Development Ideas

[Todd Kulesza <todd dropline net>]

Hello!

Drivel 2.0 turned out to be a huge release--within days the package was 
accepted into Debian-unstable and Fedora Extras, and even with these two 
 huge repositories shouldering most of the load, the number of 
SourceForge.net downloads far surpassed anything we've previously seen. 
 It looks like we have a number of new users and, hopefully, new 
developers as well :)

So, here are my thoughts on the next development cycle.  This is all 
just brainstorming right now, and I'd love to hear input from others.  I 
just want to get these ideas circulating so that if someone's looking 
for a place to start hacking, they have some concept of what the overall 
plan is.  This list might seem a tad ambitious, which is probably is. 
I'd like to stabelize for a release in 6-9 months, depending on the 
speed of development.  If there's enough interest to accomplish 
everything, then awesome.  If not, we'll punt whatever is left to the 
next development cycle.

1) Panel Applet - It would rock to have an applet in the GNOME Panel 
that would handle the authentication stuff for Drivel.  When a user 
clicks on it, up pops an editing window, instantly ready to go.

2) Journal Extensions - LiveJournal supports all sorts of extensions, 
like "no comments", "auto format", etc., and Drivel has a UI to handle 
them.  MovableType also supports extensions, but Drivel doesn't have an 
interface for any of them save categories.

3) Security Groups - It would be cool to be able to edit LiveJournal 
security groups from within Drivel.  I made a mock-up of the UI in Glade 
and it's already in CVS, but nothing works and I'm not sure I'm happy 
with the interface.

4) Security Group Selection - LiveJournal allows a user to select 
multiple security groups for each entry, but Drivel's UI only allows 
one.  If this could be fixed without resorting to a pop-up window UI, 
that would be sweet.  I'm not sure if GTK+ supports check-boxs in 
GtkComboBoxes, but that makes the most sense to me.

5) Category Selection - Same story as Security Group selection--fixing 
this without a pop-up window UI would be great.

6) Entry Preview - If Drivel could preview entries in HTML before 
posting, it would go a long way to alleviating the lack of a WYSIWYG editor.

7) WYSIWYG Editing - I wouldn't want to eliminate the plain HTML editing 
we have now, but a WYSIWYG mode would be cool.

8) Preferences - Currently, Drivel stores preferences per-account.  I'm 
thinking that should be ripped out and done per-user.  This was 
originally done because Drivel started out as a Win32 application, and I 
could see multiple users on the same PC sharing the same profile.

9) Journal Profiles - The Login window should be redesigned to support 
profiles for popular blogging sites.  Each profile would have the 
default server and API preconfigured.  I've already started a mockup for 
this, but haven't had a chance to finish it yet.

Before any of these can really be tackled, some other items need to be 
addressed:

1) The code is a mess.  I've been started to restructure the internals, 
but it won't be easy.  I'd like to get away from the huge, monolithic 
DrivelClient struct with pointers to everything.  This work has only 
begun, but I'll post again when I have an outline of how the different 
components could/should interact.  Any ideas here are more than welcome :)

As a side effect, I'd like to restructure the code to support 
command-line posting.  This might be a big step in getting the Panel 
Applet to work, as well.

2) The networking layer, which was beatiful for the 1.2 release, turned 
hideous as more and more protocols were added.  It could be drastically 
simplified.  I'm not sure whether to keep the existing libcurl backend, 
or look at moving to libsoap, as it's part of the GNOME Desktop stack.

So, thoughts, comments, ideas?

Cheers,
Todd