
-- 0.7.0 --

- Sierra/Fionna integration

- identity management ( we need a proposal for this )

-- 0.6.5 --

- Reptile should migrate to using mod_aggregation timestamps and source
  information

- Reptile should support aggregation of a remote blogroll feed as your own
  personal weblog.  This way I can keep my blogroll local with records-mode and
  then use Reptile to index everything.

- Spend some time dealing with JXTA issues that WILL arise.

- there is NO reason that Reptile needs to take up 80M.  Try to debug why it is
  taking so much memory.  I need to be able to get this down to 20M or so.

    - do some basic memory footprint benchmarks.  How much does it cost us to
      keep a Tomcat instance... what if we trim it down.

-- 0.6.0 (peerfear.org priority) --

    - Write a search for all articles that reference a specific search.  AKA
      blogdex URL-> post resolution

    - Implement the ability to view Task Logs from the website

    - different toolbar when running in rendezvous

         - (DONE) add documentation link

         - add source link

         - mailing list link
         
         - (DONE) remove manage subscriptions

         - (DONE) remove publish links
         
         - (DONE) donations at the top of the page.

         - Need an API link at the top which point to Axis WSDL and associated
           documentation

    - (DONE) Anakia documentation under /reptile/documentation
                    
    - Fix all reptile xdocs when running under Tomcat so that icons are setup
      correctly
             
    - ability to index all RSS feeds regardless of subscriptions/categorization.
         
    - migrate to running this under mysql     

    - enable the SOAP service

        - Deploy the AXIS servlet and org.openprivacy.reptile.services.ReptileSearchService

        - Write some quick unit tests...
    
    - Panther cached articles.

    - disable page actions

    - Top 40 style operation.

    - finish port (and test) mysql on peerfear.org

    - enable task logging for channels and status (need documentation on this)

    - look at my REP API from the paper and see how it relates to the API
      exposed by the search service.

    - WeblogSearchProvider should NOT be an option in rendezvous mode.

    - FEATURE: News by category

    - Advanced search pages:

        - search by URL
                  
-- 0.6.0 --

- Need an action to mark all search results as read.

- Look at the MetaWeblog API...  We might want to use this in Reptile as an
  alternative publication mechanism.

- We need to fix this bug of loosing the database.  This is a HUGE problem.  I
  don't know what to do here.  Under Windows this problem still happens with
  around > 1 connection.

    - Try reverting back to the hsqldb.jar that I originally used.
      
- JXTA integration

- (DONE) Include a TASK_LOG that should include

    - (DONE) identifier of the task
    - (DONE) time of last update
    - (DONE) last status (failed, successful)
    - (DONE) classname of the exception (if any)
    - (DONE) stacktrace (if any)

    - Need a Task garbage collector to remove old Task status entries.

    - Need to disable logging of some Tasks (IE the TaskLogGCTask)

HTCache integration

Decode content before it gets into the DB.  I don't want to see any more of this:

    &lt;p&gt; &lt;a name=&quot;10018102&quot;&gt;&lt;/a&gt; &lt;b
    class=&quot;byline&quot;&gt; &lt;br&gt; There has been much more discussion
    about the OSX 10.1.3 update on various boards. There are enough unrelated
    disasters where I would still say that new OSX users should wait for a while
    before upgrading; I'm certainly going to wait at least another few
    weeks.&lt;/p&gt; &lt;p class=&quot;byline&quot; align=&quot;ri...

- include a ChannelInitializer which

    - puts the channel into the cache

    - runs meta-update... if it succeeds marks the channel (initialized)

        ( if we are going to run meta-update... what about subscribed vs
        non-subscribed articles )

- OCS export?

- Need the ability to run a search on specific subscription layers.... should
  this just be a narrowing request?

    - How would we provide a UI for this?

- The search page (mozilla-search.xsl) should support ... category / history ...
  right now we have history... we don't have categoory.

