This file contains notes about the status and progress of TBWiki
------
{{TableOfContents}}

Note that some test materials are on a separate wiki at:
http://bird.org/cgi-bin/test.cgi/FrontPage

Before working on this wiki any more, re-read this article by a developer
who has written 7 personal wikis:
 * https://borretti.me/article/unbundling-tools-for-thought
   * the basic idea is that there are better tools for thought than personal wikis
     * but that's OK - tbwiki is a hobby project that is fun to work on

Next major items to work on [edit section]

= Next major items to work on =
 * support captions 
   * if caption="foo bar" is an attribute of an image, use figure, and figcaption to surround the image.
   * see https://www.w3schools.com/TAGS/tryit.asp?filename=tryhtml_figcaption_css

• MacroCategory.py - show a list of pages in a particular category * implement category tags, like this: {{{ {{Category(Tasklist,2023)}} {{Category(New Feature)}} }}} * implement category reference, like this: {{{ {{ShowCategory(Tasklist)}} or {{Category(show:Tasklist)}} }}}

 * MacroCategory.py - show a list of pages in a particular category
   * implement category tags, like this:
{{{
 {{Category(Tasklist,2023)}}
 {{Category(New Feature)}}
}}}
   * implement category reference, like this:
{{{
 {{ShowCategory(Tasklist)}}
or
 {{Category(show:Tasklist)}}
}}}

• Blog processor * support adding a blog entry * (done) ExportSite processor * Search macro: {{{ {{Search(word or phrase)}} }}}

 * Blog processor
   * support adding a blog entry
 * (done) ExportSite processor
 * Search macro:
{{{
 {{Search(word or phrase)}}
}}}

• do performance enhancements, possibly using the Scalene profiler * add a "blog" processor * for ideas, see: https://jekyllrb.com/docs/posts/ * save blog pages as: blog-YYYY-MM-DD * don't support comments * show index of blog posts, current post * research blog layouts: * see Blog Layouts and Features for ideas * figure out why I keep getting logged out (is there something wrong with the cookie expiration date?) * don't add an extra line when saving. * use splitlines instead of split to avoid adding an extra line on an edit?? * there appears to be no split here * do some testing to see where the extra \n gets introduced: * client side or server side? * in form population (send from server) or in form response data (sent from client)? * is there a way to detect a new newline, and remove it? * grep split tbwiki_engine.py | grep [\\]n * get css from page area of local wiki * don't retrieve from /var/www/css * Support wiki on mobile devices * (done) fix up mobile theme * change mobile.css and mobile.footer and mobile.header * some pages still show very small fonts on my phone * autodetect mobile phone browser, and use mobile theme automatically * user UserAgent field * switch to wsgi - I considered flask, see FlaskResearch * should switch to wsgi - I have experience with that * enhance blog system (see TBWikiToDo#Blog) * add page lock (to support multi-user) * add help routine to all processors - add SystemInfo processor to handle help requests - with help action item, that calls help with a processor name - processors only have to return a block of tbwiki markup * make tables robust * eliminate save_table_data - placing support in appropriate routines in the io_handler (tb.handler) * one of write() or some other class method. * fix bug 017. See BugList * fix bug with edit-in-place on QuarterlyData table * make automated test suite for testing page rendering * fix all known bugs, with all known table types * make a table test matrix * table flow control * for transcription program (ProcessorTranscriber.py) * make a view form and edit form for the transcriber database * read documentation on view and edit forms * add processor_name to the transcriber database config * add support for hybrid tables * use case is to add a spreadsheet formula to the FIXTHIS pages, to show totals * ability to have tables with records from different sources * add io_handler per-record * need permission checks (readable, writable) per-record

