you have different messages for different audiences, your website will be
able to attract and retain the attention of a much more varied audience.
Take a look at the OpenSource site (http://www.opensource.org) They have
three clearly labeled "messages" or "positioning statements" in marketing
lingo:

"This site offers several complementary views of the open-source
phenomenon. You can read a brief introduction, a techie/hacker's case, a
businessperson's case, and a customer's case. Still not convinced? Then
read some third-party case studies."

So, I propose to the community that the Zope site should presents at least
three "messages" to its visitors:
- The technical view: here you could go bananas with all the nitty gritty,
techie-focused stuff
- The ISP view: why ISPs should look at Zope as a cost-effective web
hosting platform or whatever
- The corporate view: how corporations can use Zope internally to
web-enable their techno-phobic users
- The end-user view: etc.

Mind you, these "messages" or "complementary views" of Zope are in addition
to and to complement the case studies / testimonials Hadar is collecting.
The testimonials fit perfectly under each one of the "views" above.

> Managing a community is always a delicate balancing act
> between lavishing over your existing customers (taking care
> of your base) and appealing to new customers (growing your
> base).  Since Zope started with the Python/Bobo communities,
> they deserve the gold treatment.

Agreed, just don't give the "rubber treatment" (boot-in-behind) to new and
potential customers. It takes just a little more effort, not much money.

>> b.) It took me a week of playing with Zope (installed under
>> both NT Server 4.0 and Red Hat Linux 5.2) and reading the
>> manuals before I started to understand the magnitude of
>> Zope's power. And I am still not done exploring its features.
>> This is not a good thing for potential new users who are not
>> hard-core techies.
>
> Again, contributions welcome. Can you point me at a community
> web site, such as say Frontier or Python, where the content
> quickly described the power?  We can then create by example.

The OpenSource site comes close to describing the power in a very
audience-focused way, as I described above.

BTW, I don't like the Python site because it is very techie-focused as
well.

>> c.) I would suggest positioning Zope as a web-based,
>> FrontPage-like web development environment that does NOT
>> require the user / creator to have any software locally on his
>> PC. Using Zope, anybody with access to a PC and the network
>> (Intra / Internet) can publish their own website by using
>> Zope's formatting / structural capabilities. "You focus on
>> the art and the content, and Zope will handle all the
>> nitty-gritty for you."
>> Or something like that. Sell the sizzle, not the steak.
>> People will not likely buy Zope because it's OpenSource, or
>> because it's written in Python. Corporations will buy Zope
>> because it allows non-programmer users to get their job done
>> faster, cheaper, and with less hand-holding from the
>> corporation's MIS / IT crowd.
>
> Ugh, go after FrontPage is an enormous challenge: Zope doesn't
> have a powerful client-side program.  Think of all the
> client-side power FP has that we'd have to create: we don't
> even have search-and-replace in the "editor"!

That is one of my points: not everybody *needs* the full power of FrontPage
to do a basic web page with some database linkages. FP gets in the way of a
quick and dirty website. Zope, on the other hand, could be deployed
internally by a company to allow Joe Schmuck 9-to-5 to whip up a quickie
site for the boss before lunch, and with some nifty linkages and search
pages to the corporate SQL database.

> I _think_ what you mean is focus on what Zope *does* and what
> content managers *see* rather than on *how* Zope does it?

Precisely. Sell the sizzle (what the results are, how simple it is to use
it) rather than the steak (dumping the full techno cart of OpenSource, OOP,
Python, extensible, blah blah) on the poor, unsuspecting Dilbert boss type.

> I think you also touch on making Zope useful to people that
> don't even know HTML.  This is a common theme for "community"
> applications.  It is also the focus of a document-repository
> approach for Zope, where it manages/publishes
> previously-created artifacts.

Here you go again: "a document-repository approach" and
"..artifacts." I suggest you say something along the lines of:

"Install Zope on a cheapo Linux box, let the Linux wenie define the SQL
plugins to your Mucho Grande corporate SQL database, and then let your
clueless users put all their exsiting files on the corporate intranet
without teaching them neither HTML nor buying 200 copies of FrontPage." And
you don't need to hire Andersen Consulting to do this for you. (Forward a
check to Digicool at the end of the exercise, of course).

> Still, going head to head with Front Page is hard.  I'd prefer
> to see the results of someone that tried it by creating such an
> environment.

?? Linux and the entire OpenSource movement going head to head with The
Bill is hard ;) I think you might want to flip that argument on its head:
FP is an over-bloated, expensive and inflexible tool to do a quicki site,
or to allow non-technical users to produce quick results. And those two
types of users are over 90% of the audience out there.

