/**
 * @var bool
 */
@var is highlighted. If the comment starts as "/*" instead of "/**", the highlighting is gone, which is to be expected. Wouldn't it be good to extend this distinct behavior further ahead, by adding a little more, namely: when pressing <enter> within the "/**" comment block, the new line should start with the usual space padding, followed by the " * ". The thing is that at least for phpdoc, lines within the "/**" comment block, not starting with an asterisk, are ignored by documentation generator. So one has to manually add asterisks at the beginning of each line. Auto adding it would be a great help.

The Python API should use the "property" function (http://docs.python.org/library/functions.html#property) wherever possible to make accessing and setting information about ST2's current state more seamless and Pythonic. This will make it even easier to write more and better plugins, and since it's a backwards-incompatible change, the time to make it is *now*, while ST2 is still in alpha/beta state.

All methods that take no arguments and exist only to return information should be decorated with "@property" so that other functions don't need to use parentheses to call the method. Some examples of methods that could use this treatment: sublime.windows(), sublime.version(), view.id(), view.file_name(), view.is_dirty(), view.settings(), region.size(), region.empty(), window.id(), window.views(), window.active_group(), etc. This allows for easier chaining: in the context of a plugin, it would be nice to refer to self.view.window.folders[i], instead of self.view.window().folders()[i].

Some objects have a pair of getter/setter methods: for example, view.name() and view.set_name(name). The "property" function allows you to combine these into one property, so that plugin writers can find out the name of the view by accessing view.name and set the name of the view by assigning a string to view.name (like so: view.name = "my view"). This is much more readable and Pythonic than having a separate method just to do assignment. 

Add API for Open and Save file dialog.

This is useful  when a plugin would like user to choose an converted output for current view.

Since Ruby uses the pipe symbol to demarcate block parameters, it would be nice if the pipe behaved in a way analogous to parentheses, square brackets, and curly brackets: if text is selected when entering the character, instead of replacing said text with the character, place one at the beginning and one at the end of selected text.

If I'm not mistaken, Sublime Text uses hunspell dictionaries for its spell checking mechanism. hunspell is also used in LibreOffice and can be used in Vim as well.

Unfortunately, Sublime Text marks words as incorrectly spelled although they're definitely correct. As a result, almost haft of my document is underlined in red. 

I'm using hunspell-de-de (German dictionary) on Ubuntu 12.10, which is identical with the dictionary in the github repository.

As a verification, I opened my document in LibreOffice, Vim and hunspell's interactive console and everything was fine.

Could you please look into that. Right now, the spell checking mechanism seems completely useless.

Common shift + (arrow/mouse) selections that are tedious and can be potentially replaced by a triple click:

1. Selecting the text within a single or double quoted string
   
2. Selecting the enclosure of a pair of parentheses, brackets or angle brackets
3. Selecting a name-value pair (particularly in XML)
4. Other commonly copied and pasted language constructs

It may be difficult to make these language aware, but at a minimum, selecting an enclosure based on the following as delimiters may be possible: ['"<>(){},]

Is it possible to navigate FindResults from keyboard?

like,

1, ctrl+shift+f, find "test"

2, Press 'F7' to next match, 'F6' for previous match

Currently this could only done by double clicking the result line, and I must open the FindResult tab to find other matches.

I'm using Sublime Text 2 build 2051 on ubuntu linux 10.04 LTS 64-bit .  I'm experiencing a bug where performing repeated search-and-replace where replace is empty automatically fills the replace block after the first run-through with the previous replace input.