Re: more vs. better (was Re: morning hack...)


Subject: Re: more vs. better (was Re: morning hack...)
From: Martin Sevior (msevior@mccubbin.ph.unimelb.edu.au)
Date: Tue Apr 17 2001 - 17:35:20 CDT


On Tue, 17 Apr 2001, Paul Rohr wrote:

>
> The idea which makes me itch is seeing core developers starting new features
> instead of finishing what we've already got.
>
> >But our list goes further. Martin posted a TODO list that we came up with at
> >GUADEC. I want this list to be fluid, at least for the next week or two. We
> >can all evaluate what needs to be done, and what we'd like to see in the
> >release.
>
> Yeah, I still need to read that list. After I finish wading through raw
> email, I'll start opening attachments. (Sigh.)
>

The 1.0 list we made is not very controversial in the sense of adding
radically more new features except for Bonobo support.

Of the 1.0 list the one that scares me the most is double buffering
drawing in the GTK build. I believe that is the only way to fix bug 119
and for Abi to show the performance people expect of a modern WP on Unix.

The Perl scripting and Macros code is almost done already. Dom has just
polished off 90% the multi-lingual Language stuff. With the new
ap_Preview_Abi class getting templates going is not too hard for
Dom. Templates are a really useful feature for Church secretaries.

Now regarding Bonobo, I think having this capability is essential for
Abi's long term viability. The Gnome project really needs an embedable WP
and a WP for embedding. I also really want to use Abi for my everyday
email client, which will be trivial once we have bonobo support in place.

As we've seen already from Paolo, there is a TON of high quality talent in
the Gnome community. If we can supply them with a light weight yet
sophisticated WP component I think we'll naturally drag many talented
developers our way. Everybody realizes that a WP is the cornerstone of a
sophisticated document. People will want their useful code to either embed
Abi or to embed their app in Abi. If we had full bonobo support we could
use the Charting and graphing component, GUPPI,
Gnumeric will soon be ready, gnome-db is ready, Dia the Visio clone is
ready, SodiPodi the vector graphics editting component is ready and we
could embed Abi in Evolution and Nautilus. Many of our top developers
joined the Abi project to give GNOME the cornerstone WP it needs. We want
Abi to link these great Apps together and to be used throughout gnome.

There is a political aspect to this too from Open Office but I won't go
there now.

There is one more feature Abi needs in addition to Bonobo though. It needs
XP SVG graphics support so it can DISPLAY compound documents on non-Gnome
platforms and to speed up the display of embedded gnome components. The
idea is to not only store the XML fileformat of the embedded gnome
components but also a SVG rendering of the last time they were accessed. I
believe we can do this via a generalization of our image handling code and
the framework Justin Bradford put in place. I think that with a bit of
cajoling we can get the gnome-print guys to add SVG to their collections
of rendered outputs. So we can just ask each app to make a SVG image of
itself via gnome-print.

So the final issue is how do we get native SVG support on all our
platforms? Well GNOME has this library called librsvg which been developed
by the Nautilus project. With luck we can reuse this code.

I personally plan to finish off the styles dialog, then help with the
replacement of the Lists dialog. I will keep the current Modeless list
dialog somewhere though because it will be useful for implementing a cool
way to do collapsing/expanding of list and heading structures. (This is
actually not hard but given that we want a 1.0 I'll leave it for now.)

After that I'll look at getting SVG graphics implemented. I'm not quite
sure what Dom and Joaquin plans are but I believe they're both interested
in getting Bonobo going. I can helpout on the Abi side but the GTK object
system which is needed for Bonobo has defeated me till now. I'm hoping
that Dom and Joaquin's code will teach me how it works.

So what do I see happening until 0.90? Something like this:

1 month to implement Bonobo/styles/Lists/SVG graphics by say May 15th
deadline.

In that time other platform developers should try to get their platform up
to the GTK level in Styles/Lists/background. We should not release 0.90
until the dialogs on at least Win32 and Unix are equivalent.

From what I understand of Bonobo a one month implementation time is not
unrealistic now the API has finally frozen but it is really up to Dom and
maybe Joaquin.

I guess the other platform developers will need at least 1 month to catch
up with the gtk build.

After the 0.90 release I'd like to see a weekly release containing nothing
but bugfixes. ie Every Monday we release Abi 0.91, 0.92, 0.93 with source
code and whatever binaries are easy. I volenteer to make gtk rpm's. Tom
Briggs seems to have fully automated building of Win32 binaries so we can
release those in step too.

Well that is my vision of the Grand Plan. With luck we might make a 1.0
release by some time in Summer. Actually for me, only bug in bugzilla that
scares me is 119. If someone would like to try implementing double
buffering for gtk drawing code I would be very grateful. Read Havoc's book
and look at Dia and the gnome-canvas for example code on how they do it.

I'll happily give advice and maybe write up a POW on it.

Cheers!

Martin



This archive was generated by hypermail 2b25 : Tue Apr 17 2001 - 17:35:29 CDT