>> d.) Furthermore, Zope allows for the creation of web-enabled,
>> database-driven applications from a web browser. That is, no
>> more grudgery writing ODBC stuff and such. This product could
>> find a serious following among the corporate crowd: techies
>> define the ODBC links to the corporate databases, and then
>> Joe Schmuck writes his / her own web-based queries to the
>> databases using Zope, and then publishes the results to the
>> corporate Intranet.
>
> This was the first incarnation of "Principia": as an
> environment for a database integration tool called "PyDB" that
> became "Aqueduct".

Yes, I have the manuals for Aqueduct. Very confusing docs, BTW.

> What do you think the requirements are to compete in this space?

First and foremost, change your "message" / "positioning statement."
Fine-tune it to several types of audiences. It's the same product, just
positioned in different ways to satisfy different needs.

Second, I would focus heavily on the centralized-management,
user-freindliness of the system. Meaning: set-up a central server,
distribute the editing and content management. From Principia Manager's
Guide, page 08:

"The scenario in this chapter involves the creation a fictional web site
for Plutonia Incorporated. Stan has been appointed
the task of creating the public web site for Plutonia to give the public
online information on its products, research, and
news. Stan is also busy as head of the Products department and does not
want to presume anything about the Research and
Public Relations (PR) departments, so he needs to delegate responsibility
for content to the departments themselves. Yet,
he also needs to make sure that the site looks and behaves similarly at all
points, and some common, consistent
information needs to be displayed on all pages. He also needs the ability
to make any change at any location.

Additionally, he needs to create a public feedback page where a visitor to
the site can send an electronic mail message to a
specific department to request more information or give a comment.

Each department is responsible for its own content. Each department has
varied groups and committees overseeing
different areas of that department. Finally, each department requires
individual control over who can be assigned
privileges to add and manage contents in their sub-areas."

THAT is a very good business scenario where Zope can come to the rescue.

> For instance, the wonderfully named wizard tool that was just
> posted here might cover some requirements.  What else is needed

I have not played with it yet, but it sounds like that is the way to go,
IMHO.

> to be credible?  And why have a relational database involved
> instead of Zope's database?

Come on, Paul. How many corporations out there have something called
"legacy systems?" You didn't expect people to dump their Sybase / SQL
server / Oracle databases and port their business-critical apps to Zope,
did you?

You *want* Zope to become the "universal glue" that ties all the heavy-duty
back-ends to a friendly, idiot-proof, no need to buy a license per user a
la FP, tool in the market today!

> What you're describing is similar to ColdFusion I believe. This
> is the target of the current ad campaign in Linux Weekly News. I
> *know* that some will say, "Don't pigeonhole Zope that way!"
> I'm not doing that yet, but I know that ColdFusion strikes a
> raw nerve, and I'm just the kind of guy to prod that raw nerve.
> :^)

I have never used ColdFusion. I do believe that an OpenSource, Python-based
solution could take a good portion of the market if there are other
products that have already created such a market.

>> e.) Yes, I _know_ most of the issues above are already
>> discussed in Zope's website. If only after you go over the
>> Zope Features, Questions and Answers, Zope Components, Zope
>> Zen Revealed, and Testimonials sections. What I am saying
>> is summarizing all of those benefits in one paragraph,
>> maximum.
>
> And that paragraph would be how many pages? :^)

No longer than 04/05 lines.

Again, there are at least 04 distinct markets that need a customized
"message" for each one of them. If you are interested, I can send you more
specific pointers to create such customized messages for each audience.

>> f.) Web presence providers should love this product for
>> their users. Using Zope to design and maintain a website is
>> much better than using FrontPage and then dealing with the
>> broken links and upload hassles!
>> With Zope the user can backtrack when he deletes some portion
>> of his site. With FrontPage, what's gone is gone.
>
> Hmm, not if ODBC is used.

Of course, I didn't mean the data you deleted / added / etc. I meant the
pages you over-wrote with an older version, the graphic files you
inadvertedly deleted, etc. Editing mistakes, in general.

>> g.) I know, you guys tremble at the mention of comparing Zope
>> with FrontPage. But, hey, in Marketing you need to consider the
>> issue of educating your prospective customers. Instead of
>> wasting time and resources educating the marketplace on what an
>> "open source web application platform" is, why not piggyback on
>> the widely known concept of what FrontPage is? The Bill has
>> already spent millions of bucks promoting what a FrontPage-like
>> app us supposed to do. So, let's take that and build a Zope
>> definition on top of the pre-existing, widely-known FrontPage
>> definition.
>
> Perhaps because the FrontPage definition doesn't fit?

Zope is not a FP replacement / direct competitor, that I understand. What I
meant to say is to pisition Zope as a tool that allows end users to create
/ manage killer websites *without* the need to use FP. So, you piggy-back
on FP's name recognition by claiming that end users can a lot of the things
FP does without actually buying it, and with more centralized control that
FP will ever give a corporation or web hosting firm.

