CMS-Cutre, the command line CMS-hack

CMS-Cutre is a dirty proof-concept tool for creating a maintaing a web repository in a easy way with a CMS'like command line tool.

I'm using this tool for maintaing my web. This not mean you could find it useful in anyway. Maybe future versions will have some kind of end user Q&A, but don't hold your breath.

Rationale

We assume everybody have the right of publishing on the WWW. Putting a html file on the web is really easy but having a tree of documents, updating it, and maitaining it with others could be a so complex work.

There are lots of software efforts for solving this problem, many of them with libre software. But the most of them are server side software, so complex for installing and configure, plenty of doubtly useful features, tied to database servers, plenty of code (multipying the possibility of bugs and security holes), unflexible architecture, managing information in a, in many times, unnecesary dinamic way, spending CPU cycles and obligued to create and maintain contents with the software features instead of the availaible powerful desktop tools. So complex, so unnuseful.

Many users (maybe only me?) want and need another way of work for creating and maintaing good quality webservers more easely, using the filesystem features as the content container and the power of all client site managing and editing software.

Concepts

• The filesystem is the content containet, you can organize your content in a tree form with the tools availaible in your system.
• Use always XHTML.
• XHTML can wrap every file type. You can manage any type of content creating a subdirectory with an index.html referencing it.
• You can add CMS links using HTTP redirecting. Indeed you should wrap «links» with HTML too.
• Each XHTML file should be XML valid ones, with the required <title> tag.
• Organize your contents directory by directory. That's mean: you have a directory tree. Each directory should have an index.html. That file could be hand maded or automade. If you made it, you are responsible for having a correct XHTML subtree. If you want it automade, you should have an INDEX.xml file (see example). The script will read it contents and all the html files on the directory and all the subdirectories with an index.html inside, it'll take their <title> information and build a index.html file with the correct linking.
• The syntax of the INDEX.xml file is so easy, it should have an &lt11> tag with the title of the directory and one or many <p> tags containgin arbitrary correct XHTML code. The INDEX.xml should be well-made XML.

Requirements

• Command line oriented
• Semantic web aware
• XHTML based.

Tools needed

• tidy
• xmlgrep
• bash :-)

Licensing

This tool will be libre software. At this moment is no software :-)

Show me the code

That's your problem. Remember this is a dirty shell script hack.

Look at the public CVS server.

Do you have a working example too. For regenerate the index.html file run $ cd ejemplo/ $ cms-cutre.sh > index.html And done!

TODO

• A comprehensive way of managing directories
• Commad line options
• A configuration file

Future enhancements

• Command line tools oriented to end-user
• exporting RDF/RSS
• blog-like feature
• automatic reporting of badlinks and/or lost links inside the repository
• GUI using zenity
• Adding xslt auto-filtering middle step

How modify the website or have visit feedback

For a lot of uses the static websites are better in most of the ways. Anyway, having some kind of feedback in a clean, free and easy to maintain way. We are considering the use of remote or local web-annotation servers for this.