On 9/7/06, Tony Maro <[EMAIL PROTECTED]> wrote:
On Wed, 2006-09-06 at 23:55 -0700, Adrian Maier wrote:
> If you start a new application, I see no reason not to use something
> better - that supports goodies like transactions and SQL . For a
> new stand-alone application I'd choose embedded-firebird with no
> hesitations. If later it is decided to make the application multi-user, the
> change to the client-server version of firebird would be easy.
Its a question of the right tool for the job, not a question of is a
screwdriver superior to the hammer. Each tool fits a different use.
Sure, each tool has its place (or else it would cease to exist because of
the lack of users).
The question is whether the features of a more complex tool is worth the
overhead that it adds, in a specific situation. Each should of course do
his/her own evaluation and choose what fits best.
I'd use TDbf for a networked 5 user LAN environment, but if you want
something that works under more latency, for instance over the Internet, or
if you have more users you'll probably want a client-server transaction
based DB model.
My preference for a sql-based dbms is not neccesarily related to the number
of concurrent users, but to the features that are available to the developer.
For example, joining manually 4 tables with sums and 'group by' is not fun at
all, compared to executing a simple sql query . It would have been a huge
pain to generate the reports in my applications without a SQL engine !
If you want a single user application (or a few networked users) that is
database enabled and doesn't require the end-user to know how to set up a
database server, then products like TDbf are the answer.
Or products like sqllite, firebird ... ;-)
This discussion seems reminiscent of old discussions over networked use of
Paradox versus going with... what was the C/S database engine Borland
supplied?
Are you referring to Interbase?
So if you want to compare TDbf you'd compare it against text-file based
database engines, Paradox, dBase components, etc. Other than the
possibility of porting an old open-sourced text-file database engine for
Delphi, I think TDbf is probably unique in it's class for Lazarus support.
As a producer of commercial software for non-technical end-users, I'd never
consider producing a MySQL, Firebird or MsSQL application.
Aren't most of the applications made for non-techincal users (except from
compilers, IDEs and other development tools) ? :-)
The end-user level of expertise is too high for my market. With TDbf I only
have
to let the user pick a directory to store the data in, and provide an easy way
to
do backups.
What extra actions should the end-users perform if the application uses sqllite
or embedded-firebird ? If would be even cleaner because the database
is a single
file instead of a directory full of dbf and ndx files.
On the topic of embedded firebird, I've never used it in Lazarus (is it
even available and stable in Lazarus?) but the comment on their website "The
embedded runtime is < 2 MB" is reason enough for me to stick with TDbf. My
total app with TDbf is smaller than that.
You surely know better how to choose your tools according to the
specific details
of the application(s) you build .
Cheers !
Adrian Maier
_________________________________________________________________
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
