Re: Minutes of the GNOME Board meeting 10 July 2001

[linas linas org (Linas Vepstas)]

On Mon, Jul 16, 2001 at 03:20:04PM -0500, Dan Mueth was heard to remark:
> 
> First...  Thanks Linas for the feedback.  

And thanks for the freindly reply.

> Linas - Could you send the URL or date and list of the email you are
> referring to?

April 6th, sent to board-list gnome org, advisory-board gnome org

There seems to be a profusion of mailing lists, most of which are not
mentioned on public web pages (e.g. the above two), and which do not
have public archives (e.g. the above two...)  Maybe that's intentional,
but its a bit confusing.

I am reattaching the note below. Its, uhh, hopefully provocative,
which is why I wondered why there was no reply...

--linas

Subject: Re: Minutes of the GNOME Advisory Board Meeting
Date: Fri, 6 Apr 2001 13:16:08 -0500 (CDT)
Cc: board-list gnome org, advisory-board gnome org
In-Reply-To: <20010406090948 A15333 imag fr> from "Daniel Veillard" at Apr 06, 2001 09:09:48 AM
From: linas linas org


It's been rumoured that Daniel Veillard said:
> 
> 	       Minutes of the GNOME Advisory Board Meeting
> 	       ===========================================
> 		      5 April 2001 Copenhagen

Very Interesting.

I wish I could have been there; instead, let me takle a few minutes with
this email to express some of my ideas.


>    
>    GNOME 1.4 and 2.0:
>    ------------------
>          1.4                              2.0
>     compatible platform           Coexistence with 1.4 libs
>     Nautilus replacing gmc        I18N, switching to Unicode
>     general UI                    Accessibility API (similar to Java)
>     GNOME VFS                     Usability testing
>     Bonobo                        Platform revision
>     GNOME conf                    Interoperability


The gnome-2.0 list has must-have items on it, but its, well, a bit
boring.   I think we need to have two lists for gnome 2.0:
the "gnome-2.0 yes-dummy,you-must-pay-attention-to-the-details" list
and the "gnome-2.0 grand-vision-for-the-future-list".

>     GNOME-2.0 main issue is [...] making sure
>     nothing will be forgotten. 

I think everyone understands that we must have even more
internationalization, and further-improved usability, and easier
maintenance and install, and etc.  These "details" make or break the
thing.  And I think gnome developers will indeed do a good job of these
things.

But we must also have a vision:

>     Userbase:
>     ---------
>         => growing will requires breaking on other market, basically
> 	   - office market
> 	   - desktop appliance

webtop. YOPY. that grey area between large laptops and palm pilots.

>       Developers ratios:
>         - 7 M VB 
> 	- 3 M Java
> 	- Perl/Python
> 	- 300 K Unix

The Microsoft .net strategy will allow 7M VB programmers to create
network-enabled, client-server, network-object-model applications
with the twitch of the nose.   This should scare the bejesus out of us.

>    Rob Sun:
>      - web services
>      - integration

Funny, nobody mentioned integration between the office suite and
PHP-office.  You think I'm joking?

Today, we have pure-desktop apps that are mostly single-user and 
mostly network-ignorant: gnumeric, star-office, abiword, (and sadly,
gnucash, although we are working to change that).

Then we have the pure-network apps: PHP office, SQL-ledger, bugzilla.
they require a functional browser, and nothing else. 

There is a giant split between 'web services' and 'desktop apps'.

There are a few exceptions: Quicken and MSMoney are desktop apps 
that are network savvy: they are a kind of 'financial browser for 
the network'.  Graphical CVS clients are 'network-aware desktop 
apps'. Nautilus shows some hints of blurring the desktop
boundary.  Palm pilots: network-aware, hot-syncing palm pilots 
and thier kin really are blurring the boundary.

Microsoft is essentially in the same boat.  But they have a plan:
its called .net

It will allow 7M VB developers to write 'desktop applications' that
are instantly network-enabled,  And with a different keypress, thier 
VB tool will spit out the matching "server application" to go with it.

There is no way in hell that any linux app could possibly be
'compatible' in this scenario: the .net client and the .net server will
fit hand-in-glove, two parts of an intricate puzzle that will be
impossible to dis-entangle.

=============
So when we talk 'web services' and 'integration', we must really talk
the meta-question: how can we integrate, even in principle, something
like star-office and something like php-office?   Can gNOMe create
network objects that make it easy to create 'desktop apps' that run
network-aware, are (at leaast partly) usable with ordinary browsers,
and, btw, run on my YOPY, my iPAQ?

>    Jim Getty:
>      - interoperability at all levels
>      - computing will shift over more distributed and less desktop model
>      - don't fight the battle of the past, rebuilding W95 is not the way
>        to succeed

Jim gets it. 

> 1:15 pm What is GNOME:
> ----------------------
>    Havoc: we need to do a good job on our existing customer i.e. the desktop
>           this involves fixing 

Yes, "good job on our existing customer" is vital, critical, extremely
important for success.  Its in the "keep-your-eye-on-the-details"
category,  we need to do this *and* have that vision thing.


>           short term [...] look at appliances

'appliances' will be hard until you solve the 'how do I sync the
appliance with the server' question.

>    DV: need to integrate with Windows services being deployed through SOAP/.NET

SOAP alone is not enough. Need corba-on-soap. Need 'glade-to-soap':  Use
Glade to drag-n-drop a nice GUI, and you hit a button, and glade spits
out the matching corba IDL, the matching SOAP.  Derek Neighbors, the 
GnuE guys (www.gnue.org) understand this: the have a forms designer: 
when you press the button, a GUI falls out, *and* the matching SQL 
falls out.  The GUI's they can auto-gen are gtk, java (maybe?), 
curses(???), and win32(!).  I keep telling them, the Glade guys need 
to get with them, work on this together, not apart.  And the GnuE guys 
need to expand beyond SQL, and add support for SOAP, for corba.

>    Havoc: 
>     1/ Do a good job of a Unix stable usable desktop
>     2/ Pervasive computing, i.e. extend to other appliances
>     3/ frontal opposition to Windows won't work

Yes, but the *only* path to 2/ and 3/ is through the
network-object-model, which gnome has pretty much failed to deliver on.

>       => users don't care if it is a GNOME or KDE app they want it to behave
>          in predictable ways

Users don't care if its gnome, KDE or a web-site, they want it to behave in
predictable way.   We *MUST* integrate with Wiki, and with Zope, and
with Slashcode.   Nautilus should *not only* display slashdot correctly,
but should also do the expand/collapse/threshold/moderate/post functions of 
slashdot *ENTIRELY* on the client side, so that you don't have to wait
for osdn's network bottleneck to respond. So that you can do simple
things locally and quickly.

Because, mark my words, in 3 years, someone will have this kind of
slashcode implemented in Microsoft's .net and we will look like chumps.



--linas


_______________________________________________
advisory-board mailing list
advisory-board gnome org
http://mail.gnome.org/mailman/listinfo/advisory-board