> I mean, he's spent more building the brand for Word -- why don't
> we call ourselves open source Word?

Well well, that was not nice. Zope has very little in common with Word, yet
a lot of the functionality of FP.

> There is a very definate market for an "application server".
> This market is *booming* and is becoming clearly defined by
> analysts, researchers, journalists, and reviewers. In fact, the

?? "analysts, researchers, journalists, and reviewers" did not design the
Sony Walkman, or Dodge Caravan. Those two products created entire new
markets where none existed.

> Web Techniques issue containing Amos' article was on "application
> servers".  I believe we'll start striking "web application
> platform" and just settle in on "application server".

As you wish, that is one area where Zope could find its niche.

> Note that MS is spending tons of dollars advertising their
> application server, as is Netscape, as is Sun, as is...

Good, we both agree on the benefits of piggybacking on somebody else's
advertising bufget.

>> h.) Buried on page 24 of the Principia manager's Guide we find
>> the following text, which I suggest should be on Zope's main
>> page:
>> "Netscape Publishing
> [snip]
>
>> Since Principia also targets the content manager, it would
>> make sense to let the content manager manage his content
>> using Netscape authoring too ls. This section describes how
>> to use Netscape publishing with Principia."

> Right.  There were bugs in the support for it that I *think*
> are fixed for 1.10.

So that's why I am unable to Publish from Communicator.

> Netscape Composer isn't the best IDE in the world.  For instance,
> it doesn't do HTML forms, much less style sheets or DHTML or ...
> But it certainly is a good start and I know some people here
> have tried to use it.

A lot of people in corporations already have access to Communicator.
Therefore, installing a Zope server will suddenly empower all those people
to become web publishers in a flash at US$0 additional cost to the
corporation. Does that sound like a good business proposition to you?

>> 3.- Workspace improvements.
>>
>> a.) Add the size and date of last update for each file
>> listed.

> Good.  I suggest submitting feature requests for these in the
> Collector (www.zope.org/Collector/)

I already did so.

>> b.) When in the "View" tab, there are no tabs on top. The
>> user needs to hit the Back arrow. This is consufing and
>> inconsistent with the rest of the tabs' functionality.

> I disagree.  When people "View" their page, they want to see
> *exactly* what their page will look like and want to see
> *exactly* what the HTML will look like.
>
> There is no way that we could get the tabs in there -- the page
> would have to have two <HTML> statements, kind of a document in
> a document.
> We could have *three* frames but I don't think that would work.
>
> I view it this way: the web browser has a UI that people have
> standardized on.  The concepts of back and forward are very
> common themes in web browsers.  When in Rome...

OK, no reason to argue when it's an issue of personal taste.

>> - - - - -
>>
>> 4.- Good to have.
>>
>> a.) A full, Python-based HTML editor to create/edit HTML
>
> Uh, I don't think *anybody* wants to learn another editor, I
> think they want Zope to work with their current editor. We
> *certainly* don't want to write an editor.

Not everybody has currently access to an HTML editor, you know? Just an
idea, not critical.

>> files via Zope . I think there are some options out there
>> already. It's just a matter of integrating the code into

> "It's just a matter..."  Do you realize what an enormous amount
> of work you just referred to? If you disagree and feel that it
> isn't much work, then please send some patches.

Who, me? I'm not even a programmer. I am a busines person. And I do realize
the amount of work involved in doing new things to existing software: I get
paid to support the customer when the software "fix" to the customer's
problem causes more problems than there were before the "fix" was
installed.

>> Zope. One of those options is HTMLgen. Curren tly found at
>> http://starship.skyport.net/lib.html this library of classes
>> mirrors the HTML 3.2 markup elements.

> I'm not sure you've looked at HTMLgen.  It is in no way an editor.
> It allows people writing Python code -- hardly the Front Page
> audience you describe above -- to generate HTML from within their
> Python code.  It covers the same ground as DTML, though with the
> reverse focus (a code-centric view of markup rather than a
> markup-centric view of code).

Told ya, my programming ignorance comes shining through. Sorry for the
misleading comment then.

>> That's all, folks. Keep up the good work.

> Thanks for taking the time to send such a thoughtful note!

Hey, you guys put a lot of effort into Zope, which I got for free. The
least I can do is give you some feedback on how to improve your baby's
chances of becoming a huge success.

Regards.

- - - - -

Jose C. Lacal, Senior Engineer, Data Services
Siemens Information and Communication Networks
900 Broken Sound Parkway; A2 - Boca Raton, FL 33487
Voice: +1 (561) 955-3081 - Fax: +1 (561) 955-6500
Text page http://www.skytel.com/Paging/1way.html
PIN # 1258390  -  jose.lacal@icn.siemens.com