 * do performance enhancements, possibly using the [[https://pypi.org/project/scalene/|Scalene profiler]]
 * add a "blog" processor
   * for ideas, see: https://jekyllrb.com/docs/posts/
     * save blog pages as: blog-YYYY-MM-DD
     * don't support comments
     * show index of blog posts, current post
   * research blog layouts:
     * see [[Blog Layouts and Features]] for ideas
 * figure out why I keep getting logged out (is there something wrong with
   the cookie expiration date?)
 * don't add an extra line when saving.
   * use splitlines instead of split to avoid adding an extra line on an edit??
     * there appears to be no split here
   * do some testing to see where the extra \n gets introduced:
     * client side or server side?
     * in form population (send from server) or in form response data (sent from client)?
   * is there a way to detect a new newline, and remove it?
   * grep split tbwiki_engine.py | grep [\\]n
 * get css from page area of local wiki
   * don't retrieve from /var/www/css
 * Support wiki on mobile devices
   * (done) fix up mobile theme
     * change mobile.css and mobile.footer and mobile.header
     * some pages still show very small fonts on my phone
   * autodetect mobile phone browser, and use mobile theme automatically
     * user UserAgent field
 * switch to wsgi - I considered flask, see [[FlaskResearch]]
   * should switch to wsgi - I have experience with that
 * enhance blog system (see [[TBWikiToDo#Blog]])
 * add page lock (to support multi-user)
 * add help routine to all processors
   - add SystemInfo processor to handle help requests
     - with help action item, that calls help with a processor name
     - processors only have to return a block of tbwiki markup
 * make tables robust
   * eliminate save_table_data - placing support in appropriate routines
   in the io_handler (tb.handler)
      * one of write() or some other class method.
   * fix bug 017.  See [[BugList]]
   * fix bug with edit-in-place on [[QuarterlyData]] table
   * make automated test suite for testing page rendering
   * fix all known bugs, with all known table types
     * make a table test matrix
 * table flow control
   * for transcription program (ProcessorTranscriber.py)
     * make a view form and edit form for the transcriber database
       * read documentation on view and edit forms
     * add processor_name to the transcriber database config
 * add support for hybrid tables
   * use case is to add a spreadsheet formula to the FIXTHIS pages, to show totals
   * ability to have tables with records from different sources
     * add io_handler per-record
     * need permission checks (readable, writable) per-record

Things to do (random ideas, un-triaged) [edit section]

= Things to do (random ideas, un-triaged) =
 * don't allow caching a page edit page
   * use http meta cache-control directive to prevent caching
   * set caching relatively low anyway
 * add page locking for edits - to prevent collisions
   * lock a page for 15 minutes when an edit is started
     * create a lock file in the admin directory
       * put user, start time, and page hash in lock file
     * start a timer to handle lock timeout
   * generate a hash of the page when the edit starts
   * if someone else tries to edit, give them a warning that they have lower priority - indicate who has the lock
   * when someone submits a page, compare the page hash and lock
     * if page hash doesn't match, don't save the data
     * remove the lock
   * after 15 minutes, remove the lock
 * support arbitrary field addressability in json data:
   * JsonPageFile:field.list_field[n].foo.bar
 * fix bug: edit-in-place wipes out conf entries in inline block

• add 'tbclient list' feature * modify tbclient to send 'list' request * modify ProcessorDataHandler.py to respond to list request * add TODAY() function to spreadsheet

 * add 'tbclient list' feature
   * modify tbclient to send 'list' request
   * modify ProcessorDataHandler.py to respond to list request
 * add TODAY() function to spreadsheet

• add websed over snapshotted values * add websed over diff of snapshotted values * add chart over snapshotted values * use plot charts? * add access-controlled content * the purpose of this is to add a page with contact information for other test projects, on the fuego wiki * see https://moinmo.in/HelpOnAccessControlLists * at the block level, with ProcessorAccessControl * at the page level, with accesscontrol directive * put required permissions at top of block * prevent reading or writing to block (or page)

 * add websed over snapshotted values
   * add websed over diff of snapshotted values
 * add chart over snapshotted values
   * use plot charts?
 * add access-controlled content
   * the purpose of this is to add a page with contact information for other
     test projects, on the fuego wiki
   * see https://moinmo.in/HelpOnAccessControlLists
   * at the block level, with ProcessorAccessControl
   * at the page level, with accesscontrol directive
   * put required permissions at top of block
   * prevent reading or writing to block (or page)

• add link to page for file column in attrdb and parsed data files (to go to page) * This is handy for FIXTHIS entries, and for editing * _bug0013 - can't edit-in-place the buglist

 * add link to page for file column in attrdb and parsed data files (to go to page)
   * This is handy for FIXTHIS entries, and for editing
 * [[_bug0013]] - can't edit-in-place the buglist

Major Features that are not complete enough to use [edit section]

= Major Features that are not complete enough to use =

User accounts [edit section]

== User accounts ==
 * Creating and administering user accounts
   * need to re-test user accounts, with new features (next_num)
   * Users can create accounts, but physical access to the server is required to change admin flags, or to remove an account
     * may be able to use a special version of attrdb (raw page edit) for user accounts.  It wouldn't be great but would get the job done for now.
     * is the user edit form not working??

