[REBOL] VSITE: Re: Re: Wanted: a Libmaster - no pay, but glory
From: jason::cunliffe::verizon::net at: 15-Sep-2002 10:09
Just been checking out Zope news. 'ZopeLabs' is a community script
library/cookbook site. Nice pattern structure:
Comments: [multiple entries added via web form]
For example: "Adding users using your own forms "
The site [cookbook] is actually a Zope 'product' [that zopespeak for
plugin/module/component] running on Zope of course. It is freely downloadable so
Zopistas can learn/hack/improve it.
Zope is based arounf the ZODB [Zope Object Data Base] which is published to web,
objects and methods being accessible as standard URL. Sections of a Zope site
including Zope 'products' can be imported and exported as single binary file
[object] through the web. That allows one to transfer and install anything from
a single GIR image up to complex collection in the ZODB's tree. This can scale
to be an entire site. The process is the same.
Translating this to REBOL and in particular Vanilla, it would be akin to
import/export of a set of snips, dynasnips and perhaps Vanilla lib modules. For
those not familiar with it, Vanilla snips are modular content files, each with
implicit, but invisible .metadata files.
I have been experimenting past few days manually transfering snips and dynasnips
from on Vanilla site to another. Very nice. On my todo list now are some scripts
to manage the job. For example:
- publishing links of new snips to another site,
- find and fetching [import] snips or links to snips from a remote site
- reconciling matching names or user-id conflicts
- chckcking snaip and dynasnip dependencies. Lik any translusion system, this
can soon get hairy. But basic checks are simple and useful.
The use of Vanilla for community web library might be helped by adding a new
metadata field = 'cookbook:' and writing a dynasnip [or vanilla library script
in /lib] as cookbook application with methods to manage snips with "cookbook:"
metadata. That field might be configured as:
id: 83823 ; a unique database-type value to act as content checksum and
type: 'entry' [author--somewhere--org]
category: ['cgi' 'forms']
Other cookbook meta-types might be:
'idea' ;idea for a script not yet developed
'comment' ;free seqential list of comments, tips etc.
'errata' ;versioning or other code specific variations/fixes
'article' ;An intro/essay/tutorial discusseing rebol technique and/or use of
entries in the cookbook library
'example' ; other code which uses rebol library. For example progams using
anamonitor, or View scripts using FTP, CGI or image processing.
Cookbook code would be a combination of apps/index.r and apps/calendar.r
Calendar is Vanilla's blog application and includes a basic comment structure.
Blog entries are named based on date
That makes it easy for calendar.r to find and link based on name and metadata.
Cookbook snips might be named along similar lines.
As you prbably know, I am big fan of Vanilla. I believe it makes an ideal
platform for community rebol web-based library/cookbook. Being written in rebol
it encourages developers to extend it. Vanilla makes a great showcase for Rebol
to the rest of the world. This is important to promote and demonstrate rebol. If
actions speak louder than words, the implementations convince where sales blurbs
and evangelism cannot.
It has a very elegant modular design which fits rebol thinking.
Vanilla plays nicely with Dreamweaver, CSS, XML, SWF, etc..
It needs work and features, but is already serviceable.
I believe several others are working towards similar goal [Petr?]. Let's work