Re: [Conky-development] Conky roadmap

[Jason T. <jt...@gm...>]

Hey everyone,

I have no idea who's actually reading this email, but this crossroads
in Conky's development seems like a good time to introduce myself. I
hope others do the same, so we can all get to know who we're working
with a little better.

My name is Jason Tan. I am a 19-year-old student majoring in Computer
Engineering at the University of Washington, in lovely Seattle. My
upcoming school year will be my last, and after that I hope to find a
good job where I can make tons of money and drink gin and juice.....
laaaaaaaaid back. Anyway, my interests lie in software engineering and
programming on a team. I've been using Linux for about a half-year
now. Started with Mandriva (UGH!!), tried Fedora Core 4 and didn't
like having a whole bunch of crap on my computer that I didn't know
about, and then I finally decided upon Ubuntu, which has worked
beautifully for me. I have programmed quite a bit in Java and Python,
but am not unfamiliar with C/C++.

As for my availability, I am very easily reached by email and am
almost always in #conky (even if i'm afk). My fall quarter of school
doesn't start until very late September, and I will have the most time
available right now while I only have to work 40 hours/week.

Anyway, I am ready to take on the challenge of an overhaul to C++ if
you guys are, but it'll be important to do careful planning and
designing so things don't end up as ugly as torsmo did.

In the meantime, I've been porting the docs to a system that'll be
much easier to maintain and generate all the different formats from.
What's next on my mind is to refresh my html/css and see if I can cook
up a small renovation of our five-minute website, and also look into
making a debian package for Conky.

Team Conky, over and out. let's kick ass!
--Jason

On 8/21/05, Brenden Matthews <br...@rt...> wrote:
> I want to discuss Conky's future, and how everyone can be a part of it.  =
First
> I'm going to start with a little background to make sure everyone is on t=
he
> same page.  Oh yeah, and sorry for the length of this email, but I wanted=
 to
> get everything across.
>=20
> As you may already know, Conky is based on the original torsmo code.  Tor=
smo
> is (was?) a rather popular light-weight system monitor for X.  Torsmo was
> fairly well regarded for it's ability to let you display just about any t=
ext
> on the screen, as well as built-in support for other useful things.  Howe=
ver,
> for whatever reason, the original torsmo maintainers decided they didn't =
feel
> like working on torsmo one day.
>=20
> After trying to write some patches for torsmo and being frustrated at the=
 lack
> of response from the maintainers, some people suggested I fork it as my o=
wn
> project.  At first I brushed off the idea, since it seemed a little
> disrespectful to the original developers and I didn't think I was up to t=
he
> task.  However, I ended up starting Conky anyway.
>=20
> So, with that, I made my first Conky 1.0 release from the last stable tor=
smo
> release (I never touched the torsmo CVS code).  Since then, I have added
> several new features to Conky, and have developed a (small) userbase.
>=20
> There are a lot of things I don't like about Conky the way it is.  First,=
 the
> original torsmo code was very ugly, and therefore Conky is quite ugly.  I=
t's
> written in C, and makes heavy use of confusing macros, making it quite
> difficult for people to jump into the code and hack it.
>=20
> As it stands now, Conky has a few bugs which can't really be fixed withou=
t a
> major overhaul of the core of Conky.  As well, there aren't many more
> features which can be implemented without completely re-thinking how Conk=
y
> works.  With that in mind, I'm considering rewriting the core of Conky in=
 C++
> to make it easier for people to work with the code, and to clean things u=
p.
>=20
> One feature Conky could really use is network transparency.  However, thi=
s
> could only be implemented with filthy hacks given the current framework.
> Since most of Conky is a filthy hack, I want to steer away from this and
> build some more maintainable code.
>=20
> So, with that said, I want to open the discussion on whether or not we sh=
ould
> work toward an object-oriented C++ Conky.  My thoughts are that we end
> development on the C 1.x tree, and begin a C++ 2.x tree.  We will fix bug=
s in
> the 1.x tree until 2.0 is done, but new features will be minimal.
>=20
> Now for some more technical stuff.  The way Conky currently works, it
> basically prints everything into a character array (text buffer) and then
> passes this to the X draw function.  For special stuff like bars, graphs,
> alignment et cetera Conky simple looks for a "special character" within t=
he
> character array, and then does whats need (it's just a switch/case).  Thi=
s is
> a really useless way of doing things, because it's a total pain to do
> anything that isn't text.
>=20
> So, I'm thinking of using a set of classes and one big linked-list in Con=
ky 2
> (if we decide to do the C++ thing) to store the stuff to be rendered as w=
ell
> as evaluting expressions.  With a linked list, it wouldn't matter what th=
e
> content is, because we can just have one big structure (or union) with al=
l
> possible objects as the node for the list.
>=20
> There will be plenty to do, so I would like to assign certain tasks to ce=
rtain
> people and I will count on you to get things done, or you could simply ch=
oose
> something you'd like to do, and I'll make sure 2 people aren't working on=
 the
> same thing.  There is no particular time period in which I'd like to see
> things happen, and I'm not even going to speculate, however I'm ready to =
code
> now.
>=20
> Hopefully someone will read through all of this, and I look forward to he=
aring
> some feedback.
> --
> Regards,
>=20
> Brenden Matthews
> Conky maintainer
> <br...@rt...>
>=20
>=20
>