Table usability [edit section]

== Table usability ==
 * Specific table usability features:
   * other usability feature is deciding which links to use for a table (I never use query, and always have to decide between 'edit-in-place' or 'edit')

BugList, FIXTHIS, Regressions [edit section]

= BugList, FIXTHIS, Regressions =
 * See the [[BugList]] page for bugs that need to be worked on.
  * that bug list is from a database.  Here is one that just text on a page:
     * [[Manual Bug List]]
 * See the [[CodeFixthisList]] page for a list of FIX''''''THIS items in the code
 * See the [[PageFixthisList]] page for a list of FIX''''''THIS items in the pages

{{{#.ToDoData
||Date      ||bug_count||FIXTHIS_count||missing_FEATURES||total||
||2017-03-08||15||xx||xx||=$.bug_count + $.FIXTHIS_count + $.missing_FEATURES||
||2016-08-12||14||94||2||=$.bug_count + $.FIXTHIS_count + $.missing_FEATURES||
||2016-08-02||13||92||3||=$.bug_count + $.FIXTHIS_count + $.missing_FEATURES||
||2011-04-21  ||17||78||1||96||
||2008-10-06  ||17       ||78           ||1  ||=$.bug_count + $.FIXTHIS_count + $.missing_FEATURES||
||2008-02-04  ||13       ||66 (55 cgi-bin+ 11 plugin)||3||82||
||2007-07-25  ||9        ||62           ||3?||74||
||2007-02-06  ||9        ||41           ||5 ||55||
||2006-11-15  ||8        ||35           ||6 ||49||
||2006-10-06  ||8        ||37           ||7 ||52||
||2006-08-25  ||6        ||35           ||5 ||46||
}}}

Table of combined issues [edit section]

== Table of combined issues ==
{{{#!Table
source_spec=#:.ToDoData
cols=Date:bug_count:FIXTHIS_count:missing_FEATURES:total
sortby=Date:rev_alpha
is_spreadsheet=1
field=Date
default_value=%(timestamp)s
}}}

----------
== Regression count ==
Periodically, check [[http://bird.org/test/Regression_Test|the regression test page]] to see if things still work as expected.

Here's a table, manually maintained, with regression testing history.
{{{#!Table:regression_history
||date||version||commit||OK||missing output||bad||bad page list||
||2016-08-12||1.6.2||a45cce77||91||2||2||TestMarkupVariables, TestMarkupVariablesDyamicTest||
||2016-07-18||1.5.3||61eb2f1e||39||31||4||TestMarkupVariablesStatic, TestSandbox, tbwiki.footer, tbwiki.header||
}}}

Other items to work on, by category [edit section]

= Other items to work on, by category =
== Style ==
 * enhance themes feature
   * see [[DocThemes]]
     * for an example of self-editability - see http://kernelnewbies.org/navigation_menu
   * calculate theme from user account or other variable
     * current use of page name is lame
   * allow configurable navigation links
   * allow configurable toolbox links
   * read page dressing from nested pages??
     * allow header and footer pages to be HTML or TBWIKI format
        * if TBWIKI, allow for sub-page references that are
        loaded, parsed and formatted recursively
 * tweak mobile.css
   * see w3c pages about automatic layout - http://www.w3schools.com/html/html_layout.asp
 * switch to MediaWiki page syntax?? (not yet)
   * see http://meta.wikimedia.org/wiki/Help:Wikitext
   * need to switch bullets handling
   * need to switch table parsing
 * have css come from page area??? (would allow editing css using wiki)

General [edit section]

