Vignette StoryServer
2000-09-27 15:28:25+00 by
Dan Lyke
6 comments
John Tesh's website company sues Vignette over StoryServer misrepresentation. I've been looking around at content management concepts and come to the conclusion that very few of the commercial products for publishing static sites are any better than a good Perl coder can whip up in a week.
[ related topics:
Business Web development Content Management
]
comments in ascending chronological order (reverse):
#Comment made: 2002-02-21 05:30:21+00 by:
DaveP
I've come to pretty much the same conclusion. Frontier is closest to what I want, but
on the Mac, it wedges about every 4k hits or so. Slash and Scoop are kinda close, but more discussion-oriented than
they are content-management systems. The stuff Phil Greenspun uses to
run Photo.net isn't bad, but
there's a lot to wade through, and it's a little more tightly coupled to
Oracle and AOLServer than I'd like.
Sigh. If you find something that doesn't suck, I'd sure like to
hear about it.
#Comment made: 2002-02-21 05:30:21+00 by:
Dan Lyke
One of the issues I have with Frontier and the Ars Digita schemes is that they're so based on rendering on the fly that they need way too beefy machines to back them up. Especially since modern operating systems have much of the basic web serving in the kernel, what most web publishers need is a tool that helps them build better static content, including richer linking and link verification, and automated linking to groups of documents. In the background I've torn apart my system which uses XML and am trying to return to using text files, but giving better interfaces to CVS and some interactive apps that help resolve citations and such. But Flutterby infrastructure is a back-burner project right now.
#Comment made: 2002-02-21 05:30:21+00 by:
DaveP
Agreed. The render-on-the-fly part of Frontier is why I have Frontier
running on a G3 instead of a slower Mac. Ars Digita seems even
worse, but it's hard to say for sure. I'd be interested to see the Ars
Digita code running on top of
MySQL and Apache so I
could instrument it all and have a better chance of figuring out
where the bottlenecks are.
On the other hand, having some sort of web-interface so people
unfamiliar with the server, and uncomfortable with command-lines in
general can post content seems like a pretty darned important thing.
One of the directions I've been looking at is having a CGI that will
eat new content, render it once to an html file, and then store it.
The problem that leaves is the ability to easily go back and fix a
typo or formatting issue.
Luckily, I'm also working on a better front-end for CVS for version
control within my
business so I don't have to think hard about the version-control
bits when I start a new project. I'm hoping that'll end up helping out
my web-needs at some point.
Oh well, back to the paying work....
#Comment made: 2002-02-21 05:30:21+00 by:
Dan Lyke
A CGI that both saves the original data in-place and renders that data to the final HTML seems the right way to go, one of the things I learned in retrospect from Newwwsboy (the software that originally ran Flutterby) is "don't throw anything away".
And as long as you're not rerendering on every page view an SQL database makes a fine filesystem, although that doesn't let you CVS the changes.
Speed issues with Flutterby right now come from not using mod_perl
effectively, connecting to that database takes a long time. I also need a better solution to page rebuilding than my once every 5 minutes cron
job which scans the XML files to see if any of them have changed.
Speaking of CVSing changes, is anyone working on context diffs for XML and SGML documents? Something that understands tag structure seems like a really good idea...
#Comment made: 2002-02-21 05:30:21+00 by:
ebradway
I believe Onsale has been using a rendered-static model for quite a while (and quite effectively too). The auction engine I wrote would also generate the static HTML as needed. Usually there is some event that triggers a regeneration and it should be possible to tell what needs to be rewritten.
But Dan is right-on here. Unless you are truly front-ending a database via the web, you probably don't need fully dynamic content. Oracle goes to the other extreme by building the web server into the database - but they are really trying to supplant the OS.
The OpenACS guys have been rewriting ACS to use Postgres on the backend but it still requires AOLserver and competency in TCL.
#Comment made: 2002-02-21 05:30:21+00 by:
sethg
Joel Spolsky's new company is apparently going to rent out consultants to tackle content-management issues. I don't know if they're planning to release any of their code, though....