Subject: Re: [TS] CGI vs. Zope
From: Ben Darnell (ben@thoughtstream.org)
Date: Mon Aug 14 2000 - 23:59:35 EDT
On Tue, Aug 15, 2000 at 12:22:41AM +0200, Marcus wrote:
> I haven't used Zope, but the documentation already said that it wasn't
> simple to install.
I've never tried to install Zope by hand, but installing the Debian
prepackaged version couldn't be simpler. I don't think Zope would be
any harder to install than, say, Apache. (although it can get kind of
complicated if you want Apache and Zope to cooperate)
>
> I looked at WebDAV and that seems interesting, especially given the
> fact that even M$ is readily supporting it, or at least providing help
> to the developers.
WebDAV is on track to become as ubiquitous as CGI (possibly more so,
since once the software matures it will not pose a security threat,
while CGI is inherently somewhat risky). Microsoft is really pushing it
- IE5 includes a WebDAV client (called Web Folders), and the latest
version of IIS supports WebDAV. On the Open Source side, Zope supports
it, and there is a module for Apache (which will be integrated into the
core for Apache 2.0). There are even free WebDAV-enabled web hosts
(e.g. www.mydocsonline.com)
>
> In any case, my vote goes to a widely supported and easy to setup
> platform base for the initial web version. Specialisation with it's
> possible speed and feature advantages later.
I agree here (although the speed of the initial CGI version is likely
to be ridiculous).
New design-of-the-month ;-):
Forget the structural changes I described in "Musings from France" (only
one content object per idea). Everything is structurally as it was
before (see the bottom half of http://thoughtstream.org/design.php3).
Each idea is stored in a separate XML file, which can be accessed
through the filesystem or through WebDAV (WebDAV is preferred; direct
filesystem access should only be used when there is not a WebDAV server
available, as WebDAV provides better support for concurrent access from
multiple clients) (data accessed through WebDAV need not, of course,
actually be stored in one-file-per-idea, but the WebDAV server should
present each idea as a separate resource (i.e. Zope)). These files
could be concatenated for import/export as a single XML file. Other
than that, things are generally the same as they were before.
Some other design issues:
Should configuration be done inside the system, or outside? That is,
should one be able to edit templates, create users, etc within the
ThoughtStream interface, or should these tasks require a separate
interface? I'm torn between visions of World Domination (TM) in which
ThoughtStream is used for everything, and fears of trying to pound a
square peg into a round hole. Any opinions?
-Ben
-- Ben Darnell ben@thoughtstream.org http://thoughtstream.org Finger bgdarnel@debian.org for PGP/GPG key 1024D/1F06E509
This archive was generated by hypermail 2b25 : Wed Aug 30 2000 - 22:01:01 EDT