== General ==
 * convert upload to be a processor
   * allows putting upload on any page
   * control number of files via config
   * can specify link(" action=upload&dest_filename=foo") for missing files - to prompt user if file is missing
   * upload result should be shown in page message?
 * add page rename
   * can do create, move content, delete, but that's awkward
 * generate tbwiki stats from log
 * cache page content, for static content
   * change all prints to output routines (to control routing of output)
 * allow administrators to view and edit files in the 'admin' directory
 using normal tbwiki access (just like regular pages)
     * treat it like a MoinMoin overlay directory
 * make a mobile viewer
 * add ability to grab content at specific intervals
   * kind of a combination of crontab and websed
   * allows for charting something automatically over time
   * make a crontab processor
     * present a form for when to run, and action
 * make a syntax for block names that are not included in block output
   * we have anonymous blocks, hidden blocks, and named blocks
     * named blocks show both the name and the content.  It would be nice to hide the name
   * possibly: #.BlockName ??
 * make tbwiki more relocatable (for use on laptop)
   * on bird.org:
     * move *.cgi scripts from www/cgi-bin to ~/work/tbwiki/cgi-bin
     * put link in www/cgi-bin
     * add base_dir to *.cgi scripts
   * write script for laptop that does:
     * downloads latest tbwiki and tbwiki-data
     * adjust base_dir in *.cgi files
     * adjust base_url in *.cgi files
     * adjust plugin symlinks in tbwiki-data dirs
 * make unit test system - see [[Unit Test Notes]]
   * NOTE - already have ProcessorRegressionCheck.py
   * have a page specify a test
   * have markup, and html and compare them
      * here's one unit test:
{{{
       record=bold 1
       markup='''bold'''
       html=<b>bold</b>
}}}
   * make a ProcessorUnitTest.py processor
     - read all "UnitTest.*" pages, convert markup to html, compare to html
     - report results
 * put this categorized todo list in a table?
 * change main() in tbwiki_engine.py
   * call main when tbwiki_engine.py is called from command line
   * make a tbwiki.py command?
     * program should have the ability to create a static page from the command line
       * e.g. tbwiki.py FrontPage -o FrontPage.html
     * program should have the ability to run unit tests?

Security [edit section]

== Security ==
 * check permissions on table write operations
 * allow administrators to view and edit files in the 'admin' directory
 using normal tbwiki access (just like regular pages)
     * treat it like a MoinMoin overlay directory
 * have rights affect the displayed table edit links

Processors [edit section]

== Processors ==
 * add help routines for all Processors
 * add a processor for "select an item"
   * used for choosing a group of words to practice
   * used for choosing a story to read in scriptures
   * present a list of items from a database, that are unselected
     * allow a user to select one
   * or, select an item randomly from a list of database records
     * allow user to select it or keep choosing a different one
   * have 3 categories: unselected, in-progress, finished
     * database must "state" field?
   * processor allows only legal state transitions
     * specify legal state transitions (or is this too meta?)
 * support mapping routines (e.g. to convert chapters to pages and back)
 * ProcessorScriptureLookup.py - maybe should be a macro?
   * take a reference, and replace it with the text for the versus
   * this should be very handy for church talks, lessons, etc.

Blog [edit section]

=== Blog ===
 * add right sidebar with:
   * recent posts
     * categories
     * search bar
     * tags
 * use individual pages for each post
   * create add-post form
     * page schema has title, content, date, tags, id
 * navigation to previous and next posts
   * response area
     * show threaded responses with "reply" button on each one
     * response form with "post comment button
 * show date, tags, response count
 * see example: https://lava-block.com/godot-engine-tutorial-making-a-2d-game-ui-mechanics-sound/
 * support customizable background image
 * ability to drag and drop images to upload to wiki

ProgressChart [edit section]

=== ProgressChart ===
 * add support for goal lines to the ProgressChart processor
 * handle chart attributes
   * make chart_attrs_class, to generate defaults when attrs are missing
   * create flat namespace for chart attributes
   * translate chart atttributes into call arguments during chart creation
 * un-specialize pychartwrapper.py
   * chart another data type using same wrapper??
     * generic multi-series data chart
     * scripture reading chart?
     * weight chart
     * running chart??
 * support data entry and modification for TimsRunning data
   * see ProcessorProgressChart.py
   * use Add Form for table??
     * support current date for date entry??

SlideShow [edit section]

=== SlideShow ===
 * SlideShow
   * make sure all nested parsing works
   * add sidebar list of slides (section 1 headings), with links to pages
   * support show attributes and per-slide attributes (bgcolor, bgimage, margins, etc.)