    - This is because we aren't updating the SUBSCRIPTION

- SECURITY HOLE.  Reptile could potentially receive a javascript: URL that could
  do bad things.  Gaurd against this in meta-update and in rss-1.0.xsl.

- try to update the time and date of a channel from the dublic core info
  (assuming it exists?) ? What formats are date fields.

- We need to provide statistics on channel updates

    - last update

        - FAILED|SUCCESS
        - number of new articles found
        - failed/success ratio

- When reptile links to URLs that redirect... use the redirected URL NOT the
  source URL.  We need this so that one site doesn't have our whole click stream

- Implement localhost only access

    http://jakarta.apache.org/tomcat/tomcat-4.0-doc/config/engine.html#Request%20Filters
      
-- 0.5.5 remaining issues --

- I think we are having a Classloader issue with Tomcat.  Migrate to 4.0.4 and
  move Xalan BACK where it belongs!

    - release these changes into an a6

- Make sure we dont use the Apache channel as the default (even if this is no
  longer subscribed).  Right now there is a BIG FIXME for this.

---------------------------

-- GENERAL --

- Need a BrowseArticles and BrowseChannels query... this should support listing
  all articles and channels so that we can just browse throught them.

- BUG: can we place a unique index on ARTICLE.TITLE, ARTICLE.CHANNEL.  There
  should be NO reason why a channel would ever have the same title twice and it
  is becoming irritating when moreover does it.

- (TODO) Support a 'related' query that finds all articles from the same host.

-- LONG TERM PROBLEMS --

- new weblog 

  - How do we subscribe to a remote weblog through JXTA?

      - when we do a 'Peer Search' we should have an option to 'view weblog of
        peer'

      - this will require us to have a table for user metainfo.  Specifically
        'name or handle', 'description', 'interests', etc.  We would need to
        include this metainfo in JXTA advertisements.

            - DONE: right now we have this as a setting.

      - I also think should have a quick overview of the remote peer:

            - their metainfo, name, title, description

            - most recent articles and channels they have found

            - top 5 items in their weblog with a link to view their weblog.
        
  - how do we find articles found within X and Y... IE (what did you find for
    slashdot.org between this time and that time)

        - this query is fine.  Just need to specify a criteria of

            select * from article where date_found > DATE1 and date_found < DATE2

- search

  - What EXACTLY does the XML protocol interaction between two JXTA services
    look like?

        - TODO: we are going to need some way to serialize deserialize objects
          to XML across a Pipe.  Can we do this with DOM4J or XML4J or that SUN
          project?

- (TODO) how do we search across JXTA advertisements to find FEN.  Can we add a
  message?

    - it looks like we could add Dublin Core to the Advertisements.  Go ahead
      and do this and try to make sure that we get this information in the
      advertisement cache and that we can still this info back into the DB (so
      that we can do searches on it).

    - (TODO) If we DO decide to stick this info in the DB we are going to need
      to make sure that we garbage collect it.

    - (TODO) We need a JXTA_ADVERTISEMENT table with a DATE_FOUND column so that
      we can GC them.
  
- (TODO) How do we make Reptile join the correct JXTA peer group.

    - This is just a matter of making sure that we look at the source.

        - TODO: Can I join multiple peer groups?

- (TODO) We need a proposal for coming up with the HTCache.

    - How do we cache locally... How do we know if the content is fully cached?
      (including all resources)?  How do we transfer the WHOLE cache including
      all resources to a remote node?  Maybe we should just limit ourselves to
      caching locally for right now.

    - Reptile should have a HTCACHE table that contains URLs that are cached by
      the Panther HTCache.  This way we can quicly find out if we have a
      document in cache and return it.  The cache would ONLY be populated when
      HTCache returns successfully.  

- (TODO) Need a proposal for Reptile X-node caching.

    - Should we search in the current peergroup for URLs that are cached and
      use that instead?

