97->2003 conversion "gotchas" ?

[Gregory K.]

If an Acc97 app converts to 2003 and tests successfully, can I assume that
all is well ? We have an Acc97 app with 10 users (more details below) which
must be converted to 2003. I'm suffering from nagging anxiety that even if
it runs successfully on my PC, problems might show up later (performance
degradation compared to Acc97, corruption ....) The app is used by a
transportation firm, so reliability and (reasonable) speed are important.

Regards,
Gregory K.

DETAILS: datafile ~ 105 Mb 25 tables, 10 users (most on Metaframe 1.8);
occasional slowness in response time with 97, though app has been optimized
wherever possible; migrating to Win2003 servers and Terminal Services; tech
people want 2003 on T.S. due to concerns about excessive CPU usage by 97
[though this was never a problem on Metaframe]; application is fairly
"routine" (standard Access objects, no ActiveX components, no replication, no
security used).

 

[Allen Browne]

See:
Converting from Access 97
at:
http://members.iinet.net.au/~allenbrowne/ser-48.html

The article contains suggestions on setting up, converting, and usability
issues. It does not relate specifically to Terminal Services.

 

[david epsom dot com dot au]

There have been some minor changes in Jet SQL, like this:

ACC2000: Changes in Use of DISTINCTROW in Microsoft Access 2000
http://support.microsoft.com/default.aspx?scid=kb;en-us;207761

I don't have a URL, but we also found that Jet 4.0 was more
particular about type compatibility in parameter queries.

And the 'macro security' dialogs of A2003 are causing some
irritation.

Our biggest problem in converting from A97 to A2K was the
tendency of A2K to crash instead of giving a runtime error
message. Hopefully you won't have that problem with A2K3.

We are seeing some unexplained very slow network performance
with A2K3, but at the moment it is still unexplained. We have
made all the standard optimisations, but the A2K MDE running
on A2K3 is still 5-10 times slower on some reports than the
A97 version. (Novell or Win2K server, ok with local data).

-----
Although the 'excessive CPU usage' of A97 is a well
documented non-issue, use of TS is still a valid justification
for CONSIDERING movement away from A97, because although
the CPU usage value is false, it reflects unjustified use
of the Terminal Server channel.

-----
I would be at least as worried about the move from
Metaframe to TS on 2003. There have always been
occasional problems with TS like this:
http://support.microsoft.com/default.aspx?kbid=818528)
and no particular reason to expect 2003 to be
any different.

Also note that networking to a Win2003 file server can be
very slow if the server is configured for general server
duties rather than file sharing (packet signing forces
packet serialisation, which is incompatible with ordinary
file share optimisations):
http://www.smallbizserver.net/Default.aspx?tabid=98


(david)