ImageGallery [edit section]

=== ImageGallery ==
 * ImageGallery
   * support links=1 - to show individual images in full screen
   * autodetect image size for full-screen (standalone) image gallery
     * use Javascript to detect browser window size??
   * support num_images - to show only some of the images
     * gallery has prev and next links on sides (similar to blog navigation)
   * support image borders in ProcessorImageGallery
   * support link to fullsize image in ProcessorImageGallery
   * support displaying the filenames in ProcessorImageGallery
   * add javascript zoom functions 

Tables [edit section]

== Tables ==
 * add sortby=select_order - to specify that text fields should be sorted
   in the order that they appear inm the select possible_values list.
   * this helps the user control the sort order for things like word-based
    priority: ex: very high, high, medium, low, very low
 * add messages to table module
   * allow for outputting messages about an individual table
   * table.add_table_message(msg)
   * this would be handy for table-specific error messages
   * should have block-specific messages in general
 * allow for configuration of table control links
   * have user specify which links to show, and where
 * use better filenames
     * use unique part of filename for record_id
     * allow user to pick filename when creating the record
 * support edit-in-place for tbwikidb files
   * use get_section and set_section
     * modify to read/write definition lines
 * eliminate difference between record, record_id
 * use tbwikidb or attrdb for user database??
   * this puts human readable filler on every page, but that's the wiki way
   * this allows creating a multi-file database just by creating the template
   * this is fragile - human editors will break the pages parseability
 * add table query cheat sheet to query table form page
 * make prototype family info database on bird.org
   * automatically create a User account file on login???
   * make MemberXX records for family members
     * determine the schema (what fields)
     * create a table view with contact information
 * table queries and views FEATURE
   * support <table>RecordView
 * support read-only attribute on whole table
 * Support time-sampled database
   * schedule a websed operation that grabs data at crontab-like intervals
     * or that creates a new record whenever the page is accessed (and schedule something to access the page automatically at periodic intervals)
 * support automatic data snapshotting
 * write a 'table creation wizard' processor
   * ask for field names and types, and create the schema file and cover file (or template file, if its a tbwikidb)
 * document table config concepts:
   * base config (from original conf block)
   * dynamic config (from request form variables)
   * query config (from query conf block)
 * saved queries
   * specify query list in conf block
   * each item in list specifies a query conf block
 * make backup file when editing table files
 * finish all routines for editing a table
   * use {{{Table(arg1,arg2,arg3...) notation for inline tables
     * support other args:
       * formspec, query line (row filter, column filter, etc.)
 * document Table use with data_scan module
   * [[DocTables]] has some information (one example, the FIXTHIS table)
 * multi-file tables
   * display multi-line values correctly
   * display '<no-data>' values correctly
   * auto-sense type of each field (text, select, textarea) and use
     for auto-form generation
 * handle field types with enumerated values (lists and radio buttons)
 * document table architecture and interfaces - see [[Table Architecture]]
   * document table attributes
   * document chart attributes
 * make ProcessorTablePick.py
   * See [[Table Pick Page]]
   * present simplified list of "unpicked" items
   * allow user to pick one item
   * save picked item with attribute "state" changed to "in-progress"
   * present detailed list of "in-progress" items
   * allow user to mark an item as "done"
     * save done item with attribute "state" changed to "done"
   * whether user wants to have multiple items in "in-progress" state is their choice
 * queries, query_lists 

Forms editing and Templates [edit section]

== Forms editing and Templates ==
 * use form emitter for user accounts.

Users and attributes [edit section]

== Users and attributes ==
 * Support online creation of new users
   * make a "create user" form
   * create a new user account file from a form
 * provide help if user forgets password
   * link to e-mail the password??
   * password hint??
 * create 'user.edit' action, which edits a user account file
 * support different roles (admin vs. non-admin)
 * allow user to edit their attribute values
   * make an "edit user" form
 * REFACTOR - use existing form code for user forms
 * make user account management it's own processor module
   * ProcessorUser.py - we now have sub-actions, so this should work.
     * it needs to access table operations, but in a special way

Long term ideas [edit section]

= Long term ideas =
 * add 'save expected output and show again' to RegressionTest processor
 * add link to backup page on history page
 * look at other wikis for feature ideas
    * in particular, [[http://ikiwiki.info/|ikwiki]] statically compiles the wiki on every page edit
 * documentation
 * table editing: 
   * should support "save and edit again" and "save and add another row" buttons for saving edit forms.
 * Table Of Contents
   * do auto-table of contents, like wikipedia??
   * handle included files and text expanded from processors??!!
 * check HTML and CSS on multiple browsers with BrowserCam or browsershots
   * see http://browsershots.org/http://bird.org/tbwiki/FrontPage
 * make a journal processor (based on progress chart)
 * make a blog processor (with threaded comments)
 * make a forum processor (with threaded and voted-on comments)
 * make a page cache
   * add dependencies to a makefile, and rebuild cache when dependencies change
 * put config into admin page?
   * reduce cgi-bin to just some directory references

