[Hypercontent-developer] HyperContent 2.0 status

[Alex V. <av...@co...>]

Hello all,
	I just wanted to provide a status update on the development of  
HyperContent 2.0.  As of last night,  I have created a new CVS module  
"hypercontent2" which contains the current code and library  
dependencies for HyperContent 2.  This is very much a work in progress;  
there are no build scripts currently, and I have not checked in any  
project content to demonstrate the new server.
	The way to run HyperContent 2 is through a new class,  
"org.hypercontent.HTTPServer", which uses Tomcat in embedded mode to  
offer up the server.  HyperContent 2 is backwards compatible with  
version 1.x projects, but the structure has changed: project definition  
files now reside within the project repository, at  
"/config/project-definition.xml", and another file  
"/config/permissions.xml" is used to configure permissions.  I will  
check in sample content in the near future.

A quick overview of what is there already in the HyperContent 2.0 code  
base:  a standalone web server using JAAS for authentication, which can  
serve up static or processed data via "pipelines", and supports  
advanced browser caching via the HTTP1.1 specification, including  
support for If-Modified-Since and Etag headers, as well as the basic  
Last-Modified, Content-Length and Content-Type headers.  Projects are  
mapped to base paths within the servlet context, so that the URL of a  
file looks something like  
"http://www.server.edu/hypercontent-servlet/project-mapping/file-dir/ 
file.ext". Recursive dependency resolution has been implemented, so for  
example a file which is transformed by an XSL stylesheet which imports  
another XSL file will be properly recognized as out-of-date if the  
imported XSL file is changed.  Disk caching has been implemented using  
Berkeley DB Java Edition, so that dynamic processing of a page is only  
performed once after server startup, until any dependency of that page  
is modified. As of right now, HyperContent 2 can act as a dynamic web  
server for any HyperContent 1.x project content. HyperContent 1.x style  
batch processing has not yet been migrated over, but will be soon.

HyperContent 2.0 is introducing a new type of view, which I've taken to  
calling "modal".  Any form of content presentation that was supported  
under HyperContent 1.x is considered a "static" view, such that it is  
mapped to a plain URL, and is a candidate for batch processing and disk  
caching.  Modal views of content will be mapped to a "mode" parameter  
in addition to the basic URL, and will not be capable of being built in  
a batch request or cached to disk - they are exclusively dynamic.  This  
type of view will be used for management and authoring tools.  I am  
currently working on integrating the Velocity templating engine as an  
alternative to XSLT, and with either templating language you will be  
able to import other modal views, such that modal views can be  
assembled as components and brought together in various combinations.   
For example, a directory browser will be a single component that might  
be reused in different contexts, such as a 1.x style management screen,  
or in an "open file" dialog that might pop-up when you want to insert  
an image in a WYSIWYG area.

We're still a few weeks off from doing an alpha release, but the code  
is now out there if anyone is interested in taking a peek.  I'll be  
happy to help anyone with further details on the migration process, as  
needed.

Cheers,
	Alex