[nycphp-talk] Microsoft Access application conversion recommendations
Paul Houle
paul at devonianfarm.com
Sun Mar 11 22:09:13 EDT 2007
David Krings wrote:
>
> My point is to foremost forget about the old system and make a new
> shiny one. Then see on how to get old data ported over, ideally the
> least amount with which you can get away with.
>
> Just the 2€ from someone who never did anything like this....
>
It's hard to say. You really ought to consider this option, but it
depends on how you feel about the quality of the existing database.
I recently did some work on a project that was botched by another
programmer -- the presenting problem was that a particular feature
didn't work right. Looking deep inside, I saw that certain database
structures were hopeless, making the feature impossible to implement in
a reliable and maintainable way. It took about 100 hours to understand
the requirements, change the database structure and rework the rest of
the application to use the new database structure. After all that, it
took 5 hours to implement the new feature. What's great about it is
that the new design reflects the business reality so well that it
reflects scenarios that the client and I never thought of -- and can
sometimes be accomodated in 15 minutes to 2 hours worth of work.
If the old database design is good, it makes sense to keep the
database. If the old database makes you want to punch the wall, then
you'd better consider switching. Even though desktop databases are
often overgrown spreadsheets, I've seen Access databases that are as
beautifully designed as any Oracle or MySQL database.
If you keep the existing database, you can have the PHP and Access
versions running in parallel. You might port the 20% of functions that
80% of people use to PHP first, and make a gradual transition towards
eliminating the Access version. This approach takes a lot of risk out
of the transition -- one problem with the "Build System B and migrate to
it" approach is that you might build system B from scratch and think
that it's really great, but discover that the migration isn't really
possible once you try it.
More information about the talk
mailing list