Things Completed [edit section]

= Things Completed =
 * 2021-03 Remove '?action=save' from url in edit form
 * 2018-10 Add image rotate to ImageGallery app.
 * 2017-04 add triggerd - ability to detect triggers that are due and execute them
 * 2017-03 add ProcessorTrigger - ability to trigger a snapshot
 * 2017-03 add ProcessorSnapshot - page/block/table snapshot function
 * 2017-02 add tbclient - ability to get,put and edit items in the wiki
 * 2017-01 add page trails
 * 2017-01 add relative index to cell address (so you can do $.-1)
 * 2016-12 ProcessorImageGallery - finish showing images in full size page
 * 2106-11 add ability to customize table operation links (edit_links, sort_links)
 * 2016-9  add ability to read a specific block, by number
 * 2016-08 convert to per-table messages (for debugging and errors)
 * 2016-08 support editing sections with nested sub-sections
 * 2016-08 allow default_values with generated values (%%(timestamp), %%(next_int), %%(next_page_num))
 * 2016-08 add page delete
 * 2016-07 add tbwikidb database type
 * 2016-07 fix data_scan module to parse multi-line fields
 * 2016-05 make ProcessorImageGallery
 * 2016-05 make ProcessorBlog
 * 2016-03 make better diff output
 * 2016-03 put engine_version in data_class, and add to footer
 * 2016-03 add "printable format" option for page
 * 2016-03 made a test wiki at: http://bird.org/cgi-bin/test.cgi/FrontPage
 * 2014-05 make talk pages optional - see config.has_talk_page global config
 * 2010-10 support inline tables - FEATURE
 * 2010-?? websed - add websed feature to scrape other web pages
   * 2010-?? data-scan - add as a processor first, then make into a database grabber
 * 2008-10 table - back up page on table edit
 * 2008-09 chart - draw chart of table data
 * 2008-08 processor - add a charting package (pychart)
 * 2008-02 table - added support for sorting table rows
 * 2008-02 table - support col_list and row_filter specifiers
 * 2008-02 table - added ability to edit whole table at a time
 * 2008-02 macro - create Include macro
 * 2008-02 table - separated conf into separate class
   * add field_db (parsing, handling)
     * add field types, colors, possible values
   * add routine to read block from page
 * 2008-02 table data - separated IO handler into separate class
 * move tbwiki with authentication to bird.org
 * aug03 - add user accounts.
   * put user account files in an 'admin' directory
   * authenticate user from files in 'admin' directory in action user.login
   * look up user by name to process login form (in action user.login) 
 * jul07 - add cookie code for user accounts
 * feb07 - use pages for page dressing (see tbwiki.header and tbwiki.footer)
 * jan07 - support file upload FEATURE
 * 11/15 - add support for inline images
 * change backup routine to save current version of page instead
 of old version of file
 * save edit comment and show in history
 * edit sections FEATURE
 * parse and format:
    * bullets
    * headers
    * horizontal rules
    * paragraph breaks
    * preformatting
 * write edit form for page
 * save edited data to file
 * read page from data directory and emit as html
 * write handlers for various actions
 * create add_record form (using existing edit_form code)
 * write mtable.save_form() routine
 * re-write edit_form to present whole table with edit links
 * support Table internal macro, with args
 * modified backup time in file extension to be human-readable
 * wrote mtable.modify_record() routine
 * wrote mtable.remove_record() routine
 * parse and display URLs as links
 * use pages for page dressin
 * make items in table of contents into links
   * make an anchor target for each heading
   * use href for anchor target in TableOfContents macro
 * make nicer format for table of contents box
 * add search form to header
 * support multi-page table source_spec
   * support e.g. "User*" as the spec
     * read all pages starting with User
       * finish parse_table_pages() routine
         * read each page matching the source_spec
         * parse out values using a page template
           * start with simple name=value parser
     * create page on add record
     * write page on edit record
     * remove page on remove record
 * support history action
   * show history from backup times
   * allow diff between two versions