A web browser specifically designed for reading Wikipedia. (As opposed to AutoWikiBrowser, which is a web browser specifically designed for editing Wikipedia.)
http://hackyourday.com/2008/02/05/a-new-way-to-browse-wikipedia/
Has anyone used this? What's it like for the experienced Wikipedia editor?
- d.
On Feb 5, 2008 11:11 AM, David Gerard dgerard@gmail.com wrote:
A web browser specifically designed for reading Wikipedia. (As opposed to AutoWikiBrowser, which is a web browser specifically designed for editing Wikipedia.)
http://hackyourday.com/2008/02/05/a-new-way-to-browse-wikipedia/
Has anyone used this? What's it like for the experienced Wikipedia editor?
Next to useless, I'm afraid. No browser chrome, i.e. no way to go back or forward. No links in text, which sort of defeats the whole purpose of having a hyperlinked encyclopedia. Text window too small, can't resize it, can't hide the images, can't scroll or enlarge the images... nah. Oh, and I managed to have it crash in less than no time. :)
On the other hand: TreeBrowse (http://yoy.be/item.asp?i115) *is* a wonderfully interesting way to browse through Wikipedia. Screenshot here: http://www.vuijlsteke.be/michel/treebrowse.png
Michel
You know, I never saw that. But I am on a mac, what do I know? Maybe it could get ported to Mac soon. But until then, it's still Safari for me.
-Soxred93
On Feb 5, 2008, at 5:11 AM, David Gerard wrote:
A web browser specifically designed for reading Wikipedia. (As opposed to AutoWikiBrowser, which is a web browser specifically designed for editing Wikipedia.)
http://hackyourday.com/2008/02/05/a-new-way-to-browse-wikipedia/
Has anyone used this? What's it like for the experienced Wikipedia editor?
- d.
WikiEN-l mailing list WikiEN-l@lists.wikimedia.org To unsubscribe from this mailing list, visit: http://lists.wikimedia.org/mailman/listinfo/wikien-l
On 2/5/08, David Gerard dgerard@gmail.com wrote:
A web browser specifically designed for reading Wikipedia. (As opposed to AutoWikiBrowser, which is a web browser specifically designed for editing Wikipedia.)
The idea's good. Magnus wrote something similar. I wish someone would develop it a bit further.
What's good: - images help with random exploration - list of links is a *great* way to hone in on something you only know of vaguely. - fixed contents box
What's not: - only being able to see a chunk of text at a time - seems like some of the links are actually links to images, and the result is totally mangled. - clunky, amateurish
In general, I think it makes a lot of sense to browse wikipedia wiht a fixed GUI, rather than downloading it all again on each page.
Steve
On 05/02/2008, Steve Bennett stevagewp@gmail.com wrote:
In general, I think it makes a lot of sense to browse wikipedia wiht a fixed GUI, rather than downloading it all again on each page.
Yeah. Download the wikitext and do stuff with it. Of course, this would presently require embedded PHP in the browser to implement the present parser ...
- d.
On Feb 5, 2008 12:59 PM, David Gerard dgerard@gmail.com wrote:
Yeah. Download the wikitext and do stuff with it. Of course, this would presently require embedded PHP in the browser to implement the present parser ...
If there is a context-free grammar definition of the MediaWiki syntax somewhere, it can easily be used to generate parsers in basically any existing language.
On 05/02/2008, Chris Howie cdhowie@gmail.com wrote:
On Feb 5, 2008 12:59 PM, David Gerard dgerard@gmail.com wrote:
Yeah. Download the wikitext and do stuff with it. Of course, this would presently require embedded PHP in the browser to implement the present parser ...
If there is a context-free grammar definition of the MediaWiki syntax somewhere, it can easily be used to generate parsers in basically any existing language.
I said that in answer to Steve because he's the one working on the latest shot at such a thing ;-)
http://www.mediawiki.org/wiki/Markup_spec http://www.mediawiki.org/wiki/Markup_spec/ANTLR
The present situation is that wikitext syntax is *literally* defined as "whatever the parser does." This is, of course, far less than ideal for all sorts of reasons. There's been several attempts to remedy this - because a properly-specified grammar will make all sorts of interesting applications possible without having to either reverse-engineer the present parser or just embed PHP - but all so far have failed. Let's see how this one goes ...
- d.
On 2/6/08, David Gerard dgerard@gmail.com wrote:
I said that in answer to Steve because he's the one working on the latest shot at such a thing ;-)
Yeah, your post and Chris's made me smile :)
For Chris, wikitext is context-sensitive so a context-free grammar is not possible, but a useable ANTLR grammar is quite feasible. I'm actually getting pretty close. I'm just doing battle with tables atm, which are difficult because the syntax is pretty ambiguous, and because table cells can contain *anything* - including other tables.
Steve
On Feb 6, 2008 8:50 AM, Steve Bennett stevagewp@gmail.com wrote:
On 2/6/08, David Gerard dgerard@gmail.com wrote:
I said that in answer to Steve because he's the one working on the latest shot at such a thing ;-)
Yeah, your post and Chris's made me smile :)
For Chris, wikitext is context-sensitive so a context-free grammar is not possible, but a useable ANTLR grammar is quite feasible. I'm actually getting pretty close. I'm just doing battle with tables atm, which are difficult because the syntax is pretty ambiguous, and because table cells can contain *anything* - including other tables.
When you get it done let me know. There is a C# generator for ANTLR grammars, so I could integrate this with WikiBench for, e.g., realtime previews during editing (minus templates).
On 06/02/2008, Chris Howie cdhowie@gmail.com wrote:
On Feb 6, 2008 8:50 AM, Steve Bennett stevagewp@gmail.com wrote:
For Chris, wikitext is context-sensitive so a context-free grammar is not possible, but a useable ANTLR grammar is quite feasible. I'm actually getting pretty close. I'm just doing battle with tables atm, which are difficult because the syntax is pretty ambiguous, and because table cells can contain *anything* - including other tables.
When you get it done let me know. There is a C# generator for ANTLR grammars, so I could integrate this with WikiBench for, e.g., realtime previews during editing (minus templates).
Yeah, WYSIWYG is the big winning application of such a thing. The mediawiki.org pages set out where we're at with this beast, and there's the wikitext-l list where this particular piece of vapourware is being condensed - you have an application, so joining the list (and reading the archive) would be a good and useful thing. Same for anyone else who has a use for good third-party wikitext parsers - MediaWiki wikitext appears at present mostly to be made of edge cases ...
- d.