Mathieu Tozer's Dev Blog

Posting this here because it's probably the most coherant bit of writing I've done for several weeks! And it explains what I'm doing at the moment, even with all the typos.
--
Jim,

Since we haven't met up for a while, I'll give you a quick overview of
where I'm at with the project. I've been working really hard on it (to
the unfortunate detriment of my other studies) and I'm really excited
about where its heading.

First, sorry for not getting back to you on both accounts re the meeting
and the objects! I have my application working now for onine Dict
protocol dictionaries, which work quite well, and I have recently added
basic support for stardict dictionaries, in a bundled format specific to
my application. One of the goals for Words is to make it dead simple for
users to download and install new dictionaries, and I'm working on a new
interface to do this. I'm having issues with syncs while multithreading I
think, although I haven't attempted to debug that part yet. I'm still not
clear on how to use xjdic (minus the x for X11 part of course) to support
the Edict format within the context of my application, but I'm still keen
to

I've also had some changes to the design of the user interface, which
changes the proposed interaction for the app, and I've thought of a new
direction in which I would like to take the networked side of things.

Following seeing this
video:http://video.google.com/videoplay?docid=-8246463980976635143 I was
insired to aggregate user's word lists and definitions, uncluding images
and sounds, into a central database, in order to construct a user
generated multi language dictionary.

Using some nifty networked, server side wizardry, the details of which I
haven't even begun to fathom, you could make some very fun, data
validating two player quiz games incorporating all the aggregated media.

This also could potentially solve the hard problem of parsing the output
of dictionary searches. Because Words can / will connect to many
different dictionaries, it would be infeasable for me to write a parser
for each one to fine tune the formatting. And who can decide on what
definition is 'main' and should be used for quizzing. I cannot see how a
computer can determine this. At the moment, I call what dictionaries spit
out a 'definitionsBlob' or just a 'blob', and the user has the option of
setting at the moment 4 'main' definitions, which will be used in
quizzing etc. When networked, Words will first query the centralised
dictionary, and if there is a match, it will set these fields for you,
based on other users' submissions on the same word.

And if there is a user community, I could create a plugin architecture
for other interested people to write the 'blob' formatter for specific
dictionaries for pretty printing.

I've also decided against the smaller, minimised view in favour of a
speech bubble which will appear above a word once it is highlighted and
invoked by a right click menu or hotkey anywhere in the system, which
will give you your definitions.

Comments / Suggestions always welcome! It would be great if you had a Mac
so you could test it ;-)

-Mathieu