Re: [gnome-db] Future characteristic



Adam Tauno WIlliams wrote:

I think you're right.  This really limits the incentive to use Gnome-Db/GDA, if
one still has to do all the form get/post grunt work.  And it needs to be
supported by some RAD tool, at least something like Glade.
I still remember how powerful Borland's tools were for building simple database
apps back when I developed on Windows-For-Workgroups (gawd!).  At this point on
Linux the OpenOffice form builder and database connectivity is the closest thing
we have.
1.) Define a connection
2.) Define a query
3.) Make a "form" (or screen whatevver) , attach it to a query
4.) Populate the form with widgets attached to query elements.
Lots of fancy stuff is nice, but the above satisfies 8 out of 10 situations (in
my experience).
Its not quite that easy for building DB RAD tools. Inter-form relationships, inter-widget relationships, cross dependencies of both widgets and data, cause and effect on user actions, callback functions and when to call/not call callback functions, dealing with datasources that cover many tables, data validation, default values (always more of a nightmare than you expect), and then making sure the data on the screen is upto date with whats on the database and vis-verser. The whole web post/get thing is so much easier data model to work but its these kinda features most users demand of there databases. Maybe i need to code less and sleep more. Coffee is evil. Anyway point is that you have a limited use for software if you can only do as good as what you can do on the web.

I'm not talking about web forms;  but functionality provided by a RAD to
a very-long-time-ago.  The reason to move from a Web to a Client
interface is often simple one of speed and data-density (it is easier to
pack more on a Client window than arrange sanely on a web page).

Even OOo handles inter-form relationships, detail views, etc...  but
firing up an office package to provide a form seems like whacking a fly
with an anvil (although CPU & RAM are so cheap these days I'm pretty
close to just not caring anymore).

I understand about the complexities, the joy of callbacks, data
validation, etc...  I'm currently writing an OpenGroupware client for
GNOME, yikes!  But solving many basic needs will attract, I suspect,
more inertia and interest making the ugly plumbing easier.  One of the
most frequent complaints about the Linux Desktop that I hear is a lack
of an Access equivalent.  Currently what OOo provides is the nearest
thing (that I know of).
Sorry, I can be a bit permissistic when it comes to databases,
What you think is 90% of the work turns out to be 10% of the work.
Forms are never as easy as they seem. It be good if OOo could read .mdb files. Maybe in a 3 years we can all have everything sorted.
If someone hasn't checked out the forms system in OOo they should take a
look just to see it, for potential inspiration if nothing else.  It does
linking of subordinate forms, detail views, etc...

BTW, Does bond use GDA now?  Last I looked at it it was for pgSQL only.
It uses gda for backend . but i havn't ported the data dictonary over yet and that runs postgresql specific queries so if you use gda it does bad things when you run your app. Hopefully next release will have gda support sorted.
I wanted to keep it userbase small until i got a lot of core things sorted.




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]