Database on a server

[Roger.E.Eller at sealedair.com]

Have you considered the "no overhead" database?
What I mean is simply using text files in a folder
hierarchy as database records. Then your front-end
client can simply rename the text file to something
like "record0356.locked" when the user is making
a change to it. I have done this for years with pretty
good success. The key is to make the folder shared
on the network, but don't tell the users that it is not
a REAL database so they don't play around in the
folders which contain these text file records.

~Roger Eller <roger.e.eller at sealedair.com>

> --- Jan Schenkel <janschenkel at yahoo.com> wrote:
> Hi Richard,
>
> Sorry to disappoint you, but MetaCard/RunRev is
> inherently a single-user aplication, with roots in
> Unix. It loads an entire stack into memory, but
> doesn't "lock" the file in any way. Thus, it may seem
> that multiple users can open a stack and work with it.
> However, if you want to "save" the data from memory
> back to file, the troubles start because you cannot
> know which user has the correct version in memory, if
> any one user has it at all.
> This is no problem if nothing has to be written back
> to the disk, but if you really want a database, use
> RunRev as front-end for MySQL, Valentina, Oracle or
> any ODBC-capable DBMS.
>
> Hope this explained a few things,
>
> Jan Schenkel.
>
> --- Richard Hillen <mail at richard-hillen.de> wrote:
>> Hello list,
>>
>> I made a database using the splash-screen method,
>> compiled it as a
>> standalone and put it on a server, so that my
>> co-workes can use it also.
>> They should use it reading and writing.