Hello,
Having read this letter, and none of the previous in the conversation
I am going to add my bit anyway.
I am very much in favour of 'embedded intelligence' schemes.
They provide a way of getting at-a-glance information on the object to
which they refer.
The argument given they create more work is untrue. I believe the work
involved in maintaining a database that provides the information a meaningless
name leaves out at least equals any work involved with 'embedded intelligence',
and also, it would be faster to update 500 objects within an application than
it would be to manually update a paper or electronic database with the details
were a change required.
There is also the disadvatage with meaningless schemes in that, if you
need to know a property of a particular object then it involves first
identifying the object in the application, and then referring to the database
of definitions to discover the information required. This creates unnecesary
worktime that would be completely eliminated if an intelligent scheme was used.
There is also the problem of keeping the external reference up to date.
Whenever I have tried to set up any scheme that relies on continuous update by
the users of the scheme to reflect changes, Esp. when this update requires the
user to change to a different application to make the update, the users simply
do not do it resulting in out of date information. Which is very bad, because
it is misleading...
The claim that intelligent chemes are bound to fail through unforseen
requirements is equally valid for meaningless schemes as the format of any
external database would also have to change. The whole point of analysing the
scheme and identifying present and future requirments is to reduce the chances
of this happening.
...Which I assume anyone would do, because you cant just sit there and
come up with a scheme in your head and slap it down, can you...!? (But we
all do.... don't we?).
Sorry if I butted in, but I have had a similar argument with a
|
