On Thursday 07 September 2006 00:23, Michael Van Canneyt wrote:
> On Wed, 6 Sep 2006, Micha Nelissen wrote:
> > Michael Van Canneyt wrote:
> >>> Not scalable ? In what way ?
> >>
> >> It's not multi-user and network-capable.
> >
> > This is also simply not true. It is multi-user capable, and can use
> > files from a network share, but it's all done in the client, using file
> > locking.
>
> And that's where it goes wrong.
>
> > It does not support transactions, true; but this is a feature, not about
> > scalability.
>
> I'm not talking about transactions.
>
> While I agree with you that in theory it should work. In practice it simply
> does not.
>
> I will invite you to pay some visits to schools where they use a
> administration system, based on DBAse files on a network share. The first
> rule is: only 1 user can edit a given table at a certain time. Failure to
> comply with this rule results in corrupted DBase files.
>
> For correctness, I must also say that the applications are not written
> using Delphi, but using FoxPro.
>
> Michael.
I'm sorry Michael but I have been reading this thread and had to say
something. I'm have been a VFP (Visual FoxPro) programmer for the last 20
years and I can say you are wrong - period. I have a 100 sites running fox
using dbf's all multiuser and without any major issues. Several of the sites
are 50 users or more (total data size from 3G to 10G). File locking at the
row level (in fact all versions of fox supported row locking) and
transactions are supported and have been since version 5 (currently on
version 9). The only real issue with using dbf's is the size of the indexes.
In a lan environment the index is transferred to the client workstation which
can slow the program. But it is very stable. In fact all of my clients use
foxpro for general accounting. In other words a program central to their
companies.
Now that I have said the above I have to say that dbf's do have limitations
and in general SQL engines (MsSQL, MySQL, Postgres and others) offer more.
However, I would also like to say that for small amounts of data (under a
million records) dbf's can be faster than using an SQL engine.
John
_________________________________________________________________
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject
archives at http://www.lazarus.freepascal.org/mailarchives
