An FYI on some completed site fixes.

Most of this magic is in the perl code in these two files:

 - /openejb/site/trunk/lib/view.pm
 - /openejb/site/trunk/lib/path.pm

## Bad // in breadcrumbs

Fixed the code that creates breadcrumbs links that had '//'

## APIs used

Added some code to create a list of APIs used in each example.  Basically greps 
all the *.java files and pulls the imports, then creates a list of links which 
currently link to the JavaDoc for those APIs at oracle.com

Possible improvement: If we think we're up to it, we might start documenting 
them better here.  Perhaps a page for each API class with some hand created 
text to describe it in a paragraph, then a list of examples that use that 
annotation.  Perhaps even show small snippets (3-5 lines of code) from each.

We'd need a small army of excited documenters to make a dent though.

## Sharing links

Added Google+, Facebook, and Twitter links to make it easy to share pages on 
the website. 

Possible improvement:  Thought it might be neat to have a "comment" link right 
under those buttons.  Something that would pop up an email addressed to users@ 
with the subject filled in using "Comment: $title" and a link to the doc in the 
email body.

## Sitemap.xml

Had already added a bit of perl to generate a sitemap.xml, though it didn't 
list all our content.  Fixed that up so it is properly recursive and now 
finally lists all our example pages.

Also added a robots.txt which lists the sitemap.xml.  Hopefully that will 
improve our search support which is all driven by Google.  If Google hasn't 
indexed the page, it won't come up in the search.  So orphaned pages would 
never show up.  Now they will as at least they'll be listed in the sitemap.xml.

You can also use that file to rank some pages higher than others.  I have any 
page under 'examples' ranked higher so they are revisited more often.

## Including content from other pages

Had also added a confluence-like {include:foo.mdext} "marco" to read in other 
files and include their content.  It wasn't taking into account the links in 
that content and that they might need to be adjusted.  If you include from a 
doc in a child directory, then the relative links need to be updated.  Fixed 
that.


-David

Reply via email to