ORACLE with a Windows Client
Date: 7 Dec 1993 07:29:33 GMT
Message-ID: <lockley.51.0_at_boutek.csir.co.za>
If this is in an FAQ somewhere, I apologise and request that I get pointed
in the correct direction.
We have an ORACLE forms application which we are porting to Windows. For
the sake of this discussion, the application consists of a collection
of forms which are run selectively from our own menu system. We are
presently running the forms individually from the OS shell and NOT from our
menu system. I would like to ask people on the net what are the best options
and their experiences because we are having lots of problems.
Briefly our setup. We bought Forms 4 for Windows with the intention of
connecting to an ORACLE server across our local Ethernet LAN.
Problems:
i) We have to start Windows in real mode (win /s). This is a real (pun)
problem because we need to run other applications simultaneously with
the forms which *require* Windows protected mode.
ii) The Forms 4 system is slow and requires loads of memory.
We did some basic timing and came up the following table (25MHz 486DX):
+-----------+-----------+-----------+-----------+
| 16M RAM | 20M RAM | 24M RAM | 32M RAM |
| F3.0 F4.0| F3.0 F4.0| F3.0 F4.0| F3.0 F4.0|
+-----------+-----------+-----------+-----------+
1. Loading IDE |11.54 38.10|10.80 25.07|10.59 25.57|10.64 20.56|
2. Loading form | 1.31 35.49| 1.49 10.39| 1.32 10.92| 1.38 10.35|
3. Compiling form | 4.24 51.98| 4.20 9.32| 4.12 9.37| 4.18 10.12|
4. <F7><F8> | 1.88 5.07| 0.87 1.50| 0.91 0.97| 0.90 0.90|
+-----------+-----------+-----------+-----------+
Let me explain: F3.0 refers to Forms 3 running under DOS while F4.0 is
Forms 4 running under Windows. The times are measured in seconds and
were measured using a hand stopwatch. I don't think times should
be taken as absolute, what is more important is the relative
differences between F3.0 and F4.0. Yes I know that Windows programs
run slower than DOS programs, but I can't believe that that is the
sole reason for these timing differences.
1. Loading IDE: From the prompt, calling the interactive development
environment and waiting until it is possible to start working.
It can be successfully argued that this time is not important
because it is only executed once, but the order of difference is a
little worrying. I don't have the figures for runform, but I am sure
the ratios will be about the same (educated guess; not fact).
2. Loading form: Loading a form into the IDE for editing.
3. Compiling form: Compiling a form in the IDE.
4. Loading all the data into a running form using the <F7><F8> key
combination. Although this time is dependant on the amount of data
in the table (in this case, there was very little data), the order
of difference is a little worrying for the smaller memory
configurations.
Although these timing figures are not conclusive (and in no ways
should be interpreted as such), they are worrying in that they show
that Forms 4 for Windows can't really be considered a production
environment. Again if someone can argue or demonstrate that my timing
results can be improved, I'll be really happy listen.
What we are now considering is throwing out Forms 4 and replacing it with a 3rd party product which is able to connect and interact with a remote ORACLE server.
Here are my questions:
i) Are there people who are using Forms 4 for Windows? Do they
experience they same performace problems I have sketched above? ii) What 3rd party product would be suitable for our needs? MicroSoft
Access? Borland Paradox? Lotus Approach?
I would appreciate anybody's opinion on this matter. If there is enough response/intrest, I'll summarise and post to the net.
Thanks,
Graeme
Received on Tue Dec 07 1993 - 08:29:33 CET