ref: 90545cad35115eb6c1164c0d4eede407d0e712cd
dir: /sites/werc.cat-v.org/development/todo.md/
Plans and Ideas for the Future
==============================
This page lists various assorted ideas and features that have been proposed at
some point or another, nothing (except fixing the listed bugs) is assured to be
implemented, many things here might turn out to be bad ideas.
Features
--------
* Review ideas from: http://www.w3.org/TR/chips/ and http://www.w3.org/Provider/Style/URI and see if we can do better.
* Canonize and redirect all requests, we already do this for things like missing or trailing /, should do the same for /. and /./, etc. Deleting trailing . and , should make pasted urls in emails work as links (**Mostly done in RC1**)
* Blog:
* Comment threading.
* Some form of pagination (taking advantage of history browsing.) (**Partially done**)
* Sitemaps:
* Index pages should be used for description of directories.
* Cache web sitemap generation. (**Done in RC0**)
* Layout improvements:
* Need to find a way to rig the order of items in the sidebar.
* Right sidebar.
* Top+left/right nav vars (breadcrumbs at the top, current dir listing on the side)
* Make it easier to disable all sidebars and headers/footers, eg., for full screen mode.
* More testing and optimizations for mobile browsers, eg., http://www.operamini.com/demo/ Left sidebar had some minor issues in cellphones at some point.
* Better page titles (include whole path hierarchy?) (**Partially done, is it good enough now?**)
* Allow utf-8 characters in path elements (need to make sure this is safe, and not sure how useful it will be)
* General code:
* Write a regression test suite. (*Started*)
* Replace all references to non p9p/p9 programs (*Done?*)
* Document better the 'API' for sub-apps, both a set of environment vars apps can rely on (and in some cases set) and functions they can call (WIP).
* Better generation of descriptive HTML META tags, eg.,: {META name="description" content="This is the Google Summer of Code blog for Plan9 and Inferno projects."} {META name="keywords" content="google, summer of code, inferno, plan9, programming"}
* Should add werc to certain wikipedia lists:
* `http://en.wikipedia.org/wiki/List_of_content_management_systems`
* `http://en.wikipedia.org/wiki/Comparison_of_wiki_software`
* Include txt2tags as builtin alternative to markdown.
* Spam protection:
* Mathematic pseudo-CAPTCHA ala WP's 'Math Comment Spam Protection Plugin'.
* kfx (via cinap) suggests you just put a hidden text input box (or textarea?), "Dont write in this input box: [ ..... ]", if the bots fill it out ignore the post
Future Plans and Blue Sky
--------------------------
* Wanted [apps](/apps/):
* Tagging, eg., append to _werc/tags
* Related links sidebar generation.
* [Bug/issue tracking app](/apps/gregor/).
* Hg/git repository browser.
* Implement AtomPub.
* Further modularization and extensibility (meta-templates) (Is the new app framework enough for this?)
* A way to wrap external cgi applications (eg., existing hg/git browser)
* A way to generate static sites (ie., use werc as an offline templating system)
* [9P](http://9p.cat-v.org) interface.
* New (simplified) markdown implementation, ideally in C or awk (maybe smu/libsmu by gottox?), yiyus great md2html.awk is a great start!
* Simplified
* Tables
* Images
* Autogenerated anchors/navigation boxes
* No inline html
Similar frameworks to be investigated and mined for good ideas to steal:
* TinyTim: http://www.reddit.com/r/programming/duplicates/dbaee/
* nanoblogger: http://nanoblogger.sourceforge.net/
Known Bugs
----------
* It seems that certain chars are not propely encoded in cookies, possibly broken chars: :&%[+ Fix should go in cgilib.rc^set_cookie (Have to figure out the proper way to scape cookie strings first)
* If a dir under apps/ doesn't contain an app.rc file, werc fails to start. A possible fix would be to replace $werc_apps default with `apps/*/app.rc` instead of `apps/*/`, this would be backwards incompatible, but I doubt anyone uses that option.
* Links in Blagh feeds become confused if markdown 'references' are used, markdown references suck, but I guess we need to address this somehow...
* Somewhat similarly to the abouve, relative urls in imgs, links, etc. can easily break when used in Blagh posts, so it is not all markdown's fault.
* .md files of the following form, without a new line after the last = seem to cause problems in some setups:
Foo Bar
=======
Fixed or WFM Bugs
-----------------
* Sitemap can handle $dirfilter inconsistently from other places (**fixed**).
* Finding directories that are inaccessible (-rx) can generate an infinite loop (**Works for me? Fixed with new menu code?**)
* Some cat-v.org bits still left over here and there, should make them configurable (1.0 blocker) (**fixed**?)