Automatic Saving

[Linas Vepstas]

On Mon, Jan 06, 2003 at 12:32:42PM -0500, Derek Atkins was heard to remark:
> linas@linas.org (Linas Vepstas) writes:
> 
> > yes.  I'm trying to minimize the amount of crap in pull-down menus,
> > minimize the number of buttons.   The idea is that by removing rote
> > chores, the user is freed to concentrate on the important stuff
> > (instead of thinking, 'gee I have to hit save now').   Its a design
> > philosphy, its hard to put into practice.  
> 
> Hmm.. Then how often does gtt save its data? 

Every minute.  (and the archive consists of a copy of each of the 
last four minutes, and then copies on 4-minute, 16-minute, 64-minute,
etc. intervals.)

> I like the ability to
> force it to save before I shutdown my machine (save early, save
> often). 

gtt auto-saves if you ctrl-c it, or if you just quit out.

I added the archive feature because I've had several catastrophic,
total data losses while working with development versions :-(

> I'm not sure I like this.  While I acknowledge that avoiding UI
> clutter is a good thing, the name of the data file is NOT, imho,
> something that should be ignored.  Clearly you can/should provide a
> default name, but the user should be able to easily say "I want to
> look at _this_ data file".

I'm following the example of e.g. evolution, and some other newer
microsoft-based apps, which don't allow you to do this.  The fundamental
question is 'why would you want to look at other data files?'  I'm
guessing that the only reason you want to look at other gtt files
is because of a deficiency in gtt itself, and the 'right thing'
is to fix that deficiency. 

> Cool.  Unfortunately I'm still on a gnome-1.4 system so I can't
> play with this (yet).

?  If you're on debian, you can install both sets of libraries
side-by-side.  (including the devel's).  I presume this is true for
redhat, suse, etc.

> > > I'd also like to see gnotime interface into gnucash's invoicing system
> > > and customer management system.
> 
> I dont know -- I think we'd have to agree on some interface to do so.
> This was more of a pipe-dream kind of thing, but some way to script a
> new invoice in gnucash would be a good thing, IMHO.

Some pipe-dreams are easy to realize, some are hard.  Won't know which
till you try.

Consider the problem of fetching customer info from gnucash.

plan z -- gtt reads gnucash files and picks out customer names.
plan y -- gtt uses a gnucash libarary to read customer names
plan x -- gtt fires up a gnucash bonobo gui componenent that allows
          user to select customer, and then works with that customer
          name.

Right now, plan x seems to be the 'best' plan.  But I know little about
it.  The bonobo stuff is less than easy. 

> Well, I think we'll have to do that.  right now there isn't any way to
> import invoice data into gnucash.  OFX wont help; OFX only inputs
> transactional data, not invoice data (which is separate), and has no
> way to tie it to company information.

Is this a deficiency of OFX, or of the importer?  

> Well, all I would think of importing into gnotime would be a list of
> customers/projects.

I'm hoping to add a generic import, of things like ascii bulleted lists,
of html <dl> lists, <ul> and <ol> lists, etc.  But this would only
import lists of projects.  I don't really want to invent a customer
management system.


I sure wish the evolution coders were a little older/more experienced.
The GUI looks nice, but its not configurable; we can't modify thier
addressbook to act as a generic customer management system.    That's
one of the things I'm looking for.

--linas

-- 
pub  1024D/01045933 2001-02-01 Linas Vepstas (Labas!) <linas@linas.org>
PGP Key fingerprint = 8305 2521 6000 0B5E 8984  3F54 64A9 9A82 0104 5933